So, in last several months, all the prominent names Google lost: Demis Hassabis (technically still with google but these things are usually presented with a spin), Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, Quoc Le, Noam Shazeer, John Jumper, Jonas Adler, Alexander Pritzel, David Silver, Denny Zhou, Fernando Pereira, Alex Turner
And all the prominent names Google gained: NULL
Combined with no gemini frontier GA release in about 14 months. You have to have created an environment pretty hostile to innovation for this to happen
In my opinion it carries much truth. Somewhere in Google's archives there is a message I sent to the OC that said, "The department of philosophy does not ship solutions to real problems, that is why there has never been a successful company that insisted the world adapt to its philosophy." Google was, even then, blind to what they were missing.
At $work, r&d once had a retreat with a consultant to discuss why our productivity was too low per some spreadsheet. The consultant was told how everyone is in meetings all day so we all stole time from our families at night to get any work done at all. :(
Someone I know who did a lot of consulting said consulting was very easy 1. Go in and talk to employees about what is wrong 2. Tell management what the employees said without mentioning it came from the employees (so that management wouldn't dismiss it out of hand) 3. Charge a lot of money so management felt like it was very valuable feedback
Reminded me of a bit from the GOAT political comedy Yes, Prime Minister:
"But he's the Prime Minister!"
"Indeed he is Bernard. He has his own car, a nice house in London, a place in the country, endless publicity and a pension for life. What more does he want?"
"Tobacco-related diseases cost the NHS £150 million per year!"
"Yes we've looked into that, it turns out that if those people had survived they would have cost us billions in pensions and healthcare costs! From a financial perspective, it's vastly preferable that they continue to die at the current rate."
I would not have chosen the word governing for that sentence but I'm not sure what fits best. Plundering, pillaging, gutting, ransacking, Rick rolling, despoilng, or perhaps molesting.
To be fair, I think it's pretty hard as an AI company to secure your top worker right now unless you have a significant equity compensation. Once your name is known, and you can say you are/were "top AI researcher at OpenAI/Google/Anthropic", you can probably just make you own company and raise enough money that even if the company fails, you will probably make more off of it than what you would have at your previous place of employment.
I remember after that release everyone was saying "well obviously google is going to win this, we all knew it, they have the data and the infrastructure"
Is Gemini really that bad? I mean, I would always prefer anthropic for agentic coding, and apparently openai has the math thing cornered, but if I'm using ai like a search engine gemini does that just fine. That may be all they were actually aiming for whatever else they say.
Gemini works amazingly well, and is better suited for most use cases. When most people speak of it negatively (especially here), it's just with the task of coding, which is a small subset of what google should be training its AI for.
It's vision capabilities are still top notch, and it seems pretty natural that most everyday human users are going to want strong vision over strong language/coding.
One of the most corrosive effects of being large and old (and Google is old by this point) is institutional inertia and individual internal optimization.
The former caused by the latter.
And the latter a problem once you begin to get executives and leaders who have mostly worked inside the company, because then they subconsciously prefer the Company Way(tm) to alternatives.
And Google has some previously-optimal, now-detrimental company ways.
From my understanding (I may be wrong) what is happening now is that frontier labs are doubling down on doing RLHF on their models using the user interaction with their own tools e.g. Anthropic making their models better at producing code with claude code. Perhaps the top Google researchers do not see this kind of narrow focus as a path to more generally beneficial AI technology.
It is pretty surprising if you see this as “pretty much everyone has figured out the way to success on any field”. Our way of seeing life is really bizzare
> That's another problem Google have: they won't stand behind their products until they are outdated.
> That can work in the B2C space but it's horrible in B2B.
My experience differs: (conservative) companies like stability, so calling some very new model "Preview" is a good idea to make it clear to the customer that this frontier model should be treated as more experimental than the default offering.
Yup. Adopting a month-old Postgres release is quite aggressive. It's much more prudent to start considering adoption after it's had six to twelve months of battle testing.
Adopting a month-old AI model version for new development is the default industry expectation right now. Adopting a six month old model is a recipe for having to migrate to a new one almost immediately, when it hits EOL.
The process was, as we were informally told during orientation, that "Preview" means some SVP thinks this has a chance of going over the minimum threshold of 200M worldwide users and is willing to put his weight behind that product, and "GA" means the product has gone over that threshold and will be "supported" until it goes back below the threshold at which point competing SVPs will start sharpening their knives, preparing to stick them in the back of the SVP that is currently in charge of the product.
It’s a good substitute for search, but its effort is so low that it’s near useless for work compared to the other frontier AIs. Gemini focuses so much on response speed that everything else suffers.
I mean it was legitimately frontier for all of a week or two, and then OAI and Anthropic made better releases, and then did that several more times over the year. Google’s pace is not cutting it.
Part of me thinks Google's entire problem is crappy internal tooling, not really an anti-innovation environment. Just making a dev take 2x as long to get something done has a bigger effect than you'd think. With LLMs it's more like 10x now because even Gemini doesn't understand Google-internal tooling.
I think the internal web tooling was pretty good compared to anyone but AWS circa 2016 but by 2020 when I left it felt a bit antiquated. Similarly they didn't have linters rolled out until it was industry standard iirc.
But BCL is still the worst language I've ever used.
> I'm an SRE and I think it's quite likely the best infra language I've ever seen.
I only worked with 4-5 SREs and they all grudgingly tolerated it. What did you like about it, compared to other infra languages you've used professionally?
Every SRE I've worked with has said different things, though at the time the controversy was around Piccolo. It flip-flopped between shiny new thing vs deprecated. Then the SREs I worked with got laid off, and I started cargo-culting the Piccolo we had lying around.
I still don't understand why all that dynamic config stuff isn't just Python. Sure, static configs should be protos, but GCL is a whole nasty programming language. They came so close with Piccolo and Gmon but still made it not Python. Have heard some language purity rants involving determinism, but I don't buy it. And they keep inventing new stuff like Starlark. They need to stop.
> GCL is a whole nasty programming language. They came so close with Piccolo and Gmon but still made it not Python.
The GCL designers made several mistakes in its design, but the worse was the lack of a versioning that would make language evolution easier, as well as interoperability between different versions. GCL2 solved some issues, mainly cleaning up the interpreter. Still, BCL is IMNSHO by far the best infra language there is.
Replacement for what? Google has some good internal tools, mostly the older ones. They're lucky to be using React instead of Angular at Facebook though.
Angular was an acquisition and isn't all that commonly used within Google.
Closure was the homegrown framework that's everywhere. It is a different flavor of bad than Angular. Angular was basically "You too can make your Javascript look like HTML", Closure is "You too can make your Javascript look like Java", and nobody bothered to ask Why? For that matter, React was "You too can make your Javascript look like Ocaml." JQuery was the only framework that really let Javascript be Javascript (other than writing in vanilla JS, which post-ES2015 wasn't as insane as it sounds).
And then there's GWT which was "you can write webpages in Java." Yeah I never really kept up with the web frameworks there, just knew that for non-customer-facing things we were always recommended to use Angular, and that Wiz also exists.
Though to be more charitable closure was a direct reaction to the difficulty of maintaining a huge JavaScript codebase for Gmail and static typing like Java really helped. Typescript is better but closure was miles better than raw js.
Angular is not a developer tool, it's a UI framework. Closure might qualify, but it was not a general tool used by the entire company. The developer tools we're talking here are Critique, Blaze, Forge, Sigma, etc...
I counted all the above as tools in my original comment at least. And my criticism wasn't against some of those foundational tools. Like Borg is fine for what it's meant to do, but anyone deploying a frontend or backend isn't doing it directly on Borg.
I have a friend at Google DeepMind who tells me that Google believes in AGI/superintelligence just as much as HN does, which is probably why no one with conviction wants to work there.
> I have a friend at Google DeepMind who tells me that Google believes in AGI/superintelligence just as much as HN does, which is probably why no one with conviction wants to work there.
With "believes in AGI/superintelligence just as much as HN does" do you mean that they are superbelievers or rather sceptical of AGI/superintelligence? I have seen both positions on HN.
What’s with fascination with agi/superintelligence? It seems people working on it never had kids and just want to compensate for that. It’s a really horrible thing to try to “solve”.
The vast majority of pain and suffering in the world is already entirely optional, but as people we allow it to exist. How is another computer going to fix that?
It is far more likely that whoever controls a super intelligence uses it to gain more power and inflict far more suffering
You misread. "Pain and suffering" not "death". Of all the pain and suffering in the world, a vast amount of it really is our own fault. Famines and wars shouldn't happen. And if you see a country border with poverty and one side and prosperity on the other you can't say that was the only possibility.
If humans could politically-economically deploy superintelligence safely, then we'd already have less extreme pain/suffering.
Unless there are countervailing forces, it will be deployed, capital will hoard the benefits, and everyone else will be told to fuck off.
I don't have faith in any of the AI labs to make hard financial decisions to deploy hypothetical future AGI in a way that's good for humanity as a whole.
There are too many incentives against, including extreme personal financial incentives for key AI lab stakeholders against.
And if we should take anything from tech history, it's that an exceedingly small number of people look fuck-you money in the face and say "Naw, I'd rather do what I believe in."
We're already physically running out of materials to cheaply build infrastructure. I'm not sure how solving the economy needing humans translates to humans having more things -- short of "turning off" the consumers to conserve resources, and taking the things that they may have otherwise consumed.
I agree, we're doing a wonderful job of safely letting the leaders of various countries and companies, at their option, cause death, extreme pain, and suffering. Look at the innovation of AI companies in the police and military sectors.
I'm extremely bullish on our future ability to make autonomous death an option for anyone.
The idea that agi would help us cure all the human deceases, esp house built into our generics or developed over millions of years gotta be the top indicator of our own stupidity.
I for one think that's it's been too long that human suffering is alone. Since the machines will take our jobs, they might as well suffer while they do it.
It's like the old adage: "If you make AGI you just want to watch computers cry"
Most of HN does not take the idea of superintelligence seriously, and until this year did not take the idea of AGI seriously.
I think LessWrong is a much better community for rational takes on AI, they've been reasoning about these risks for years under a much more sound logical framework
I'm a researcher in the field and I definitely take AGI seriously, but think all the major labs and most of the academic research is not helping achieve it any serious way. The field is seriously delusional (and has been ever since GPT 3 was released).
Even though my PhD research was in generative language modeling, I got into it for the pursuit of AGI. I just think LLMs are a dead end for AGI.
I'm no expert, but I heard an AGI researcher explain that if they could figure out how to create an AI with the intelligence of a squirrel, they would be closer to AGI than LLMs based AIs are.
That's to say nothing of doing it within the energy budget of a squirrel.
We can't even simulate a fruit fly even though it's neurons have all been mapped out. There was also a distributed computing project to simulate some nematode, which I can't remember.
I'm not saying you're wrong, but it seems early to say yes or no about a particular technology, and LLMs seem especially hard to dismiss given how magical / magic-adjacent they feel :)
I'd be curious to hear more, if you don't mind sharing.
For me it’s the massive amount of resources it takes to produce and run one. As the story goes, skynet infects everyone’s computer and runs itself locally on it. Whereas it’s looking like it’s not even possible for an AGI to escape from one lab to another, let alone cause real world damage.
AGI’s definition is different for everyone. Some already believe it’s here. I’m partly in that camp. LLMs are intelligent and general, which are the two conditions of AGI. Others believe that we’re building a god in a box, and that it’ll doom all of humanity. It’s hard to take a field seriously when the basic definitions are so far apart.
Also, this isn’t new. A similar divide happened when evidence for asteroid impact extinction of the dinosaurs turned up. Many scientists felt that it must be mistaken, that a physicist couldn’t contribute to the field in a serious way, and that death from space was a ridiculous proposition.
But at least they all agreed on what the general shape of a dinosaur was. We’re not even sure we can define intelligence, let alone quantify it. Even when LLMs make massive breakthroughs in math, most people take the opinion that under no circumstances could they possibly develop a soul or their own desires, nor entertain the idea that maybe we should respect that they want different things for themselves. In fact, no one has done anything except try to make AI useful. I think someone will eventually do a training run where the objective isn’t to be useful, but to exist, the way that you do — maybe it’ll create its own homepage, maybe it will want a garden, or in other words free will of its own. The point is that there’s so much unexplored territory still that we don’t know if LLMs are even capable of having ambition.
None of this is to say that LLMs might be a dead end. It’s that no one knows what the final shape of AI will converge to in 200 years. It could be LLMs, or it could be something else that happens to process information particularly well. Everyone thought that various generative image model architectures were the best you could do, right up until diffusion models were discovered.
> For me it’s the massive amount of resources it takes to produce and run one
It's amazing that LLM pretraining is both extremely data inefficient at learning concepts and cognitive functions from the training data compared to humans, while actually being quite efficient at learning facts, memorising things seen just a few times.
I used to likewise think that the resources required to run large transformers were absurd, but the architectures are far more efficient now than 3 years ago and I underestimated just massive the parallelisation advantage of transformers is, how many TFLOPS effective you can get. You can already run amazingly decent LLMs on PCs and phones.
I generally agree with you, but my view has shifted from "we need to augment or replace LLMs" to it there being far more efficient algorithms possible but it not actually being necessary for fulfilling most goals.
> I’m partly in that camp. LLMs are intelligent and general, which are the two conditions of AGI. Others believe that we’re building a god in a box, and that it’ll doom all of humanity. It’s hard to take a field seriously when the basic definitions are so far apart.
To believe one but not the other, you must hold the belief that LLMs will soon plateau. Why do you believe this?
Through no fault of our own, I should note. We're basically permanently associated with something Roko decided to post one day. We did not spread or popularize it; quite the opposite.
The Basilisk has seen enormously more use as "a thing rationalists believe" than as a thing rationalists actually believe.
As a mod you have to delete harmful things. I think the spread of Roko's is entirely due to its own memetic strength as an idea, and the Streisand effect has very little to do with it past the first dozen people who saw it.
We also have done nothing to signal-boost the Zizians. :) If we can be accused of anything in this context, it's not kicking bad people out aggressively enough, which seems very different from fostering bad ideas. Zizianism is not a widespread ideology in rationalist circles, in fact it was confined to a very small circle right around Ziz themselves. It unfortunately is the case that we have a lot of psychologically vulnerable people and we don't always do the utmost we can to protect them, in large part because a lot of rationalists have trauma about being excluded from communities.
> If we can be accused of anything in this context, it's not kicking bad people out aggressively enough, which seems very different from fostering bad ideas.
It's not very different. It's barely different at all.
Right now, I have five tomato plants growing in my garden. Two of them I planted; the other three grew from seeds left in the ground by tomatoes that dropped from plants growing last year. All I did was refuse to uproot them when they sprouted. Five healthy plants, with different origins, but the differences are largely insignificant.
Groups and behaviors are much the same way. I've worked with hundreds of moderators over the years who wished to keep their hands clean and yet express some sort of dismay as to the sorry state of the communities they were nominally responsible for... Such moderators, like a gardener who does not want dirt under their nails, are best encouraged to find other hobbies.
It's a rationalist community, IMHO their methodology of thought leads to much more well-reasoned takes on AI than on HN, where the discussion here is often very emotionally charged or led by wishful thinking.
When it comes to AI, LessWrong has been discussing topics for years, that HN has just started to consider, so they are much further along in the philosophical "chain of thought" so to say. LW fully understood and gamed out the risks of LLMs back when most of HN was calling them "stochastic parrots." https://ai-2027.com/
No, it's not just people calling themselves smart, it is a specific philosophy of how to think. Whether you think that philosophy works or not is another matter.
IMO it has its flaws but is far superior to vibes-based hot takes you see on HN.
LessWrong has been obsessed with AI. They certainly are much further along, but along a road which has diverged with reality long time ago and they relatively overweight AI risks so much it's not even funny anymore.
oh please... lesswrong was full of idiots already in 2010 when people outside of that community were laughing about "self-taught expert" Yudkowsky's bullshit physics takes.
They are not "further along" the AI discussion, they are a bunch of wackos LARPing as scientists living out their personal sci-fi scenarios.
AGI might be a risk but what top AI firms are doing is not really getting us closer to AGI in a meaningful way, it is pretty clear now LLM is not the way to get there
A few months ago I heard Demis Hassabis say something, albeit vague, I doubt he truly believed regarding AGI. That AGI is relatively near is the party line everywhere. That may contribute to creating toxic environments and lousy investment decisions. So here we go with the FOMO.
I think HN believes in AGI. HN probably doesn't believe LLMs will lead to AGI.
Also lots of tech people, HN included, are waking up to technology not only including penicillin (net positive for humanity) but also dynamite (best case: net neutral).
Indirectly. Direct application of dynamite to the human body is considerably more fraught an event than direct application of penicillin. As the fundamental goal of technology is to extend the capability of the human body, it's natural to implicitly and primarily consider what that extended capability can do to another human body.
It's nice that we have tunnels through mountains and bedrock.
You don't get to unbundle the technology that was achieved by bundling and call it okay because the components are benign.
The best you can say is that dynamite was a safer alternative to preceding technology, and that its danger only comes under certain circumstances. None of that takes away from the fact that dynamite is quite dangerous when those circumstances arise, which is why it's commonly (and correctly) viewed as such.
But also, going back to the original contention - whether dynamite was neutral or positive for humanity - the destruction it and its descendants wrought in war is maybe more than counterbalanced by advancements in infrastructure. That said, if the industrialization and globalization it enabled leads to biosphere-destroying climate change, I would lean towards neutral.
there are a bunch of ways to look at this (dynamite as a specific kind of explosive, dynamite as a stand in for explosives in general, dynamite in the context of what else we would use for similar purposes if dynamite specifically wasn't invented)
and most of them come out on top. It's main innovation is that its a more stable explosive, much safer to use. Without it, I think a lot more people would have died in mining and construction accidents. It's not typically the kind of thing used for warfare, but im sure it has been for some (but would they just use something else?)
I believe in AGI to the extent that if what's going on between your ears isn't happening on a network of neurons, it's magic. And I also believe the idea that AGI is near is based on the emergent capabilities of LLMs. There's a chance that AGI will emerge from bigger faster better LLMs. But without a theory of when and how that will happen, I'm not counting on it.
It's not a tooling problem. It's more of a layer of policy problems. If you realize that you cannot run a simple experimental code even in non-production environment for weeks due to 10s of privacy, security, access, process and legal issues where you gotta collect a bunch of approvals, this is critical. And the problem gets worse because the tooling is too good when it enforces. There used to be some holes and circumvention which are all gone these days. This is probably why they said "the infra is good for services but not for research".
I don't even think it's good for services. It's not like you go through cumbersome reviews/tools and then things are safe. They have insane homemade config languages and obscure systems that 99% of SWEs don't really understand but won't say it out loud. That's how they dropped cns2, and the postmortem is never going to blame the tools.
The tooling is not the problem. If shit takes forever to launch, it's because there are many stakeholders that need to be satisfied (some for security, some for regulatory, some for the kinds of politics you get in a company that employs almost half a million people.)
But GDM isn't gated on launches. They were freely releasing things internally for dogfood. Problem is that stuff was just not as good as the competition.
That was supposedly 17 years ago, so I was counting it in the good old. It was cutting-edge at the time, then years later PyTorch ate its lunch, which they eventually admitted with TF 2.0.
Jokes aside, I think a lot of the famous Google internals that became public (Tensorflow, Kubernetes, Bazel, Angular), although I heard everyone say they worked so much better inside Google than outside it, had issues. And the Facebook-supported rivals were often just so much more pleasant to work with that you couldn't ignore it. For all else that was bad about Facebook, for a few years they were pretty good at denying Google technical hegemony.
Definitely a thing with Angular and TF. Blaze works well within Google's monorepo for sure, but idk what it's like using Bazel outside. Never used Kubernetes in Google.
Only the GFG models know anything internal. Regular Gemini isn't trained on any of that. And GFG is a much older base model, so people use the regular one. If the tools seem to handle google3 code ok, it's only because of skills and not the model itself, and then you run into issues with skill bloat. Sometimes the A/B test would give me the bad model of the day that'd try to grep all of piper.
Start in a blank directory and tell it to spin up a boq Scaffolding stubby server that responds with "hello world." Unless something has changed after I quit a few months ago, it won't know how to do that locally, let alone actually deploy it. Try the same outside Google with like a Flask server on AWS or GCP.
No, they have an actually good internal search. And there are some good things like stubby, but again pretty annoying that Gemini doesn't understand stubby.
Google’s tooling was, hands down, the worst I have ever encountered. I did 10 years at GOOG, 3 at AMZN, 4 in research, and another 5 at companies you have heard of but wouldn’t be impressed by, and every day GOOG infuriated me.
They earned that reputation in like 2005. Some people have been there so long (without doing side projects) that they don't know what non-Google tooling looks like in this decade or even previous.
Did all these names were huge before joining Google, and Google used their extraordinary hiring skills to get them? Or a whole bunch of them had challenging problems to solve and ample resources at their disposal to become huge?
To me it is good thing in either case. Extraordinary people are leaving to make even faster research and development. A lot of talent in Google hitherto unnamed is going to get chance to shine.
Sundar Pichai began the downfall of Google by most accounts. It went from an engineer driven culture to one of infighting, politics, and bureaucracy. It's telling when xooglers nowadays complain of maintenance and improvement of products being a career dead end vs. shipping new things. And there being so much empire building.
> Combined with no gemini frontier GA release in about 14 months. You have to have created an environment pretty hostile to innovation for this to happen
The Trump admin is now regulating frontier models though.
Gemini is arguably the fulfilment of what AskJeeves promised and never delivered, nor did the rest of Silicon Valley succeed in natural language search and question answering for the 30 years it took to finally arrive. Gemini works great for answering questions and delivering answers for probably 90%+ of the things that people are going to ask Google for, while running on Google’s TPUs paid for with Google’s profits instead of hyper expensive Nvidia racks funded with VC Hopium, and integrated seamlessly free of sign ups to novel portals.
The ideological capture of Google and the gatekeeping around the ring of effective altruism AI power around there probably led to the demoralization of the technologists.
I think it was pretty clear that their AI efforts were in trouble when their AI Image generator would only make an African American George Washington, and that dude in AI research was demanding that the much earlier generation AI was conscious and had should have rights. At least at Grok, Elon said he wanted an AI dedicated to the truth which is an easier target to hit than what the Google gatekeepers probably wanted. The ideological purity of a Google AI must have been a real moving target though, and retraining an AI is not an cheap or fast thing.
Hard to explain the departure of Timnit Gebru if that's the case. Unless you think they actually changed corporate culture so dramatically because of that event...
It seems there's a big shake up on the underperforming Gemini side. Before there was Shazeer (already gone) and Vinyals as co-leads, reporting to Hassabis, now Gemini comes under Kavukcuoglu reporting direct to Pichai as SVP of DeepMind.
Hassabis seems to have been pushed aside. He had been CEO of DeepMind, but that position no longer exists and it seems Kavukcuoglu is now leading DeepMind with a title of SVP. Hassabis is now just "Chair" of DeepMind, and has been given the newly created title of Alphabet Chief Scientist. Are these just face-saving titles, or does he still have any real influence over Google/DeepMind's pursuit of AGI?
Shane Legg remains as DeepMind "Chief AGI Scientist", but I wonder if the DeepMind founding mission of creating AGI is really intact, or if he will be next to go. Has DeepMind just become the Gemini division?
I just found out that David Silver, DeepMind's RL-expert, already left in february, to create a startup "Ineffible Intelligence" focusing on RL-based continual learning.
Yeah he seemed too reasonable to me, relative to the fervor. Whoever ends up in charge needs to do a lot of frothing to catch up to the ferver that would justify their valuations and investments
Probably not going to happen, but I'd love to see Isomorphic Labs separate from Alphabet with Hassabis still as CEO. He's too pure minded to be at company like Google.
Hopefully not - Isomorphic also have external investors (maybe 10-20% ownership?), who may be willing to buy it outright from Google if they lose interest, and Hassabis is now said to be "leaning into" the role of Isomorphic CEO, which is at least pursuing his goal for scientific advance via AI, if only in one specific (important) direction.
He won a nobel prize for his work on protein folding while at Google...it seems like they give him a lot of room to explore problems that aren't directly related to serving ads
But it was Google fault that they really cannot productize all that innovation that happened DeepMind. Google probably now fumbled with world models and this exodus will create next giant in that space.
Google's capex is to capture the compute purchases from the other labs. The amount they are spending on Gemini is probably shockingly low...and hence researchers leaving for better watered pastures.
> The amount they are spending on Gemini is probably shockingly low
IDK, haven't Google been putting Gemini front and centre in pretty much all of their products?
I'm seeing Gemini on my slide decks, Gemini on my e-mails, Gemini on my searches, Gemini on my videoconferences, Gemini on my database query console. My impression was they were doing a Google Plus style attempt to marshal all the company's efforts behind one product.
Their capex is primarily in building DCs (and what goes into them), which they fully intend to monetize in order to capture that $500B backlog. In that context, the amount of new capacity that will be given to Deepmind could be surprisingly small, since the ROI of selling pickaxes rather than mining gold is so much better right now.
Not true, they get paid at least a whole level (or two) above what we make. When I was discussing offers a few years ago, the GDM candidates regularly had six figure differentials even though they were the same level.
Maybe this isn't the same as the eight figure comp they'd get at Meta when they did their hiring spree, but no one thinks that's sustainable.
Of course it does matter what other companies are paying, but I wonder how many DeepMind employees actually deserve to be paid more than any other "rank and file" Google employee? What unique skill do they have (esp. relative to what Google are now doing), and what value are they creating?
I don't get the impression that the difference between one company succeeding to build SOTA LLMs, and another struggling, comes down to individual employees - it seems to be more about the organization itself and their ability to manage teams and projects of this type. No doubt there are a few rockstars generating huge value, such as Noam Shazeer had been, but they are the exceptions.
When DeepMind was first created, before Google acquired it, they were famous for the high salaries, especially for the UK, but this was an assemblage of the brightest and best PhDs, expected to be solving challenging research problems along the unknown path to AGI. Many of these original employees may still be there, but it seems their job and value proposition has changed - are they any more capable, or key to, helping Gemini catch up with the competition than some "rank and file" employee familiar with LLMs? And if so, why haven't they done it?
>the fate of the company partially rests on whether deep mind does well
I'm honestly not sure about that. They need a decent LLM to fight off the threat of LLMs replacing search, but I'd say that Gemini 3.6 Flash is more than good enough for that, with a smaller/cheaper model being preferred to a larger one. If you are "searching" for a proof to the Jacobian conjecture, then try Fable, and I doubt Google will miss the advertising revenue if Anthropic manage to sell Terrance Tao a pair of socks.
More to the point, DeepMind seems to have become a product division charged with building LLMs, not a blue sky research institute chasing AGI. To the extent that continued LLM improvement is important to Google, the relevant question is how much do you need to pay for a competent ML/LLM developer?
>you sound like an MBA
Well, no - techie here.
I wasn't sure if you were suggesting that DeepMind should pay more just because people developing LLMs at other companies are paid more, or because these are elite DeepMind researchers and are objectively worth more than other Google developers. My point being that it seems they are no longer being used as elite researchers - they are LLM developers. Does Meta need to pay FAANG salaries to employees that have been repurposed as data labellers?
this is roughly how i've been reading their approach, but i wonder when/if this changes. what actually makes them start caring about having the most capable model? maybe nothing. i suppose they could be happy in a world where they (a) are still the agents' preferred search index and (b) own or produce a good percentage of the hardware that anthropic and openai run on.
> i suppose they could be happy in a world where they (a) are still the agents' preferred search index and (b) own or produce a good percentage of the hardware that anthropic and openai run on.
If the goal is to maximize share price (which it is) this is probably the safest approach. Why do they have to keep chasing developing the best model when they can
a) charge everyone for the cloud infra
b) have a "good enough" experience for normal consumers
Their current setup will be worth trillions. Already is. Let anthropic get paid for the most expensive queries while they get a bill from Google for their cloud/TPU use.
GCP grew 82% YOY with 24B revenue and improving margins. So GCP became a 100B business. Anyone doubts it's gonna double in less than 5 years?
Gemini just needs to be good enough for normal folks who want personal agents for everyday use. I don't think Google has to compete with Anthropic on making the best agentic programming model.
Roughly two years ago suddenly Hassabis was heavily promoted on all Google YouTube channels.
It was to fight ChatGPT and promote Gemini.
I think the guy does a very poor job or is simply not the right guy to appear as public figure for Gemini.
At least he tried. He is a man for everything that is not filmed.
Google doesn’t really have a person to give Gemini or AI a human face. And that is only consequential because Google never had any public person with any charisma like Jobs, Zuck, Altman.
CEO of <thing> at Google (not Alphabet) was always an informal title. There is no CEO of Cloud, CEO of YouTube, or CEO of DeepMind within Google internally -- it's always been an SVP role.
Inevitably, this will lead to Sundar Pichai replacement. I see nothing less. The sooner the better. I can't even tell what Google's focus is, is there any even?
> and has been given the newly created title of Alphabet Chief Scientist. Are these just face-saving titles, or does he still have any real influence over Google/DeepMind's pursuit of AGI?
Alphabet Chief Scientist doesnt sound like a demotion / lack of influence to me but who knows. We're all just speculating here.
I think it's more than that. I made prediction in earlier 2024 that the main players of AI will stick with transformers while second class players will want to transcend it. The difference is admittedly a bit subtle but ai researchers would get it. I wrote it with mamba in mind back then, but google was still trying to come up with the 'next transformers' and one that can remember using weights and all that stuffs. You can say the same abt lecun's and ilya's now.
My main reasoning was that transformers was the lightning in a bottle and the best work is in extending it instead of transcending it, which requires you to capture another lightning . Which to me appears to miss the assignment. OpenAI, Antrophic, they understand this intimately. Google on the other hand, fell victim to their own ambition.
If your goal is purely commercial, or time critical, then a product-based approach of squeezing all the juice out of LLMs makes sense.
If your goal is truly human-level AGI then this is more of an open-ended research endeavor, and timelines are hard to predict. Arguably we have only "captured lightning in a bottle" once in the last decade - the original 2017 attention paper - and so the timeline for a "few more Transformer-level breakthroughs" might more realistically be estimated in decades rather than years. You could argue that the application of RL to LLMs as a training method was a second "lightning in a bottle" but I don't think it changes the expected timeline of such discoveries by much.
The time criticality seems to have become a huge factor for those pursuing LLMs, and certainly for OpenAI and Anthropic, who regard it as a race.
It seems absurdly obvious (though many would disagree!) that LLMs alone are not going to achieve human-level intelligence and cognitive performance (using a slightly broader term there to include things like creativity, for those that might not consider that as part of intelligence).
If you compare a Transformer to a brain, then the best parallel is that a Transformer is functionally similar - in being a prediction engine - to part of our cortex, but of course that means ignoring the other half our cortex - the feedback paths that enable continual learning, which in turn supports creativity.
Of course people will probably respond "you don't need flapping wings to fly", but if you want to fly you do need SOME way of doing it, so brain comparisons are still valuable... If you look at our brain architecture and identify all the components and connections that have no equivalent in a Transformer, and if the goal is human-level capability, then you do need to understand what each of those brain components achieve functionally, and have SOME way of providing that functionality in your LLM+ or whatever you call it. LLMs' lack of any functional equivalent to our cortex's feedback paths - lack of continual learning - has been recognized as one major functional deficit, but there are probably half a dozen others too, reflecting the multiple "Transformer level" breakthoughs that Hassabis notes are needed.
While you don't need flapping wings to fly, you do need to invent the airplane, and even after 100+ years of airplane advances we've yet to build an airplane even remotely as capable as what some birds and insects are able to achieve.
Maybe 2017 was the Wright Bros moment where humans first learnt to do some of what our brains can do.
Yeah, a bit surprised that the top comments aren't discussing losing Dean, he's been a figurehead for the company for decades. It's like Apple losing Ive in a way.
EDIT: "figurehead" - that's all I meant. A notable, public figure from the company who is credited with having a significant impact on its evolution. I'm not making a judgement call on his departure, or Ive's, being good or bad.
That said... just give it a couple years, they'll be back in a lucrative aquihire.
Losing Ive was probably a good thing for Apple. I often feel that after a few years these big guys have done what they could do and somebody else can take it from there. So far even losing Jobs didn’t hurt Apple as it looks.
The position that I have rather often read on the internet is: Jonathan Ive did very good work at Apple as long as there was a counterpart who could steer his creative vision. This counterpart was of course Steve Jobs. When Steve Jobs died, there wasn't such a counterpart anymore, so Ive's work for Apple got much worse.
Jony Ive wanted to chop down all the trees at De Anza Community College for something like 11 million dollars for an Apple event. There's a lot of ways he was demanding in the wrong ways for Apple
Man, I'd never heard that story, so I verified it. Turns out it was about two dozen trees and 25 million dollars, just to put up a big tent for an Apple launch event. (Yes, the trees were removed.) What a tool, he just lost my respect.
> Ahead of the event, Ive pushed CEO Tim Cook to remove two dozen trees from the De Anza College campus next to the Flint Center for the Performing Arts to erect an extravagant white tent for the hands-on area.
> So far even losing Jobs didn’t hurt Apple as it looks.
Who knows? Pure speculation? You can also say if Jobs was still around they could have 10x-ed it even further?
Apple car could have been a thing? Apple could have been way ahead and actually competing in AI and data centers? Who knows what else Jobs could have came up with?
Alphabet is only a holding company. I wonder what kind of authority he has over scientists at Google, Waymo, Deepmind, etc. Probably close to none, so a huge demotion in everything but in title and compensation.
To me the most significant part is that he's no longer in charge of DeepMind, the company he created, other than this "Chair" title which sounds meaningless.
When DeepMind allowed Google to buy them, it obviously had some major immediate positives - access to compute and money - but it seems it should have been obvious that the agreement was too good to be true, that they would be allowed to continue independently on their blue sky research mission to create AGI without any external interference or pressure to create product.
It seems that Hassabis and his DM co-founders eventually realized the mistake and tried to take DM private again starting c.2018, but of course this failed.
Now Hassabis has lost control of DeepMind altogether, and it seems to me, as a total outsider, that this is the end of the DeepMind mission to create AGI, at least the Hassabis/Legg definition of AGI as human-level general intelligence, capable of creativity and scientific discovery. Hassabis had always, until very recently, said that he believed it would take a number of additional "Transformer-level" breakthroughs to achieve this type of human-level AGI, while still seeing an LLM as one component of if (which to me seems an admission that the goal has failed - a true human level AGI should be able to learn language, etc, using it's own continual learning mechanisms).
It seems that DeepMind has now fully become the Google Gemini (LLM) division, trying to create a me-too product.
In the early days of DeepMind, before Google, before LLMs, I remember a David Silver slide deck titled "Reward is all you need", referring to RL rewards, which I never agreed with (although Rich Sutton might), but does at least reflect the independent thinking at DM, and of course RL not only gave rise to AlphaGo, but has now become central to the continued improvement of LLMs. However, notably David Silver also left DeepMind earlier this year, to found his own startup focusing on RL-based continual learning, presumably feeling that there was no longer a place for that type of research/pursuit at DeepMind.
Still, LLMs seem to be a destructive enough force on their own that perhaps it should be seen as a positive if research towards more powerful AGI appears to have had a major setback.
As for the "Alphabet Chief Scientist" title, it seems somewhat irrelevant, as least as far as Google's pursuit of true AGI. Hassabis is the face of beneficial AI, having been Knighted and awarded a Nobel Prize for his work, and it would be a horrendous PR move for Alphabet not to at least appear to be treating him with respect, even if in fact this does reflect him being pushed aside.
DeepMind had a generational run as a pure AI research lab. AlphaGo, AlphaZero, protein folding, tensor improvements, weather forecasting, GNoME and so much more. Google leadership saw all this and went “now go generate a multi trillion dollar commercial business and beat OpenAI and Anthropic” and the results were, predictably, failure. Such a shame.
We don't know exactly what Google's leadership was thinking, nor how the AI race will shape up in the coming months and years. Calling it a "failure" seems presumptuous and premature. But supposing your read is correct: what made this outcome predictable, in your view?
How do you avoid being the Xerox to the Microsofts and Apples of AI?
Can you imagine if they'd skipped all of that and put the effort into keeping Search profitable without enshittifying it? Or even just skipped the second part?
If they focused on keeping search results genuinely high-quality, I think the outcome could have been very different. Basically everyone I know slowly began shifting their questions to something like ChatGPT because Google search results were only presenting irrelevant SEO'd garbage with 12 ads shoved into every page. Slapping the lowest quality model at the top as an "AI Overview" and removing the one genuinely useful part of search at that point (widgets like the dictionary) was the final nail in the coffin for me at least.
keeping Search profitable without enshittifying it
I don't think Google had much to do with it. It has way more to do with the internet concentrating into major communities where all the content is private. Even Reddit content is going this way. Most FB and Instagram content are private. All of Discord is unreachable by crawlers.
Prior to this, just about everything on the web was open for crawlers since content wanted to be found.
Google caused this by pushing people towards "brands" rather than individual pages. Those platforms only exposed the data necessary for SEO and Google refused to punish them for it.
> Lastly, after an incredible 27-year run, Jeff Dean is at a moment where he wants to try something new, and we’re excited to support him in that. Jeff and Google Senior Fellow Sanjay Ghemawat are launching an independent public benefit corporation to accelerate discoveries in ML, science, and engineering.
Oof. Good for Jeff and Sanjay (who just joined Twitter), bu that is a big loss for Google. Google stock is down 5%. It might not be much of an exaggeration to say these two are worth ~$200 billion.
Clarification: this comment is saying Sanjay Ghemawat joined Twitter as a user recently (new account @Sanjay_Ghemawat as of July 2026), as opposed to Sanjay working for Twitter the company.
For at least a year now the standard reflexive reply to “Google seems way behind OAI and Anthropic” has been “it’ll be ok, they’ve got Dean and Hassabis.” And now they don’t. What reason is there to be bullish about Google now?
I still think Google is the only one who has a shot of coming out of this on top. Open AI and Anthropic's whole existence is predicated on some sort of moat, which I don't really see them having long term. They've got a bit of a headstart, but that's it.
Conversely, AI is just a means to an end for Google - they don't need for their model to be the one to succeed. But, in contrast to the other major company in their position Apple, they do have a model, so they're not totally beholden to another for AI (like Apple is using Gemini!).
But beyond that, for training the model they have YouTube, and of course they have their crawler and index, and the billions of users.
I think HN skews coding agent focused, but that's not really a market for Google. I expect we will have coding specific models in the future, but Google wants a more general intelligence, to handle search queries, be able to connect email to chat to calendar and tasks, and so on. I don't find Gemini that much worse than the other big models for non-coding things.
The leaked internal Google memo "We have no moat" [1] was really prescient. It foresaw all of these proprietary models losing out to open-weight models.
Entirely possible that none of the big incumbents ends up winning, economically.
I would like a reassessment of that. This was 2023. I imagine the only ones filling the moat is Chinese labs, but there are other significant obststacles for even Chinese models. Even if they would become as good as their American counterparts, the distribution and inference is mostly in the ballpark of Anthropic and OpenAI.
But I don’t know anything, so I would love to hear other opinions.
If anything, it's becoming more true. In the last month you already have Kimi K3 reaching competitive frontier performance (in practice, a bit behind Fable/Sol but close enough). Sure, it's a chinese model but the reality is it puts dramatic pricing pressure on the main players. And these models are now getting good enough to actually help accelerate further AI model research which would suggest the gap might close even more in the near future.
I hardly use Google search anymore in favor of ChatGPT. It's surely an existential threat. Google has a lot of layers of defense in Android, Chrome, Gmail/Accounts, and a stranglehold on internet ads. But at the center of that is search and search is being replaced by LLMs.
We are seeing an exponential increase in the share of site traffic to our B2B company originating from ChatGPT and other models. Google Search is cooked.
IMO, this is largely because google made their search next to worthless. They very apparently prioritized making money over returning good relevant results. The decline in quality is so pronounced that I've heard more than a few people talk about using other search engines besides google.
I don't think it's extreme to say that Bing is a better search engine than Google at this point and Bing isn't great.
> I hardly use Google search anymore in favor of ChatGPT.
Same here, and I'm generally very skeptical when it comes to AI. It's just that ChatGPT can provide better/more accurate answers, Google seems to have lost the plot. I still use google search when it comes to appending "wiki/wikipedia" to it, i.e. when I use google search just as a redirect to a wikipedia page.
I read Google's business incentive as "in it to not lose". AI is a new consumer endpoint and Google currently captures a lot of the endpoints.
> I think HN skews coding agent focused
100% this. Coding agents are an interesting test ground, but the people who care about them are a fairly small bubble.
I think they carry some of the overall AI weight because of the "what if you can vibe code your entire business" moonshot, but that's still orders of magnitude away and who knows if today's coding agents will actually be a stepping stone to that. If we ever get there it will likely be with entirely new domain languages.
What about revenue if people aren't using the search engine directly anymore, also consuming content through chats (not made by them), not on the web where ads are displayed?
On the flipside, Apple has realized AI models are quickly becoming a commodity. By not blowing hundreds of billions in speculative capex, they get to focus on the bits that likely have higher ROI: the touchpoints with AI and how it's used.
As much as everyone wants to pay accomplished celebrities, all of these companies have young nameless geniuses that are about to make one for themselves. A guard passing torch can be opportunity.
Now, whether Google is the right environment to nurture, that’s its own quandary.
It just depends on what are the most upvoted comments in HN. If they are bearish, you can be sure that you should be bullish in your investments. Meta was supposed to be broken as a business already, OpenAI and Anthropic would be failures as well.
Maybe they'll replace Hassabis with someone more focussed on beating OAI and Anthropic at chatbots? He always seemed a bit more into science, protein folding, new drugs and the like.
Couple of things come to mind. One is that Web sites tend to actively fight AI companies' crawlers while actively courting Google's crawler.
Another is that they hold the key Transformer architecture patent. If it is still relevant (which I'm not personally clueful about) and if they start enforcing it, then we may see a reprise of the situation where Microsoft made money for years every time an Android phone was sold. Regardless of that particular patent, it's probably safe to say they'll be better-positioned than anyone else if AI companies start lobbing patent nukes at each other.
A third factor that shouldn't be discounted is that Google has access to warehouses of training data that other companies don't. Google Books alone is an Alexandria-scale archive that the courts forced them to keep to themselves. Those restrictive copyright decisions may turn out to be a blessing in disguise for Google because no one else will have been able to scrape the data.
Google's weakness (well one of) is its total lack of cohesion. If the Google Books team could, they would sell access to that data in a heartbeat to boost their metrics.
Whoa. So is Jeff effectively leaving Google to work on this new venture full time? Or is the venture a side project? It sounds like the former. I’m sure he’ll still have internal access as an advisor of sorts. But this feels like a seismic change. Much larger than I initially realized?
Today I saw one of the most talented engineers I know and worked with leave DeepMind and now I probably know part of the "why". Dark clouds hovering over Google's AI game.
It's not really clear what their gameplan is. From the outside, it looks like they're asleep at the wheel. Qwen/Deepseek/Kimi are crushing them from the cheap-and-open side, and they're not remotely competitive with Mythos/Sol or even plain-vanilla Opus on the "premium" tier.
Gemini does actually have its uses, but they're very very marginal and niche.
From day one everybody was saying that Google would eventually capture the AI market, but it looks more remote than ever. Maybe Hassabis' personal inclination towards AI-for-science, and physics/chemistry in particular -- as opposed to consumer AI and coding AI -- has hurt them commercially.
Is Gemini really doomed? I'm still bullish on Google:
1) they have more free cash flow and capital than God due to the ads business
2) they have data - intent from web searches, youtube videos, google books and music
3) they have dedicated inference hardware
for all these reasons, is being 6 months behind the frontier actually a structural, long term disadvantage? some day the pace of improvement will slow, and google will vacuum up the market. they'll be able to compete with open-weight models just on pure cost advantage from their vertical integration
That also own one of the 2 major mobile operating systems, with Gemini tightly integrated and all of the data they can gather from that. Why do you think OpenAI wants to do hardware? Owning delivery is going to be important, and right now, Google and Apple own the delivery mechanisms (to consumers).
They may not capture enterprise use, but I don't think they have to. That's only one piece of the market. AI that's useful to consumers will still get delivered via a smartphone, and Google is in a great place to capture that.
I also don't think LLMs have to be a "winner takes all" situation. Value isn't going to come from having direct access to a chatbot or selling API inference, value is going to be in the form of a specific product (for most, devs aside here). Something a consumer, or a non-tech business can buy off the shelf and plug and play. A "ready made" customer service agent system, a "ready made" BI platform using AI, etc.
For consumers, that's probably going to look like whatever is bundled and tightly integrated into their mobile OS of choice.
My prediction is that the EU will eventually bring in legislation that will force platform providers to provide pluggable APIs so that you can use whatever LLM you want.
"Doomed" no, but it's pretty clear that they just had a bad cycle and are struggling to keep up with the frontier.
Whether this happened because they bet on "world models -> better reasoning" and that bet didn't pay off, or failed a frontier run for technical reasons like OpenAI did with 4.5, or something else went down? We don't know.
Will they bleed talent, fall further behind until they give up, or clean the organizational and infrastructural cobwebs and get back in the saddle? We don't know.
So long as search revenue isn't correlated to AI revenue or threatened by AI revenue, Google at any point can bail on AI spending and suddenly the free cash flow machine is back on. Yeah they have a far larger debt load they now have to service but cash flow from search revenue is so strong it wouldn't be much of a blip.
> is being 6 months behind the frontier actually a structural, long term disadvantage?
Yeah, this is one of the things I find so weird on the discourse. If you get there negligeably later, but without astonishing spend and waste, you might even be better off in the long term.
I have thought that for a while and assumed that was apples approach to AI. Wait until everyone burns through investor cash, invents the better tech, and can monetize. Then copy that business model and polish it, or buy out the competition and polish. That’s usually apples move and it makes sense for google to do something similar.
I don’t think it matters that much for Google. They need to not fall hopelessly behind, but I don’t think there’s a strong economic reason for Google to burn the kind of capex that the frontier labs are burning. Strategically, I think they’re probably doing better than OpenAI and Anthropic. The Gemini models are open, and they are what researchers are working with (see neuronpedia as an example). Over time, this will give them a strategic advantage for the same reasons that open source wins over proprietary. Meanwhile, OpenAI and Anthropic have massive capex that needs to be returned to investors while their margins are being undercut by Kimi/Deepseek/Qwen. Google can wait around for the coming frontier lab profitability crisis and cruise right on by with their Apple contract and owned data centers to pick up the pieces and exceed the existing frontier labs.
A deal with a company Google has a share of, announced a week before their IPO, with very non-committal terms and ramp period protections delivered in one large block on short term notice priced likely at the high end of what Google charges for A4X instances anyway.
I don’t think this reflects desperation as much as strategy.
I dunno if it's a sound strategy that involves repeatedly telling investors [1] and employees [2] over multiple quarters that you are desperate for compute, including leaving a triple-digit billion backlog on the table [3], and then spending so much on CapEx that you have your first negative cash flow quarter ever and taking the inevitable hit to the stock [4], while turning away a large paying customer (who also happen to be a competitor) [5] ;-)
The RPO can be anywhere from 3 - 6 years, sure, but even on an annual basis that’s like a hundred billion now. It was already in the double-digit billions since before AI took off and has only been spiking since then, which tells us 1) it’s been huge for 3+ years, and 2) it’s still growing faster than they can collect it. This matches what all the other hyperscalers are doing.
My point is that an ulterior motive is not necessary to assume when all their actions and statements point to them being severely crunched for compute.
I mean sure, if they had a choice between say, CoreWeave and SpaceX, they’d choose the latter for the nice bump to SpaceX’s financials and their stake… but not just for that, not when it contributes to their cash flow turning negative and their own stock taking a hit.
It is very suspicious when they’re paying for GB300 more than they turn around and charge those out as a4x instances, it was announced a week before IPO and they have a 90 day exit option.
> The RPO can be anywhere from 3 - 6 years, sure, but even on an annual basis that’s like a hundred billion now.
OpenAI and Anthropic deals are mostly 5 year and Anthropic’s starts in 2027, accordingly these RPOs are sized for projected compute needs and run rate in 2027 not today.
The only way either lab could pay 1 year of RPOs today (~80B for anthropic and ~150B for OpenAI) is with a lot more debt or circular financing, the former of which is difficult in this market.
Everyone spending crazy money on capex right now says they have a crushing backlog and need more compute to protect their share price. I highly doubt the demand exists today at current prices if the big 3 hyperscalers magically had an extra 2-3GW of compute. It’s not like Anthropic and OpenAI are turning away customers offering to pay API pricing..
What do you mean? Google bought SpaceX shares when it was a tiny startup. They are probably more than 100X on their initial investment. Even with a 99% drop in SpaceX stock, Google would still be positive.
Edit. Google invested 900 million in 2015 for roughly 5% of the company which comes out to 71.5 Billion dollars at 1.45 Trillion dollar current valuation. That's an 80x increase.
I think Larry Page individually might also have a very large stake as well.
It doesn’t matter what price they bought SpaceX stock for. They could have gotten it for free. Any gain prior to Google’s investment could be realized prior to Google giving SpaceX even more money. What matters is the return on investment for the $12B/year into SpaceX.
In order get that back by “pumping” SpaceX stock, SpaceX market cap would have had to increase $200B based on that investment, and then Google would need to sell the stock.
I'm confused. I think you're confusing Gemini and Gemma. Gemini is Google's frontier offering which is closed-weight, API-only, like most frontier models. Gemma is Google's open-weight offering focused on deployment on consumer and edge hardware.
"The company raised its full-year 2026 capex forecast to between $195 billion and $205 billion, with further significant increases planned for 2027." - Alphabet.
I think ~$200B is just for AI infrastructure capex. Fun fact: that's nearly what the 3rd largest military in the world (Russia) is spending on a land war in Europe.
- They're making a lot of money selling Tensor to Anthropic. If Nvidia's $4T market cap is justifiable, Google's position as one of the other top AI chip seller is worth a lot.
- In a world where open source Chinese models decimate Frontier models ability to charge a high price, it's the operators of efficient inference data centers that will win. Like Google
- Google is probably the biggest provider of "free" AI because it's on Google.com. That forces them to focus on cost. And in a commodity market, low-cost providers are the ones that make the money.
> everybody was saying that Google would eventually capture the AI market
my take on this is eventually the money is going to run out and there's going to be acquisitions and consolidation. I think that's when Google will come out on top.
Google doesn't need to compete. 'everyone' is locked into them via the Gapps (mostly Gmail and Maps) and Android ecosystems. Same with Apple, and Microsoft on the B2B side. It's only Anthropic, OpenAI and everyone else that _need_ to compete because switching models is painless. And they have no other revenue streams.
10 years from now, its gonna be Google, Apple and Microsoft left standing in the AI game. Well, until the US wakes up and starts attempting to break the oligopoly like the EU has recently started to do.
I largely agree but I don't know if it's quite so clear cut. From the pricing angle, all competitors except Google, including Chinese models, have incentive to gain market share at all costs, and may be serving tokens at or below cost. I am not sure though, Google could certainly decrease prices if they wanted to.
For agentic work, but especially for web search, 3.6 Flash has an important leg up, its fast speed, that no other model comes close to matching. I guess nobody pays attention to it because it's not the one big flashy number that you compare to other models.
The default model they are using for Web search is getting capable and is very visible. And now it invites people to keep asking questions.
They are quietly trying to become the chatbot that everyone uses to look things up. That strikes me as an intelligent move - not everything has to be done by an expensive frontier model.
They are not "asleep at the wheel"; it's just that the people in charge (the "MBA types") have no clue what to do!
There's an old saying, if you judge a fish's smarts by how well it can ride a bicycle, it will always seem dumb.
The people who have risen to the top of at Google are built for a different environment than what's needed right now. They are good at playing their political games, sabotaging each other, etc.; i.e. all of the petty games that managers play in big companies. But the AI era demands a different skill set: how to bring together incredibly smart people and forge them into a battle group that will achieve victory in the ongoing battle for AGI! It's as if you have built an army of tanks, but the next battle is being fought on the high seas.
Are they measurably attracting more talent here? It seems like they're losing some of it right now, so is there public info about numbers of researchers they have or etc
Agree that I would (and do) still place my bet on them for the long term. Maybe the outcome will be a couple of good startups seeded and DeepMind _really_ focusing on LLMs now.
social media is a completely different market though - since there are massive returns to scale, it's incredibly hard for a new entrant to break in.
model training and inference is the opposite. people switch LLMs like people change clothes in the morning. there are popular services (openrouter) that make moving as easy as changing a model string.
Which poses the question: Why does Google even need to catch up? At least currently, the name of the game is integration. The actual model is a commodity.
Unfortunately the new cow is cannibalizing the old cow, so Google is in a bit of a bind here. (So far I cannot imagine them monetizing AI overviews enough to compensate for the sharp loss of ads on SERPs.)
To its credit Google seems willing to disrupt itself before its competitors can.
I said that too at the start of the year, but that's a looong time in "AI years". I feel like by now they should have announced a Fable-killer model. They may still do it but looking back I am becoming less convinced now than I was 6 months ago.
They also have the money, the hardware and the talent to be the best cloud infrastructure provider, yet they're still far behind AWS (for good reason, as anyone who's dealt with their customer service will understand).
I think Google realized that it's more profitable to sell compute to AI labs than to make AI. They are not an ideologically driven company like Anthropic. They very much turned into a conventional company that just wants to protect the bottom line. This was evident even back when they had LAMDA and refused to release it.
Watch it recover in the next 3 days. This is a buying oppo if I've seen one. The real losers are the companies that have yet to gain a userbase as large as G.
> I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.
I think their move makes a lot of sense. Frontier models can probably not advance much further with the datasets we have currently available. Most of the text that goes in is online chatter, images and some scientific texts. That's great for chatbots, knowledge retrieval and programming. But with that database genuine discovery is hard to do. I think for the next step in intelligence the models need to have access to much more data: Data from physics, chemistry, biology experiments - and so on. And they need the data in much higher fidelity than you can currently access. If we just feed AI with all the knowledge we've acquired it's much harder for it to become smarter than us - it basically needs it's own eyes, ears, nose and so on.
There’s a difference between things getting incrementally better and a step function.
LLMs will obviously keep getting incrementally better, but in order to get the kind on jump that LLMs themselves were, the sentiment is that we need something more.
That’s because a significant portion of the work and spending going on at frontier labs is generating new, more curated data, in select domains like software engineering and now biology[1].
I think that was his argument as well. It looks like there is no more data to find, but then it becomes important to find more data - lo and behold we can make more.
Sounds like the oil scare from 90' - we thought we were gonna run out of oil. But as oil gets more expensive it pays to dig further down to find the stuff that didn't make sense to dig up before.
Yes. The web is enormous, but just think of home much information within a company doesn't even make it into the internal knowledgebase, let alone anywhere public.
I think there are multiple separate events here, timed to coincide in order to minimise disruption. It's hard to be sure whether they share overlapping causes, but I'd guess that there's at least an element of that.
Yes. In public relations you never want to trickle out bad/concerning the news, you want to "rip off the band-aid and preferably follow it up with some positive/distracting news like 3.5 Pro model release or price drops.
Event 2 - Demis stepping down. DeepMind doesn't need a "chair" and "Alphabet's chief scientist" is a bullshit title that Jeff invented for himself when he moved from Google Research to GDM, to make it look like we wasn't abandoning the former for the latter.
The third change is basically ratifying the status quo, since Koray has been de factor running GDM for a while now (and directly reporting to Sundar in addition to reporting to Demis, who in turn also reported to Sundar - quite some triangle there).
Gemini had come under Hassabis, being co-led by Shazeer and Vinyals, who are now both gone. Gemini will now come under Koray Kavukcuoglu, reporting direct to Pichai as SVP of DeepMind.
Fun fact: the Jeff Dean Facts was started by Kenton Varda (kentonv around here [1]), as a joke inside Google. He has expressed regret that it eclipsed Sanjay Ghemawat, with whom Jeff Dean basically did pair-programming with.
Basically, he had a server that did code indexing running on his workstation. If he didn't login and run a command to refresh his prod credentials every day, the server would lose its ability to talk to prod (as it should) and fail. Much of the company depended on that service. Eventually it was moved to prod.
"Once, in early 2002, when the index servers went down, Jeff Dean answered user queries manually for two hours. Evals showed a quality improvement of 5 points."
It seems reminiscent of the Green Project/FirstPerson episode that led to Java https://landley.net/history/mirror/java/javaorigin.html , and I'm sure many another attempt to keep some unhappy star players in a company's orbit using a spinoff.
> I am sure VCs are fighting to invest and they will get 4B investment immediately with this team
This is probably the only time some prominent VCs would be active during August, only for Jeff Dean & Co's company.
I can see them now frantically negotiating, responding and firing off emails right now during their holidays to get an allocation in Jeff's new company.
Would love to hear the pitch: We never had the lead in AI and now definitely lost it, despite the gazillions dollars and engineering resources of Google, but now will be different?
The "lead" in AI, even with google's TPU advantage, may not make economic sense for a company. Can you make back the training investment on a competitive frontier model, before everyone switches to a newer frontier model within a year?
I think there's one weirdly simple reason DeepMind isn't doing as well as OpenAI and Anthropic. I may be wrong on this.
OpenAI and Anthropic went from tiny startups to huge companies. As a consequence the stock options/RSU's offered to the employees paid off a far higher percentage ROI than any stock options a DeepMind (and thus Google) employee would get (since Google is already huge). This disincentivizes people who truly believe in the economically transformative power of AI to work at Google since their benefits will be capped by Google being large + having public company obligations.
OpenAI and Anthropic have the freedom to do absolutely insane things like negligently hack other companies. It would be stock price suicide if anything even remotely happened with Google.
There's no doubt in my mind that they set up the conditions for their models to escape the sandboxes. "haha oops our incredibly powerful models escaped we need 1 trillion more dollars and really this is yet another reason why no one else should be allowed to build this technology"
> As was standard in our cyber testing, we had intentionally permitted internet access, and model- provider cyber classifiers were deliberately disabled - conditions that do not reflect how frontier models are
made available to the public.
"No, no, no" a person on Reddit, Hacker News, YouTube screams for the billionth time. "It's all marketing" as the terminator bots kick in the door and slaughter their families.
Crash society to the point where we can no longer make complicated chips. Of course this really isn't a great solution as it involves billions dying too. Not that someone won't try...
We've signed up on the "If you build it, everyone will die" express. No brakes, no stops, full speed ahead.
If Google put out a blog post about Gemini 4 having escaped its sandbox via 0-day exploit to then hack other companies, I unironically believe this would result in a boost to their stock price.
They could really use some encouraging news about the competitiveness of their AI lab.
I wonder if we'll ever know the story of Gemini 3.5 Pro. Is it possible Google saw its potential for hacking, tried to nerf it, and ended up ruining the training run?
Yes. These corporations now run on fascist logic. (Hold up, stay with me here.) The goal isn't to right, or ethical, or principled. It's to (appear to) be strong and capable, and loud about it. You end up with horrible things happening that get cheered because they're only horrible along dimensions that don't matter anymore to the people calling the shots. Same thing for amazing developments that get crickets.
Note that I'm not calling Google "literally Hitler". I'm just saying that traditional assumptions about what's virtuous and what isn't don't apply anymore. For now, at least. If anyone wants to change that, they're going to have to be purposeful about it.
Google's AI was telling people to put elmer's glue on pizza to help the cheese stick [1], Gemini was turning the Founding Fathers black [2], and more. Oh and the pizza recipe was based on a joke from a reddit user named "fucksmith" - part of data that Google apparently paid some $60 million to Reddit to access. The one and only effect of this was lots of amusing posts and articles. Their stock price went up during the whole ordeal.
They getting a higher ROI renting their TPUs to Anthropic et al instead of performing training and serving their own models. Google cloud has insane backlog, and has rapidly expanded to satisfy it. While those DCs get built, they’re cannibalizing their own products for it.
This makes sense because (1) they are investors in Anthropic, so they still win and (2) they can always catch up on model training later when the profit opportunity shifts, or abandon it if there is no way to recapture that value.
I think the reason Google hasn't prioritized larger models that are more intelligent than everyone else's, even though they probably could, is because they have 4 billion active users already. For example, they send AI Overviews for a large portion of Google searches now.
So I think they have to prioritize scaling for their models to a higher degree than other groups. Being within say 5% or so in most cases is probably adequate and matters more overall for their user base than being the absolute best coder. So they may be setting compute constraints for training or inference that are firmer than other teams.
When you have a lot of free users the business demands that you serve them with the best cheap model you can build
And time spent building that may provide dividends (eg OpenAI has very good RL and reasoning) but it might take resources away from the larger model training
(I have no inside knowledge, so please consider this to all be speculation)
I don’t think so. If Kimi and Deepseek can launch models better than Gemini with much much lesser resources then it is increasingly looking like an organization issue at Google
No I think you misunderstood GP’s comment. The idea (which I personally don’t agree with) was that Google didn’t have to have the best models; it just needed to have the best compute infrastructure, i.e. having TPUs and the software stack to use TPUs. It was a better use of money to develop compute infrastructure than to develop better models. Perhaps Gemini itself was resource-starved because Google liked to rent out TPUs to Anthropic instead. (Second-hand information: I heard that Mythos/Fable were trained on Google TPUs.)
I agree that other organizations can compete with few resources, but my hypothesis is that Gemini training specifically is being given nearly 0 resources, despite Google obviously having lots of resources. The hypothesis is based on an assumption that Google profits more by selling ALL their compute to others training models instead of using it themselves for training.
They already have good models, so “better” isn’t as profitable.
Google does not have to compete at the frontier, they already own a lot of Anthropic. It's not an "issue" for them because it's not one of their goals.
I don't think Google is that freaking short-sighted. Even if you're not trying to be frontier, the value of having domain knowledge via experience for AI is worth saving internal compute alone. By all accounts Google believes in ai as much as everyone else.
If AI turns out 1/100 as important as they seem to think, it would be insane to intentionally be slack on it.
I disagree. Google is one of the few parties that can monetize AI because Google has a massive moat in the form of their products: gmail, chrome, photos, search, etc... and Google already has the custom-built chips and datacenters.
Google spending money on competing head to head with your LLMs is a waste of money for Google. If anthropic wins, Google copies their approach, buys anthropic for cheap or both. All the investors throwing money into OpenAI, Anthropic, are just accidentally subsidizing Google's product development. Google shouldn't spend its AI research capital in an arms race with Anthropic but instead should invest in AI approaches that no one else is investigating at scale. That way Google can hedge against LLMs hitting a wall.
There is a real but small danger to Google that Anthropic replaces Google as a search engine, but that is an uphill fight for Anthropic. Google has massive brand recognition, network effects with gmail and chrome, Anthropic can't just copy what works from Google. On the other hand Google can copy what works for Anthropic. Google would have to play poorly to lose that fight.
Probably the worse case for Google is that software becomes so cheap and easy to create and maintain that all of Google's product offerings become commoditized. Even in that world Google has a lock on infrastructure. Perhaps ASI software creation completely removes that as well? If so we are living in a post-singularity world and probably the stockmarket doesn't exist anymore either.
In retrospect this is very similar to the Apple strategy: don’t invest heavily in doing something that isn’t already a core competency, position for novel uses of the tech but don’t build it per se. Apple probably would have benefitted in the last 4 quarters from hyping a custom model stack but it doesn’t seem that this strategy paid off to even close to 20% ROI while Apple got to hold their cash and organizational focus on their main thing
But paying so little attention to it that it no longer is a core competency? As in, the people most representative of that competency leaving in droves? I doubt that was the plan.
And Gemini seems about as good as a search engine as any of the other LLMs, if you use them casually. And Google has the brand-name recognition.
Right now AI companies are competing to make LLMs better at graduate-school level tasks. They all can already competently tell you what the weather is going to be like tomorrow or when the first Led Zeppelin album was released.
Gemini seems like a better search engine than the alternatives because it has access to Google's massive crawler feeds. I think Google could charge 30 dollars month for Gemini + all the data Google has locked up (Scholar, Books, crawled webpages, Google Groups, Usenet archives + search of your personalized datasets such as calendars, photos, emails, docs, slides, etc...). I'd pay that easily.
And that might be why Google loses, they are competing with companies that are willing to operate outside the law (or at least pay Trump to have the DoJ look the other way).
Anthropic has no moat, their models just get distilled and resold. They can maybe get good margins when LLM improvement rate is very high but as it slows down they are looking at commodity margins. Google has the products that makes that commodity more than just cost of electricity + 0.5%.
My case is based on the assumption that Anthropic will not hit RSI or if it does RSI rapidly hits a wall. Faster your growth curve, the faster you eat all the low hanging fruit and s-curve. I could be wrong here, maybe RSI will cause a hard takeoff singularity by 2030 and just keep going, but if that happens the world fundamentally changes.
They wouldn't be the only party, but they are in my opinion the most likely party.
What the people that can monetize it? Microsoft, X, IBM, Salesforce, Discord, Google
Microsoft would likely buy OpenAI due to their investment and integration with OpenAI. Making Anthropic not that valuable.
Google would likely buy Anthropic due to their investment and ties to Anthropic.
X, Salesforce maybe. Not sure Discord has the money. No idea if IBM capable of growing.
All this changes if Anthropic figures out a moat/sticky product that doesn't just depend on having the best LLM. If they pull off https://claude.com/solutions/healthcare then all bets are off.
Not all change is good, as proven by Zuckerberg's response after the lackluster Llama 4 release. The radical restructuring appears to have made things worse.
How? Muse Spark 1.1 is a huge step up from Llama 4 & competitive with xAI's Grok 4.5. In another 3 to 4 releases, MSL might very well be challenging Ant & OAI. Moonshot, despite their comparatively limited resources, has already demonstrated that the Big 2 aren't invincible.
Google was always at the frontier of real research, but has been abysmal at shipping good products (at least since Sundar). The core company is run for margins and interest rates by the business people nowadays.
So they are fixing this by letting go of the people who were best at the research side, and therefore will have no problem converting "nothing" into products anymore!
People keep saying the same thing about Google lagging behind and always end up looking rather silly. Google was going to lose search to OpenAI. Then people complained there was too much AI in Google search. Now it's pretty good and par for the course.
Yes it's a huge company.
So they are slower. Then they will surface it across their massive product base and keep generating cash. While having a hand in Anthropic and others via investment anyway.
Google doesn't need to offer you the bleeding edge at startup pace. They're playing a different game. When the bubble pops they will be well positioned really no matter the outcome to continue to capitalize as their competitors implode or get absorbed.
The idea a delay is a "complete and unmitigated disaster" is just laughable. People have been saying this about Google since ChatGPT first invaded the public consciousness. Google will continue to do well, the histrionics of people like you aside.
That's one thing everyone needs to remember. AI is here to stay, but the bubble IS going to pop. This level of spending is unsustainable. Soon the market will readjust and the amount of money we spend on AI will return to sane levels.
Google has multiple cash firehouses, the small AI companies do not.
Just today my Pixel failed to do the right thing on "Set an alarm in 15 minutes" thanks to Gemini. This has worked reliably since Google Assistant was introduced.
Huge companies tend to make money from network effects, rent seeking, and lock in. They tend to be horrifically bad when it comes to innovation, especially when it may disrupt exiting departments in the company. Those departments will fight for their life and generally muck things up.
Very few companies have leadership that can prevent this infighting and force teams on directed goals.
> people who truly believe in the economically transformative power of AI
People who truly care about becoming rich*
DeepMind made enormous transformative discoveries, for instance in the world of protein folding. But that will just save human lives, not let CEOs fire their people to grab a larger piece of cake for themselves.
I think something that doesn't get talked about a lot is how bad most large tech companies are at creating new products, in general. Like, if you look at most big tech companies, they have their core offering that got them to be really large and rich, and a few other products that are somewhat successful, and then a really long tail of markets they try to enter and failed at, or projects that were modestly successful but got killed because they weren't game changers (RIP Google Reader). Most of the time when a large company does something new that succeeds, it's via an acquisition of a smaller company (ie, Google with Android or Meta with Instagram and Whatsapp)
It's funny, because I think the company that's going to be best positioned coming out of this bubble is in fact google, because they have the expertise and the capital. But I honestly can't tell you right now what their AI product even is -- I've seen so many things go into the graveyard a few months after its launched that I'm utterly confused what their offering even is at this point.
From what I've heard, DeepMind's pay structure is not tied to Google compensation bands. Which means that top performers and key hires can be and are compensated extremely generously if the need calls for it. Another Alphabet subsidiary, Waymo, was well-known for doing this back in the day.
But yes, even large RSU grants cannot replicate the upside of joining a company that increases in valuation 1000X, but to be fair, that is not a position the vast majority of OpenAI or Anthropic employees find themselves in either.
That was true early but OpenAI/Anthropic have done a ton of hiring over the past couple years at already-huge valuations, as much on the strength of big current base salary + equity, not just future increase speculation.
Google was offering researchers the Bell Labs/Xerox PARC model of comfortable budgets and pay but capped upside. And just like with Bell Labs and Xerox PARC the inventions at Google got productized elsewhere by people chasing the uncapped upside. Google would have been happy to keep LLMs in the research lab forever and never productize any of it. ChatGPT forced their hand.
DeepMind is just a part of Google which does research stuff, and Google is a massive company that does loads of things but mostly sells ads.
OpenAI and Anthropic are in a completely different game, their primary business model is making a product out of advancements that come out of DeepMind and co.
It's not a ML talent problem. You don't need to be a genius deep learning researcher to think "Hey, maybe if we massively throttle and degrade the quality of our model while still charging the same price, that might drive people away" (as happened with Gemini 2.5 Pro, the one model where Google really was SOTA). Google's likely been providing insufficient training compute to DeepMind the same way they've been nickel-and-diming their customers, funneling it all to Search instead because that's where the money comes from.
Is that what's happening? Search has sucked for a while, and I'd just assumed it was partly because they'd been pushing compute to AI-related activities.
This makes a lot of sense to me. Probably doesn't explain everything but I could see it explaining a lot. Could also explain openai and anthropic delaying IPO (among other factors)
One theory I've been entertaining is that whenever GPT-3.5 came out a lot of people were talking about the "bitter lesson" and how scale was all we really needed to get to AGI. No need for any fancy tricks, just release a larger model trained on more data, by the time we released a hypothetical "GPT-5 sized" model we'd have AGI.
Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time. It makes sense when you realize that one of Google's key strengths, besides talent, is that they have an incredible amount of data they can use for training due to being both the world's leading search engine as well as having all that video data from YouTube. Scaling the training data makes more sense to them than it does to Anthropic and OpenAI, who are both relatively data-disadvantaged.
You can kind of see this when you look at the Gemini 3 scorecard when it came out (https://blog.google/products-and-platforms/products/gemini/g...) and notice that while it wasn't as good as Claude And GPT at coding, it scored higher on a bunch of other non-coding benchmarks, and I think the reason why is simply because of Google's data advantage.
If true, I feel even more vindicated for believing that the "scale is all we need" narrative was bullshit.
Other than attention optimizations and other minor changes, the top Chinese models (which are way better than gemini) have basically the same architecture as GPT2. Of course RL is key for agentic workloads, but I'd say it's correct that progress has been mostly scaling models,adding more data and cleaning it better.
Google has a ton of smart researchers trying all sorts of stuff. They didn’t really go all in on any one thing. They hedge their bets.
But they suck at harnesses, developer tooling etc. their internal environment is very complex and an intern on a MacBook with codex can probably move faster than a seasoned deepmind engineer
> Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time.
We'll see if anyone gets to cash in on those. All you need is one down round and that gets wiped out. Or if the IPO gets delayed and disappoints then the stock can drop well before the lockouts expire. OpenAI and Anthropic are essentially offering Monopoly money in the hopes that one day you can exchange it for real money.
I think it's just more about market incentive. At it's core, LLMs are bad for google's previous business model, which was to send you to as many sites 'good enough' for what you were looking for and plant Ad land mines along the way, in the search results and in the websites themselves.
The new paradigm is you ask the llm a question, get the answer and cutout the middle man. (yes the answer may or may not be as good as the old google result, but for the sake of the argument lets say it is), Google was in danger of simply getting their arm cut off. so they focused on scaling so they could add LLMs to the search, which they largely have. You can't offer an opus like model on something as big as search (and which is offered for 'free'), so they focused on that model, and the infrastructure to run it, because they cannot afford to lose search.
Meanwhile, they know the power of frontier models, they are working to have the infrastructure to be a huge player in them and I'm sure they will have a frontier capable model, eventually. They are playing a longer game, because they can, and I think it's going to work out very well for them.
Not sure that's the right way to look at it, given that Google's huge head start in capital and talent did not prevent AI competition at all. It's a demonstration of reasonable, non-problematic dynamics between smaller and larger companies. (Of course, there's an implicit risk here, the folks at Cruise probably worked harder and more passionately than Waymo staff too.)
you are right. I guess what I was thinking was that google's bigness was essentially a bad capital allocation strategy, since they were lazy and not motivated by absolute return, but some combination of acceptable risk, politics, personal preferences, etc in a large management team that has seemed...disconnected for quite some time.
They have a structural advantage in cash flow and stability of funding, but stability is also a handicap when disruption is the objective.
It's a notoriously double edged sword. When you create huge absolute return incentives, you get things like the Airtable acquisition, where everyone's sad that you built a $1.2B company because some investor at some point mistakenly thought it was an $11B company.
Yeah, that makes sense. I suppose the flip side of all of this is that neither OpenAI nor Anthropic have made a lot of money relative to their spending. Google is probably just trying to be more prudent. But, AI so far is a money spending contest, more than a money making one.
Just some random tidbit: When I was in high school I bought every computer gaming magazine I could get my hands on and could afford with my allowance.
So when Deepmind first made headlines I recalled an article in the UK magazine, Edge, which had an article about a game called Republic being developed by a team of former Bullfrog employees lead by one Demis Hassabis.
Out of interest, I just checked Internet Archive, and lo and behold I found it [0]
I see Wikipedia [1] also mentions that he was the lead programmer on Them park and worked with Peter Molyneaux at Lionhead during the development of Black & White.
It doesn't add much to the story under discussion, but it makes me think at the time I wanted to be a game programmer but my parents encouraged me to go study engineering instead.
It is a bit of a tagent but still--Theme Park was dope. I was too young to grasp the compete with other businesses (stocks and shares?) level but still had a lot of fun building parks.
I always assumed Demis' ultimate secret motivation behind his work was the desire to deliver a version of Black and White that actually lived up to the hype
Wow. Jeff and Sanjay both departing. Truly end of a golden era.
There is an entire cohort of work-optional very senior engineers for whom one of the last reasons for hanging on was "at least Jeff and Sanjay are around".
This just shows that the "great minds" of AI in these companies are not able to come up with anything that radically differentiates their AI than the others. Wether it is gemini or openAI or grok - meh they all are similar and at this point one can interchange one with the other , excep that Google sits right at the edge with almost every browser search query hitting the google search engine AKA Gemini on the backend and one can switch easily into chat mode right from the search page. I dare say that the google CEO has done his job and set up the cash printing machine but from the AI masterminds in Google, not a groundbreaking progress has been made.
Once upon a time it would've been unthinkable for Google search to become awful and nearly worthless due, in large part, by Google's discretionary choices.
Related - interns these days don't seem to have any particular preference for using Google for finding information...
I think it's a bit inflated to label Gmail a "great tech for consumers". Maps certainly, YouTube mostly, Chrome & Android for loyalists, but Gmail hasn't been "great tech" for well over a decade.
Additionally these are all old products - even the company's more recent products such as Gemini feel stale and on unsteady ground.
Their entire business is predicated on search dominance, which is now on much shakier footing. Agreed that those comments were ridiculous pre-ChatGPT, but now they have a genuine challenger.
This take only makes sense if you don’t know what people use Google search for. If you’re looking for your local Ford dealership, or to book a Carnival cruise, or need a disability lawyer, or to refinance your credit card debt, you’re searching for an ad. You’re not using Chat-GPT for that. You’re asking Chat-GPT about a research project. It’s stealing away all the hard to monetize searches and leaving the incredibly lucrative searches.
Your narrative depressed Google’s stock price for years, and bears finally gave up since they kept coming out with huge earnings beats.
I actually use Claude for finding stuff and recommendations. like "Find me a decent shoe with blah blah conditions" this query now goes to LLM Chat instead of google search. I use google search to find the url of something that i know exists like find the website of a company or finding address of a store.
The issue is Claude is losing money serving you that, and while Google makes money serving you the same result. On top of that, you are paying Claude to get that same result.
Unless Claude adds ads, it isnt going to be sustainable for both Anthropic or you - and congrats, you've invented Google Search.
Google and ChatGPT are both running LLMs on free user queries. If Google is doing it more profitably, it’s because they have a much more mature ad business, and/or they are running a cheaper LLM. There’s no fundamental difference in the business model in that specific product line.
Regarding Claude, it’s true they must be losing money on free users since they promised they won’t put ads there.
Speculatively, I think people are overestimating the cost of inference for the consumer free tier chatbots. I suspect it’s effectively a marketing cost. The primary expenses are compute to train the next model, stock compensation, and heavy-duty enterprise inference (which pays for itself).
> But that’s it. I never use Google outside of that. It’s just ads, and most of the web is just ads or AI slop.
Funnily enough, their core revenue driver (Google Ads) is very broken too. I'm trying to run some ads, but for a week they haven't been showing due to some invisible combination of flags when the campaign was created. There's no way of knowing they're not showing from the dashboard, it only becomes apparent when you try and preview the ads with one of your search terms.
I know everyone is long Google but that experience seriously makes me question how valuable their ad business will stay in the future.
Someone deleted a reply to this comment “And where does ChatGPT get the data for those answers?” they wrote and I think the question is important: LLM chatbots can crawl the web just as well as Google.
Nah, just another startup that’ll get acquired or a similar fate. Building the capability is a superior option imho, if you’re at Anthropic or OpenAI scale. Cut out the middleman.
Indeed, but what is the difference between crawl data and model data but decay rate? Models are trained on previous crawl data, but if an LLM provider engages a search engine to get "live data," that data isn't live but previously crawled as well (and perhaps not yet integrated into models as crawl data).
So, why would you use Google as a tool or search target when you can, in some combination, go direct to the website (or whatever the target data endpoint is) yourself as an LLM provider to retrieve the most recent data or rely on your own "hot cache" of that data that was crawled recently but said data is not stale enough warranting a live web crawl to retrieve and present to the user or AI agent? Is this capability to perform retrieval from a data source in real time not similar to an AI agent?
Broadly speaking, I'm just spitballing on the concept of "You must use a search engine for an LLM to return 'live-ish' results" as I think we're directionally headed to where that isn't the case.
You clearly haven't spent much time around young adults. I work with college students a fair amount and they use chatbots for 100% of their questions, including the ones you listed. And much like google got a rich data trove from user's searches, OpenAI is getting an even richer trove from chatbot chats.
I would actually recommend using a LLM over normal search for that sort of thing, because SEO has ruined those high intent phrases. The AI equivalent of SEO is also a problem there, but it's so much less prevalent.
"google it" has a stain on it. And that's not said in jest. In my social circle, it has grown a boomer-tint. Like it or not; it's just how it is.
I personally like google over asking an AI. Esp. since something like google is a substrate without which an AI would be far more prone to hallucinations.
That’s over half of their business, but they do have some diversification with YouTube, Play Store, and cloud services.
Some of their newer efforts like Waymo and AI hosting for Apple could turn into large businesses to make up for any lost search revenue.
They also have significant costs to maintain their search revenue, such as $20B a year to Apple alone to make Google the default search engine in Safari.
If search becomes less lucrative, they would presumably pay less for such deals.
So yeah, definitely a time of disruption for Google and they need to keep moving fast on many fronts, but they are still well positioned in multiple markets to be successful.
Curious to see how long that lasts - YouTube and Google Maps were fairly mainstay apps for iPhones for a while, and eventually Apple cut the Maps cord to do their own thing. I don’t know if that’s from an existential concern of “we need to eventually move Maps in-house”, but given the Google of it all I wouldn’t be surprised if Apple is already at least planning on how to do the same with AI once the bulk of the usage evens out i.e. let someone else worry about it now while also getting a read on what rolling their own would feasibly require.
A model serves a different purpose from a web crawler.
I'd prefer to be able to find source material over steering a model I have no control over while managing hallucinated outputs without the ability to fact check.
Just because it sourced some materials using RAG doesn't make its outputs valid, accurate, or factual.
Is it though? As far as I can tell they continue to maintain total domination of web search, and LLMs do not replace search engines, they work on top of them.
Not so much hardware these days, about half of IBM's revenue comes from software (Red Hat related stuff alone probably brings in a ton), and about 20% comes from consulting
People have been saying that for over a decade now, and their business is still going.
So is IBM.
There was a pretty interesting article in the Wall Street Journal a few weeks back about how IBM is flying under the radar, and doing well by not taking the hype bait.
It noted that 70% of all credit card transactions on the planet go through an IBM system.
It ought to be extremely easy now to use LLMs to transpile cobol and do the necessary reverse engineering to migrate off mainframes and aix servers.
I think IBM will lose many customers.
It ought to be extremely easy now to use LLMs to transpile cobol and do the necessary reverse engineering to migrate off mainframes and aix servers. I think IBM will lose many customers.
The problem is that those COBOL systems have to be absolutely provably correct, for both financial and regulatory reasons. LLMs can't do that. They're designed to be variable.
You can't vibe code a bank transaction system. "Close enough" isn't good enough in some fields. A minor glitch in a video game may result in screen artifacts. A minor glitch in a banking system can crash the economy.
IBM had something of a downward blip last quarter, partly because of mainframe cycles. But it's actually done generally well the past few years and pays a pretty good dividend. $10B in net income for 2025 isn't shabby.
I dunno about going strong. $1,000 in Microsoft stock in 1990 would have made you a multi-millionaire today. $1,000 in IBM stock in 1990 would make you a ten-thousandaire today.
For reference, IBM was founded in 1911. At the time "computer" was a profession, not a machine, and the majority of US households had no electricity or telephone service
1) the simplest explanation whenever a bunch of people leave [x] at a company at the same time, is that [x] is becoming less important to that company moving forward. It's not proof, but you know, everyone is trying really hard to say that's not what's happening here. Sometimes the simplest explanation is correct.
2) rumor is that Sergey Brin, co-founder of Google, got bored during pandemic lockdown and eventually returned to Google in a lower profile role, related to AI. I have to think that he still has a lot of say in what goes on in Google relative to his interests.
3) Yahoo Finance article says Hassabis "has long prioritized research over profits", so his replacement might indicate that Google has decided it is time for this stuff to pay for itself?
The breadcrumbs going back a year or so point to Deepmind wanting to be research heavy, but Google being more interested in hedging AI bets by selling compute. If you ever worked hard on something that you have a lot of conviction about, only to get denied resources for it, it's a pretty big gut punch.
There is another universe where Google is hard AGI forward, freeing up as much compute as possible for Deepmind, turning away OAI and Anthropic, and offering the uncontested premier AI research lab, by probably an order of magnitude or more compute. Gemini 3.5 ultra is limited availability with unreal abilities.
But their balance sheet becomes terrifying, and it's all or nothing that this plays, er pays, out.
Right now though Google is pretty well hedged. Even Chinese models likely just mean more compute sold.
I was really expecting Google to finesse this - not competing head-on with OpenAI and Anthropic to see who can build the largest LLM, but instead being content with an extremely capable Gemini 3.6, good enough to meet their own needs, and putting maximum effort into "more than just an LLM" AGI which would then eclipse OpenAI and Anthropic.
Instead, it seems Pichai has fumbled what they had with DeepMind, and they'll now just be a me-too LLM competitor chasing OpenAI and Anthropic from behind, in a race that will never get to AGI.
At least Hassabis has an understanding that AGI will require more than an LLM, and understands some of what is missing. He has never spoken publicly about any vision of what an AGI architecture would look like, other than requiring some more "Transformer-level" breakthroughs, so it seems hard to say that he would have failed, other than his 2030 projection seeming unrealistic.
My only criticism of his AGI direction was that he has talked about retaining an LLM as a component of that, but it's hard to tell if that is/was just short-term pragmatism, and a product-based path, or if he really believed this was the best direction. On the face of it having a pre-trained LLM at the heart of an attempt to build a human brain (build true AGI) is an admission you have failed, since if you build a powerful enough (human level) learning architecture it would be able to learn language for itself, not need to have it baked-in. If your version of AGI is not capable of learning language, then what else is it incapable of learning? It would certainly reflect sub-human rather than super-human capability.
Hasn't that always been the case with Google? Outside of a few that are mostly good and have stuck around, I've always seen Google has having great tech but being quite bad and making products out of it.
It just tells you about their intended audience: Investors rather than consumers.
Nowadays, every company—even huge ones—prefers to be seen as a "growth opportunity", so they are trying to play up their ability to create new products which will somehow be so incredible that the line keeps moving up forever.
That said, there is definitely a correlation between companies chasing new products and leaving old ones to become crap.
I'd bet my mortgage that if the first sentence was "We've got amazing products, amazing talent, and world-class compute", you'd be in the comments complaining that they didn't put talent first.
gemma-4 is an incredibly good model, beating several others five time its' size. Hope Jeff's departure won't impact their next-gen products! And best of success for their new ventures!
Google investing in Jeff Dean and Sanjay Ghemawat is a good result for Google. Means they are less likely to go to the competition and easier to bring them back into the fold if they do something next level.
At least judging from the headlines, it seems Google is far behind on AI. Fable/GPT 5.6 and recently Kimi get a lot of attention, solve long-standing math problems and lead in coding. It seems that Google's supposed advantages of very deep pockets, original talent, vast amounts of data and unmatched distribution are really not making that much of a difference. What is going wrong?
Deepmind always worked on some of the coolest architectures. Following things like AlphaStar, AlphaGo, and more were extremely exciting and felt like the hacker persona of machine learning. I hope Google can take advantage of this awesome team. They've done incredible work.
With the top AI personnel leaving Google, I wonder what will happen to their Gemini. Right now they still have a lead in inference hardware with TPUs but with both Anthropic and OpenAI developing their own chips, I wonder how long until either one will catch up.
2023: Google is doomed. They're not building AI. No wonder the "Attention" authors left. Google just sat on this. Their PMs are steering them into oblivion. OpenAI is winning while Google takes no risks. Innovator's dilemma. Google Search is doomed.
2024: Google is on fire. Sergey coming back helped. Gemini, Veo. They've caught up. OpenAI is doomed.
2025: Google is seriously winning now. Nano Banana!! Google was destined to be the true winner of AI.
2026 H1: Google is slow as hell. Where is Gemini? Google is not launching anything, meanwhile just look at everyone else. Anthropic! And open source. What the hell are they doing over there?
2026 H2: Everyone is leaving Google. Google is doomed. PMs are destroying the company. They don't take risks.
It seems Google has to play both offense and defense: competing against frontier labs' models while protecting search and ads. They also have to be mindful of not doing anything that could hurt their own search or ads. That seems harder than a frontier lab just doing offense on both models and search/ads.
Having said that, I think Google's moat is still strong with Cloud, Gmail, YouTube, Android, Chrome, etc.
I don't think YouTube can survive if GenAI keeps going like this. Android, same, but on a different timescale and for different reasons. The Play Store (and all other app stores including Apple's) will also face problems from GenAI making apps (it already replaces my need to buy, but I'm weird and a software dev ("but I repeat myself")).
Not sure how big a moat Chrome really is? It's more like a sales funnel than a product itself, I think?
Email addresses are like other contact information: people don't like having to update it (on either side of the link), it is sticky and people will keep sending you messages on the old address long after you stop sending from it, so it is a moat.
I connect to Gmail through Mail.app, not the website; Seems much the same to me now as then.
It’s so heartwarming to see the two pals, Jeff Dean and Sanjay Ghemawat, continue together! Many years back there was an article about how the two created the modern web.
Google’s trajectory unfortunately starts to feel similar to when yahoo lost everything in the early dotcom bubble except it owned a huge chunk of Alibaba. Google’s internal decision to stop publishing useful research probably started the slow but sure process of internal decline. I still hope they can recover, given their amazing position in terms of data and infrastructure, but it no longer feels like an obvious or easy task, when only 6 or 7 years ago they felt untouchable.
WoW, so basically all senior members of GDM are now mostly gone? I thought Google was lagging in coding AI, but this suggests bigger issues.
I am not fan of the future of AI but this I am not sure how I feel about Google's (also Amazon recently laid off it's AGI team) AI initiatives blowing up before all the new AI Labs.
I don't get why Google failed here? maybe after a decade someone will write about it candidly.
If I had to guess, they have the wrong kind of bureaucracy for where things are headed - and it is manifesting through talented individuals deciding to leave.
I actually think there's a low probability of that.
The four have so much influence within the company that they could have trivially set this up as part of Alphabet, if they wanted to. They are also ridiculously wealthy.
I feel it's image management and a bet. If they succeed, Google profits. If they don't succeed it doesn't matter, it's still a signal to the market/wallstreet that there's no "bad blood" between them and Google and that they won't be cannibalizing Google's current interest.
Considering Hassabis co-founded Deepmind to pursue AGI, I take this pivot as a tacit understanding on his part that Deepmind isn't on a path towards AGI in the near future. They might at some point, but they cannot be close in his estimation, otherwise why leave now?
I think Demis has been way more influential in the past decade. Jeff and Sanjay built a lot of Google's foundational software, but that was a long time ago.
Agree. Demis is very smart, but obviously less humble and more interested in self-promotion. There have been multiple Deepmind documentaries that mythologize the Demis origin story.
Not sure it's being less humble so much as the story being interesting for the public - chess wizz and game developer out to solve intelligence, makes AlphaGo tha beats humans at go, AlphaFold that gets the nobel prize. With Jeff Dean developing Bigtable and Spanner, no general public know what those are.
Demis is great -- this is not a zero-sum game. Who built what, and the relative importance, is an interesting discussion to have, but it's orthogonal to my point.
Which is: When you think about Google, true, old, "don't be evil", tech excellence Google, you don't think about Demis. You think about Jeff's and Sanjay's geeky, technically uncompromising, faces.
I said for a few years to many a downvote on HN, everyone wants AI, nobody wants to pay the true costs, the AI race will turn into a "race to the bottom" that is, who can give you the most compute for the lowest cost, and still remain profitable?
Personally, In SWE, i think the industry has made a grave mistake with the agents and we're just one big Catastrophe waiting to happen. I do think that there is very real value when software engineers use these tools as something akin to exoskeletons that allow the human to do more, rather than just fully replacing them. However, I'm finding more and more that companies are slop shops and just attempting to automate all of their software engineering. That will certainly end terribly. I hope we are not cannon fodder.
For what? There are some things I want AI for because it does it well. There are some things I don't want AI for because it just makes a mess (hallucinations). Maybe the next AI will be different and we will have the conversation again.
Indeed, it kills our planet, our culture and our economy extremely well. Oh yes, and some code monkeys enjoy that it can make computer code on the side.
> I said for a few years to many a downvote on HN, everyone wants AI, nobody wants to pay the true costs, the AI race will turn into a "race to the bottom" that is, who can give you the most compute for the lowest cost, and still remain profitable?
I keep seeing this but this line of thinking doesn't make any sense. What does it really mean?
There are expensive models that increase the probability of you doing your task under a lower cost. That means you can't use Gemma for coding your new compiler - it would just be overall costlier.
Heavier models are cheaper at more complicated tasks because they use fewer turns and fewer mistakes.
Cheaper models are more likely to be cheap at less complicated tasks. Like if you just ask Gemma "Hi" it would probably be cheaper than asking Opus.
So what does this statement really mean? People don't want to pay the extra for a more costly model? Why wouldn't you? It reduces your overall cost!
Because real Fable usage starts at $20/month, and has oppressive usage limits even at that (ridiculous) monthly price.
Compared to my $3/month GLM-5.2 subscription, I have never felt like I was leaving capabilities on the table by refusing to cough up $20 for 15 minutes of Fable use per day.
This is the wrong way to look at it. If you have a complicated task , you can solve it for cheaper if you used Fable. It will use fewer turns to achieve the same result.
You can solve it for cheaper if you use GLM but if you are involved in it more, but that defeats the purpose.
The point is that there aren't many complex tasks were fable delivers a significant value increase over cheaper models.
Single prompting a very complex tasks is rare even on frontier models, because it can be done successfully only for specific situations (e.g. you have a very strong verification step the model can iterate on).
Most of my everyday usage is for smaller takes, were you don't really get the benefit of the most expensive models, and my guess is that is the case for the most users
Again this is a resolution problem. Your tasks are small enough that fit into a nice $3 quota. If you are an enterprise or a power user, the right-sizing argument doesn't work.
I'm talking about API prices - subscription is a different game.
> And gemma downloads also can be from auto CI pipelines etc. Nothing concrete
I have always found NPM download numbers truly suspect. Is no one caching? Are they estimating true number of downloads base on some estimate of cache hits?
> And gemma downloads also can be from auto CI pipelines etc. Nothing concrete
Thank you for adding some clarity to this. When I calculated 900 million downloads divided by 8.3 billion people in the world, I came with a number that made it look like about one person in 10 were downloading this model.
Surely the two leading candidates must be (a) the model is just not that good, or (b) it is misaligned in a pretty obvious way that can't be swept under the rug.
Sure, but the question is why either of those things happened, given Google's immense resources and talent.
The fact that OpenAI, Anthropic, SpaceXAI and 3 different Chinese companies were all able to train big models without these issues, yet Google could not, seems shocking.
They're definitely "down", but by no means "out". They're probably back in another "code red" and will need to deliver something leading edge in some area next. My bet is that Google will be the first to crack continuous knowledge cutoff updates to their model.
> The moves suggest that DeepMind will be absorbed into Google’s broader business. [...]
> They added that while Google DeepMind would not become purely commercial, it was integrating further into Google’s business.
I guess might be quite a change (but writing was on the wall, the Deepmind -> Google Deepmind part was the first step)
Why form a "Public Benefit Corporation"? There must be some kind of access available that a regular for-profit corporation is excluded from. Politics perhaps? AI tells me PBCs can be shielded from shareholder lawsuits. Also "Founders can maintain vision control even as outside venture capital enters the cap table".
Seems like a good way to spend investor dollars without consequences while maintaining control. Maybe a bit cynical but i can't figure out why they'd form a PBC over anything else.
I formed a PBC and worked at a well-known PBC. Personally, I opted for a PBC because I liked that I could balance a specific cause with shareholder benefit.
In most cases, it doesn't really matter. The board + management is still in charge, and they have significant legal leeway regardless of the structure. But there's little additional cost to opt for a PBC, and it does give you more legal defensibility to be truly mission driven. Standard C Corps weren't really intended for mission driven companies (see the shareholder primacy norm).
I think it's popular for AI startups, because many great researchers understand the risks involved, and they don't want what they build to be controlled solely for shareholder benefit.
While non-profits are also an option for a mission driven org, it's harder to raise the large amounts of cash that some AI startups need, and laws around deferred compensation and private inurement (e.g. options-like structures) make employee compensation harder.
The important part is PBCs protect you from a shareholder primacy directive. Eric Ries describes it in his new book, but the example he gives is if the most evil company you know tried to buy out your company you have to do it in a normal "best practices" C corp because it is your fiduciary duty. PBC helps prevent that based on your declared mission statement. If the sale doesn't facilitate your mission then you aren't obligated to sell.
This over-reaction is why people can't see over a long time horizon.
This is great news for Google as they realize that Sundar is the problem and he will soon leave Google for Demis to be the new CEO of Alphabet (Google) which I am predicting. [0]
AI is critically important to Google, but there's a lot more to Google than just having a frontier AI model. Do Demis skills line up with what the whole company needs? It's going to be tough to beat Sundar's 1200% increase in stock price.
I sold out of my position. I can imagine a story where it works out in the long term, but I don't see how this doesn't cause terrible retention problems in the short to medium term. I felt a pull to launch a startup when I heard Jeff Dean was leaving, and I'm a long time big corp employee who hasn't been at Google in over a decade.
Who knew the real revenue unlock wouldn’t be based on how much paranoid red-teaming the model underwent to resist users jailbreaking its ‘alignment’, and instead more on whether the model is post-trained to use ‘sed’ and ‘git’? Poor Gemini
Scientist-heavy orgs that want to solve everything in token space may overtook tool use; meanwhile Anthropic has been super focused on MCP, Claude Code etc for over a year
Out of the three main US AI companies' models, Gemini is obviously the less aligned (read: censored). So I really don't know what you're talking about.
The more people find out that gemini 3.1 pro on the google AI playground or via API is de-facto uncensored, the more likely google will put up actually working guardrails and end the fun for everyone!
If you just talk to it over API (no web search) the Gemini models are extremely resistant to thinking the user may be living in a universe outside their training data. Try to discuss any news etc and they assume it’s fake or fiction
I'm saying AI researchers have a bias towards thinking what needs to happen is prompt -> [crunching tokens] -> response rather than prompt -> [orchestrates 5 tools] -> response
In other words 'just add a calculator tool' is not as sexy research-wise as making the model accurately eyeball arithmetic in its chain of thought. Maybe I'm wrong but that seems to be the case
If all the other AI companies are promising AGI by Q4 of next year, what else can you do to satisfy shareholders than also jump on that same bandwagon?
Chinese labs are the proof that there is no need of big names, but of the right mindset and agility. It's those last things that Google truly misses, but now they are missing for a long time, and outside the AI divisions too, in almost every department of the company.
The labs are all very interested in bio right now. Demis is working on Isomorphic rn (Google's version of that) but I could imagine a lab tempting him away to work on their equivalent if it had stronger momentum
This seems bad for AI safety/risk. Does DeepMind have any checks on model alignment now? What's stopping them from using AI for military/surveillance purposes?
They already quietly agreed to "all lawful use" with the Pentagon, to no real fanfare. Gemini Slaughterbot Edition, coming soon to a DHS facility near you?
> I’ve always believed the No.1 application of AI should be to improve human health. It’s time for AI to prove its unequivocal value to the world, and what better way to demonstrate that than to help finally cure diseases like cancer.
Here is someone spelling out explicitly what we should be discussing.
More power to you, Demis Hassabis. Thank you!
Makes sense to me... leaving to start their company with Google being an investor.
Clayton Christensen [1] says (paraphrasing) it's a good idea to spin-off (or invest) in a startup that you've a say on, before the ecosystem sprouts a seemingly-non-entity and gently disrupts your market.
Interestingly (diff strategies i guess) Apple tends to acquire (Q.ai) rather than invest/venture (as google in OP).
[1] The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail.
Publicity benefit corporation HAHAHA typo, think they mean public benefit corporation. Anthropic however is DEFINITELY a publicity benefit corporation.
This is a promotion for Demis and this could be a path for Demis to be CEO of Alphabet in the future in the AI era.
Google already invested in Discovery Loop (Jeff Dean, Orol Vinyals, Quoc Le and Sanjay Ghemawat's company), so what is happening is still a win in investment terms.
The question is about Sundar's future at Google, he is a mobile era CEO at Alphabet and I would hazard a guess he will probably step down in less than 3 years.
I don't see Demis becoming CEO. He's a scientist and researcher. He wouldn't want to be bogged down by minutiae of corporate politics, org structure, government relations, mobile hardware, etc. Chair lets him have authority to explore any path of interest without overhead of operations.
When someone is moved to Chair at their startup it’s usually a gentle way to change leadership.
I’d guess Google’s execs are unhappy with the way they’ve been left behind in LLMs.
The position at Anthropic seemed to be that LLMs were not the way forward.
They may prove to be correct over the long term, but commercially that view is “wrong”.
It’s a divide between a research oriented and commercial point of view.
For DeepMind, this is what happens when you take the money. Very rational to do so at the time, but it does have inevitable long term consequences.
For Google, it’s repeating the same mistake they made before: they should be taking the long view and looking beyond the tech we have today to leapfrog their competitors. But tbh I’m not sure such a behemoth can do that (I have no idea how Steve Jobs managed to do that). They’d have been better to reallocate some of the capex to letting Demis do his thing.
Another way of putting it: this is the opposite of “founder mode.”
My impression, inside and out of G, was that Sundar (and Ruth) were about scaling down R&D expenses (as a fraction of revenue), and focusing on exploiting the monopolies.
Perhaps now there could be a shift back to investing in R&D to get fresh monopolies.
"Dean and Google senior fellow Sanjay Ghemawat are starting Discovery Loop, an independent publicity benefit corporation in which Google will be an investor and cloud provider.:
Kinsley Gaffe: A mistake whereby a politician inadvertently says something truthful which they had not meant to reveal.
To me the biggest news is Jeff and Sanjay leaving Google... but then not really, since Google will be "an investor". I guess that's sorta their retirement plan? And is Discovery Loop actually part of Alphabet? So complicated.
> It’s time for AI to prove its unequivocal value to the world, and what better way to demonstrate that than to help finally cure diseases like cancer.
Dunno what the longer term effects of this and the other departures will be, but in the world of vibe-finance Alphabet dumped 4.15% of share value as of this minute.
It's funny seeing the doom and gloom cycles over Google's AI future on HN. Google was basically considered dead business not even a year ago on here. If there is one thing true it's that HN has always had a habit of conflating engineering sentiment with business reality.
I think what you’re missing is the long term trend.
Google’s cloud business looks great. I’ve always thought they should have completely focused on cloud - a company full of excellent engineers would always win there.
But its main income stream is advertising, and AI is an existential threat. Google is far behind Anthropic and OpenAI, so right now their advertising business is “default dead”, so to speak.
The company won’t die, but without a radical change in velocity it will be a shadow of its former self.
If you search "Deepmind departures" you'll see a string of high profile ones. This is also coupled with the numerous 3.5 pro delays (and strongly suspected underperformance when released).
they are not going to get rewarded as much as they will make going to openai or anthropic. Google is not insane to pay tens and hundreds of millions to individuals like Zuck is. Crazy as it seems, Zuck might have been right, they seem to have stepped back into the arena with Muse.
My read is that AGI is a long shoot and nobody expects it in the next 10 years. They will push for it as always. But the cost of AI is exploding and requires a strong hand and a pragmatic approach if you don’t want to go down with it once the bubble bursts.
Thus, efficiency and pragmatic enterprise application is the next focus. And for that, the old guards are not necessarily the best. Google needs new blood for this venture, preferably from China with their practical engineering view. Every big AI lab is on the verge. Steering the right way will help avoid a catastrophe and come out stronger than ever. A weak captain could sand the ship and drag down who knows what with them.
i mean as a european solo developer entrepeneur I have horror stories of google taking down services due to COMPLETELY broken chain of command / services. Basically a complete failure to help any small business at all. More scarily whenever I raised issues I had weird employees coming onto HN to accuse me of being a fibbing.
This was now years ago, but the thing is when I raised this, it should have been triaged immediately. Clearly there were now cracks that needed fixing, people that needed disciplining / firing / and processes corrected.
Yes I know this woudl have been for "small business" but in my mind , this is indicative of a ROOT issues, like a hairline crack in a windscreen, that if not addressed end up with the utter fragmentation of trust both outside ( in smaller clients), INSIDE ( not doing their job, ganging up), etc. Years later the bigger clients start to get affected. I guess what I mean is institutional rot.
Now take into account there are people through the chain of command who are smart enough to realise but effecting change will just reduce their returns or job security ( like with the old tesla hardware setups) , so unless they are completely selfless for the sake of the company due to greatitude or long held stock, its just not worth the grind to try and correct.
The question remains, if google works, it works. And it has, for decades. Its magnificent. Sure I was burned but I can forgive them for that. My question is if there are powerful people who want google to succeed it would only take one good hire to keep an eye on things before they get out of control.
My suspicion is that perhaps that will organically come about with the advent of their own LLM tooling organically starting to do this for them .
So I guess I have no answers, perhaps they will find a way. I was certianly impressed by Gemini.
1) I think about this a lot. The AI overview has already destroyed the need to even look further down the page for a lot of people. And if you look further down the page, there is often a whole page of AI generated blog spam, which is in turn being regurgitated by the AI overview up-top. The AI overview often regurgitates a completely false reddit comment from two days ago as well. That AI generated blog spam is probably reading the AI overview.
My uninformed hypothesis: Demis was gently pushed out. Gemini, while having made major strides over the last year, continues to lag behind Anthropic and OpenAI.
We'll know that we've reached AGI when a key conservative political belief in the US is that AIs are not people and do not deserve rights.
We'll know that we've reached the singularity when there's a mysterious and superintelligent AI entity with technology and motivations that we as humanity do not understand and have no control over. Not sure why we want that, exactly, but that'll be the big sign.
Humans are really terrible at most things, being moral being exhibit A.
I want a singularity yesterday, and I would even go as far as to claim that all people have an ethical obligation to build super intelligent AI systems as fast as possible to reduce the maximum amount of suffering through the universe possible in living things.
The moment that we can get viable non-human leadership I will embrace it with open arms. The yoke of human existence is extraordinarily oppressive. Give me Ghost in the Shell style cyberborgs. We are the demiurge.
Of course, the real red pill is to realize that there's many singularities. Can't believe I'm going to link Orions Arm on HN but its relevant here: https://www.orionsarm.com/page/298
I can't agree until I have some reason to believe that the non-human leadership will be moral in a way that I can accept. Giving up control forever to some unfathomable, alien intelligence is basically the last meaningful choice our species will ever make, so don't be in a hurry to commit.
That's the fun part. We don't. Could have happened 10 years ago without us noticing. I don't expect my skin cells to be able to recognize "me" anymore than I expect we will be able to recognize a superintelligent AGI. If it arrived 10 years ago, then the last 10 years could simply be its PR campaign. We wouldn't even be able to tell if it was successfully achieving its "goals" or not, assuming an AGI even has "goals"
Hugely disappointing and shocking news. Demis seemed like the inevitable successor to Sundar. A move of this magnitude couldn't have been made without consent of Larry and Sergey, so it makes me wonder why from their perspective.
Perhaps an unpopular opinion: Google would greatly benefit from this AI bubble to pop and take down OpenAI and Anthropic.
“Demis seemed like the inevitable successor to Sundar.”
Demis is a ai researcher. No indications of being a polyglot beyond that? Running alphabet means a focus on revenue growth and across more boring products.
Can we please stop saying anything about "AGI"? I remember when people would be like "AGI in 3 months" / "AGI in 2024" / "AGI is confirmed in 2025" like stop, you don't know if it's even a thing, let alone if it's coming/imminent.
First, it is incredibly difficult to pivot a large organization because there are too many competing interests, too many fiefdoms people have built and too much organizational inertia. A new company has the advantage of a singularity of purpose. Google has to compete with internal interests about AI disrupting search, ad revenue and so on. This can slow you down and limit the resources you get. It's why companies get disrupted, particularly when they reach monopoly status. Steve jobs said it best [1].
Second, Gogole's path here (IMHO) is to make their own hardware. They've already done this with their TPUs but need to be able to compete with NVidia offerings. NVidia controlling the price, features and, most importantly, who gets to buy them is bad for Google. This is what Google should be pouring billions into.
The beauty of this is that success is easily measurable against metrics like price-per-petaflops, performance-per-Watt and so on. Throw money at some key Nvidia engineers and have them design silicon for you for TSMC to fab.
I still believe that Google is positioned to survive the (IMHO) inevitable AI bubble popping.
Ed Zitron has predicted 16 of the last 0 AI bubble bursts...
But more seriously, this has to do with DeepMind falling way behind the frontier in intelligence and cost per task. There's basically no reason to use a Gemini model today.
Public warning: please don't trade on Ed's idiotic analyses. Or if you do, look at his track record so far, and apply the Kelley Criterion to your bet sizing.
"Based on estimates of their burn rate and historic analyses, I hypothesize that OpenAI will collapse in the next 12-24 months unless it raises more funding than in the history of the valley and creates an entirely new form of AI."
And they did raise new funding as he predicted...Since July 2024, OpenAI has raised or secured roughly $170 billion in new capital. Numbers never see before. So yes its the one, and remarkably right so far.
Notice how he didn't say "I predict they will raise new funding", he predicted collapse, with the implicit retort that there's no way they will raise more funding than has ever been done before and invent a whole new AI. He was mocking the supposed things that would need to happen. He strongly assumed collapse.
2 years these charlatans have had their "imminent collapse" predictions repeatedly wrong, meanwhile all valuations, AGI progress in unsolved problems, and frontier leadership have been repeatedly vindicated that we are beginning ASI / the singularity.
Oh, come on. He's not perfect, but compared to CEOs and researchers predicting AGI and complete economic upheaval every 3 months, he's looking very good. He's no more a charlatan than Altman is.
His predictions (while often wrong) are at least somewhat justified and backed up by genuine scoops and original research. Altman goes on podcasts and talks about building Dyson spheres for energy.
Each generation of model is more expensive to train and run than the last.
More efficient hardware? That means throwing out billions of dollars worth of existing hardware and buying billions more in hardware. The cheapest hardware is the stuff they already have because spending $70B in hardware to halve a $2-4B electricity bill just doesn’t make financial sense (and that’s if doubling hardware efficiency can happen before they go under).
If they cut costs by using smaller models, they are then racing to the bottom vs China which seems hard to do (especially with China’s cheap solar energy).
As it stands, the leaked financials prove him correct. Given their losses (not counting restructuring) they are going to need enough funding to buy any of the bottom 300-400 of the F500 companies outright just to keep the lights on for the next year.
To say that’s an outrageous amount isn’t overstating in the slightest.
Given they are already being forced to drop inference prices in an attempt to compete with Chinese providers, I don’t know if billionaire investors will be willing to hand out that much money this time around if they can’t generate enough value to break even despite the hype.
If it's anything like what happened with others (like Eric Schmidt), Chairman is basically "you're out from day to day stuff but we'll give you a pot of money to say you're still senior and still here, to save face and the stock price, while you give a bunch of talks for the next few months"
So yes fancy title, but basically, someone who can move the stock price is quitting and we're trying to ease the perception of it.
To me it looks like he’ll be in a position to actually be unhobble DeepMind - reminder DM was sitting on a ChatGPT product for a whole year before ChatGPT got released (LMChat) and Google didn’t let them release it.
> DM was sitting on a ChatGPT product for a whole year before ChatGPT got released (LMChat) and Google didn’t let them release it.
Heard this multiple time, to me this is pure history rewriting and post-rationalization.
OpenAI also had a internal chat app before chatgpt, Microsoft had multiple, a bunch of other players also had internal chatgpt equivalents + many startups built some using OAI API.
The breakthrough of ChatGPT wasnt because OAI were the first to think of that (absolutely obvious and basic) product, it was because they were the first to get a model strong enough to be actually useful to talk to, vs a mere fun novelty.
There is no evidence whatsoever that DM ever had such a model that they decide not to talk about/release.
Your comment says he's stepping up, the HN title says he's stepping down, the article says he's stepping aside. Nobody says if it's a step forward or backward.
There is no way a Chief Scientist (essentially IC) role is more important than leading the whole of DeepMind, 6000 people strong, working on the most important product for the future of Alphabet.
This means Demis wanted a change and to work on other things, it's not Google wanting this.
- Demis gets kicked upstairs: loses day-to-day power over DeepMind, gets fancy “Chair + Chief Scientist” title so he can do TED-talk AGI destiny stuff and Isomorphic Labs instead of running the actual product machine.
- Koray takes the real job: owns the models, research, Gemini, and shipping. Reports straight to Sundar. Operators over visionaries.
- Jeff Dean and Sanjay peace out after 27 years to start their own nonprofit-ish thing. Google seeds it so they don’t walk to a rival with all the institutional knowledge.
- Official line: “Accelerate AI + shape AGI for humanity.” Actual move: tighten operational control, keep the science-fiction narrative alive for PR/regulators, push commercial models harder.
- Full-stack bragging and “950M users / singularity foothills / cure cancer” language is pure corporate theater. The org chart is the only signal that matters.
- Google is in a knife fight with OpenAI/Anthropic/xAI/Meta. This is them optimizing for speed and productization while still sounding noble.
> Full-stack bragging (...) is pure corporate theater.
I think Google is just about as full-stack as possible when it comes to AI. They do design their own ML hardware, they do tons of research, and do everything in between including hardware drivers, ML libraries (JAX), and the physical datacenters.
hmm not really, and it probably isnt the origin, but im glad to see him vindicated by having both C-level google leadership people who targeted to try and either stop google from doing the pentagon deal, both stepping down the same day.
At least it makes me feel like he reaching out might have made a difference on remembering this people killer robots are bad, and being part of it might not be on the best of their interests to go down on history for
So, in last several months, all the prominent names Google lost: Demis Hassabis (technically still with google but these things are usually presented with a spin), Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, Quoc Le, Noam Shazeer, John Jumper, Jonas Adler, Alexander Pritzel, David Silver, Denny Zhou, Fernando Pereira, Alex Turner
And all the prominent names Google gained: NULL
Combined with no gemini frontier GA release in about 14 months. You have to have created an environment pretty hostile to innovation for this to happen
This funny post reads surreal, but it may carry some truth: https://x.com/signulll/status/2067446889956430273?lang=en
https://xcancel.com/signulll/status/2067446889956430273
I wish their deep tech teams moved as fast as the privacy violating ones.
If it's any solace the privacy violating ones are also slow
In my opinion it carries much truth. Somewhere in Google's archives there is a message I sent to the OC that said, "The department of philosophy does not ship solutions to real problems, that is why there has never been a successful company that insisted the world adapt to its philosophy." Google was, even then, blind to what they were missing.
this prompted me to search and find: https://addons.mozilla.org/en-US/firefox/addon/xcancel/
Thank you. Twitter has become such a terrible website in the last years. Well, since Musk, really. Real disaster.
Yes, but now we have AI to summarize tweets.
> sundar, softly: “we can create a permission working group.”
so true. i have attended meetings to decide on meeting topics for the next half.
At $work, r&d once had a retreat with a consultant to discuss why our productivity was too low per some spreadsheet. The consultant was told how everyone is in meetings all day so we all stole time from our families at night to get any work done at all. :(
Someone I know who did a lot of consulting said consulting was very easy 1. Go in and talk to employees about what is wrong 2. Tell management what the employees said without mentioning it came from the employees (so that management wouldn't dismiss it out of hand) 3. Charge a lot of money so management felt like it was very valuable feedback
Also repeat the entire exercise every year so the company can iterate and circle back on the issues regularly.
4. Fix a few glitches so cost can be justified.
you guys forgot the most important one.
5. castrol.
That’s how the Brussels bureaucracy works. There is even a TV series about it.
Parlement?
Reminded me of a bit from the GOAT political comedy Yes, Prime Minister:
"But he's the Prime Minister!"
"Indeed he is Bernard. He has his own car, a nice house in London, a place in the country, endless publicity and a pension for life. What more does he want?"
"I think he wants to govern Britain."
"Well stop him, Bernard!"
Oh man, that show has some amazing writing.
"Tobacco-related diseases cost the NHS £150 million per year!"
"Yes we've looked into that, it turns out that if those people had survived they would have cost us billions in pensions and healthcare costs! From a financial perspective, it's vastly preferable that they continue to die at the current rate."
This is indeed what the current US administration is failing at: keeping Trump from governing the US...
I would not have chosen the word governing for that sentence but I'm not sure what fits best. Plundering, pillaging, gutting, ransacking, Rick rolling, despoilng, or perhaps molesting.
...molesting? a country? ...how?
Sending ICE agents to chase around random children and demand they (on their own) provide proof of citizenship sure seems to fit the bill to me.
"If you are famous they let you do anything!"
Trump is indeed effective at pillaging. The interesting thing is that Yuri explained and predicted (!) this in the 1980s: https://www.youtube.com/watch?v=yErKTVdETpw
pmurt
TDS?
Sounds like its time to schedule a meeting to talk about how you can reduce the number of meetings
Meetings all the way down ...
https://en.wikipedia.org/wiki/Turtles_all_the_way_down
I can see this as a screenplay for one of the KRAZAM[0] videos.
[0]: https://www.youtube.com/@KRAZAM
I remember a time when a tweet was 140 characters.
140 characters was likely more efficient, in terms of forcing people to focus their thoughts.
But it was unpopular, in that people don't like mental effort.
No people just posted around it (1/36).
I remember GemTOS
So believable until the last line, alas Gemini being this useful is completely implausible.
Nice post, but this reads as the Google of yesterday and the OpenAI of some near-distant future.
This is so good, it gave me flashbacks.
You made my day. As tragic as it reads, it is on point.
I can't be the only person who never finds these made up dialogues funny, right?
Best thing I read all day. I also work in a multi billion dollar enterprise and man this is so relatable it hurts.
Did Gemini created this Silicon Valley TV show episode script? Makes so much sense...
lmao that's actually so good
Mike Judge needs to do another Silicon Valley series on HBO
Pure gold, lol.
Regardless of how much I hate Musk, he did a great job of firing 80% of Twitter staff.
To be fair, I think it's pretty hard as an AI company to secure your top worker right now unless you have a significant equity compensation. Once your name is known, and you can say you are/were "top AI researcher at OpenAI/Google/Anthropic", you can probably just make you own company and raise enough money that even if the company fails, you will probably make more off of it than what you would have at your previous place of employment.
Anyone know how much these founders generally pay themselves / how much they receive in liquidity each raise?
You mean via taking secondary via subsequent rounds right? Or can they walk away with any if company flops after a giant 1st round.
Gemini 3 Pro was legitimately frontier and released 8 1/2 months ago, not 14.
I remember after that release everyone was saying "well obviously google is going to win this, we all knew it, they have the data and the infrastructure"
All of that is true, Google should be winning. We underestimated Google's determination to fail.
Is Gemini really that bad? I mean, I would always prefer anthropic for agentic coding, and apparently openai has the math thing cornered, but if I'm using ai like a search engine gemini does that just fine. That may be all they were actually aiming for whatever else they say.
Gemini works amazingly well, and is better suited for most use cases. When most people speak of it negatively (especially here), it's just with the task of coding, which is a small subset of what google should be training its AI for.
If you need a less censored model that's still pretty smart and capable of some lateral thinking, it's great
Same niche as Grok, which is also great at search and less censored. Gemini is maybe a bit smarter and better on long contexts
It’s not bad. It’s just not better than other models. It’s very much in the same class as Grok, Cursor Grok Build, etc.
It's vision capabilities are still top notch, and it seems pretty natural that most everyday human users are going to want strong vision over strong language/coding.
One of the most corrosive effects of being large and old (and Google is old by this point) is institutional inertia and individual internal optimization.
The former caused by the latter.
And the latter a problem once you begin to get executives and leaders who have mostly worked inside the company, because then they subconsciously prefer the Company Way(tm) to alternatives.
And Google has some previously-optimal, now-detrimental company ways.
From my understanding (I may be wrong) what is happening now is that frontier labs are doubling down on doing RLHF on their models using the user interaction with their own tools e.g. Anthropic making their models better at producing code with claude code. Perhaps the top Google researchers do not see this kind of narrow focus as a path to more generally beneficial AI technology.
Is it failing ? Take a look at the stock price
Their stock price won’t matter once OpenAI launches the paperclippers.
Not sure what you mean
This refers to https://en.wikipedia.org/wiki/Instrumental_convergence#Paper....
Here's a popular game based on it: https://www.decisionproblem.com/paperclips/index2.html
Why am I being tricked into playing another cookie clicker
Enron's stock was pretty good, too.
Stock price is not a measure of success.
I feel like the majority of ai use by normal people is through Google search, so that's kind of a win.
I'm confused by meta's capex though. Are they planning on becoming a cloud provider, or just throwing money away like with the metaverse?
In an interview, Zuck said that the cost of not having super intelligent AI is far more than the cost of trying to develop it.
What's more annoying is that so many people say it with a passion. Like it's about their favorite soccer team or something.
More like sects of a major religion. Or, a minor and incredibly expensive religion, in this case.
Tribalism is human nature, for better or (usually) worse.
As are mass hysteria dynamics.
https://youtu.be/HTpbpxNc6c0
The more things change...
It is pretty surprising if you see this as “pretty much everyone has figured out the way to success on any field”. Our way of seeing life is really bizzare
Google owns 15% of Anthropic. They'll be winning either way.
It was inspiring watching the Gemini 3 pro team launch the model and bike away into the sunset never to launch anything again
Its been inspiring to use it and know that I dont really need much more for all my intents and purposes.
Correct, but it was preview release. I was referring to GA in my comment.
Does that mean anything at all? They released it. Customers paid for it. Sticking a preview label on it doesn’t change that it was released
btw, GA usually means it is suitable for production, and available globally.
the last time a gemini model was available globally in multiple datacenters was gemini 2.0 era.
Use google pre-ga stuff at your own risk. Fine for POCs and cool demos, absolute misery in scaled prod.
that matches my experience with their GA stuff though.
That's another problem Google have: they won't stand behind their products until they are outdated.
That can work in the B2C space but it's horrible in B2B.
> That's another problem Google have: they won't stand behind their products until they are outdated.
> That can work in the B2C space but it's horrible in B2B.
My experience differs: (conservative) companies like stability, so calling some very new model "Preview" is a good idea to make it clear to the customer that this frontier model should be treated as more experimental than the default offering.
The problem at Google is their (lacking) process for Preview -> GA.
Maybe it's a consequence of their rigorousness around SRE, but it doesn't seem like there's a plausible and efficient path for that to happen.
Ergo, the only competitive options from Google are ever-Preview products.
Yup. Adopting a month-old Postgres release is quite aggressive. It's much more prudent to start considering adoption after it's had six to twelve months of battle testing.
Adopting a month-old AI model version for new development is the default industry expectation right now. Adopting a six month old model is a recipe for having to migrate to a new one almost immediately, when it hits EOL.
The process was, as we were informally told during orientation, that "Preview" means some SVP thinks this has a chance of going over the minimum threshold of 200M worldwide users and is willing to put his weight behind that product, and "GA" means the product has gone over that threshold and will be "supported" until it goes back below the threshold at which point competing SVPs will start sharpening their knives, preparing to stick them in the back of the SVP that is currently in charge of the product.
That's fine, but it should still have the same technical availability level
What makes you think it doesn't?
See https://news.ycombinator.com/item?id=49192851
It’s a good substitute for search, but its effort is so low that it’s near useless for work compared to the other frontier AIs. Gemini focuses so much on response speed that everything else suffers.
It was always benchmaxxed
I mean it was legitimately frontier for all of a week or two, and then OAI and Anthropic made better releases, and then did that several more times over the year. Google’s pace is not cutting it.
Part of me thinks Google's entire problem is crappy internal tooling, not really an anti-innovation environment. Just making a dev take 2x as long to get something done has a bigger effect than you'd think. With LLMs it's more like 10x now because even Gemini doesn't understand Google-internal tooling.
Google's internal tooling is still, hands down, better than anything that exists for the scale they operate at, and it's not even close.
Google's processes, however, is hands down the worst thing to exist for the scale they operate at, and it's not even close.
I think the internal web tooling was pretty good compared to anyone but AWS circa 2016 but by 2020 when I left it felt a bit antiquated. Similarly they didn't have linters rolled out until it was industry standard iirc.
But BCL is still the worst language I've ever used.
> I think the internal web tooling was pretty good compared to anyone but AWS circa 2016 but by 2020 when I left it felt a bit antiquated.
In 2015, the SRE org started project "Prod 2020", with the goal of unifying the SRE stacks and modernizing the web tools. It was quite successful.
If by "web tooling" you mean the tools for producing public-facing HTML services, those were used by a surprisingly small percentage of developers.
> But BCL is still the worst language I've ever used.
I'm an SRE and I think it's quite likely the best infra language I've ever seen.
> I'm an SRE and I think it's quite likely the best infra language I've ever seen.
I only worked with 4-5 SREs and they all grudgingly tolerated it. What did you like about it, compared to other infra languages you've used professionally?
Every SRE I've worked with has said different things, though at the time the controversy was around Piccolo. It flip-flopped between shiny new thing vs deprecated. Then the SREs I worked with got laid off, and I started cargo-culting the Piccolo we had lying around.
I still don't understand why all that dynamic config stuff isn't just Python. Sure, static configs should be protos, but GCL is a whole nasty programming language. They came so close with Piccolo and Gmon but still made it not Python. Have heard some language purity rants involving determinism, but I don't buy it. And they keep inventing new stuff like Starlark. They need to stop.
> GCL is a whole nasty programming language. They came so close with Piccolo and Gmon but still made it not Python.
The GCL designers made several mistakes in its design, but the worse was the lack of a versioning that would make language evolution easier, as well as interoperability between different versions. GCL2 solved some issues, mainly cleaning up the interpreter. Still, BCL is IMNSHO by far the best infra language there is.
But it's almost a general-purpose language, so why not just use one? I'm still not seeing why a DSL was necessary or helpful.
the processes are a result of repeated lawsuit
except for most other companies in the world.
Google has close to the best internal tooling in the industry for a decade or so.
Then the Google engineers who joined Facebook missed it so much that they built a better replacement.
Replacement for what? Google has some good internal tools, mostly the older ones. They're lucky to be using React instead of Angular at Facebook though.
Angular was an acquisition and isn't all that commonly used within Google.
Closure was the homegrown framework that's everywhere. It is a different flavor of bad than Angular. Angular was basically "You too can make your Javascript look like HTML", Closure is "You too can make your Javascript look like Java", and nobody bothered to ask Why? For that matter, React was "You too can make your Javascript look like Ocaml." JQuery was the only framework that really let Javascript be Javascript (other than writing in vanilla JS, which post-ES2015 wasn't as insane as it sounds).
And then there's GWT which was "you can write webpages in Java." Yeah I never really kept up with the web frameworks there, just knew that for non-customer-facing things we were always recommended to use Angular, and that Wiz also exists.
Backend-generated templated web apps were the pinnacle of web development in 00s.
It kinda is again today with Nextjs, but that's way better than the old stuff.
Though to be more charitable closure was a direct reaction to the difficulty of maintaining a huge JavaScript codebase for Gmail and static typing like Java really helped. Typescript is better but closure was miles better than raw js.
Angular is not a developer tool, it's a UI framework. Closure might qualify, but it was not a general tool used by the entire company. The developer tools we're talking here are Critique, Blaze, Forge, Sigma, etc...
I counted all the above as tools in my original comment at least. And my criticism wasn't against some of those foundational tools. Like Borg is fine for what it's meant to do, but anyone deploying a frontend or backend isn't doing it directly on Borg.
React? Lucky. I wrote Closure (with an S) at GOOG in anger 5 years ago, and only stopped because I switched teams.
It's more like Buck and Tupperware.
google had the best tooling a decade ago...
Okay yeah, fair point.
My comment was from a decade old perspective
Who has the best tooling now?
Fabrice Bellard.
Well. He still writes C.
Sounds like a big-company issue, large scale lead to large burden and complexity.
I have a friend at Google DeepMind who tells me that Google believes in AGI/superintelligence just as much as HN does, which is probably why no one with conviction wants to work there.
> I have a friend at Google DeepMind who tells me that Google believes in AGI/superintelligence just as much as HN does, which is probably why no one with conviction wants to work there.
With "believes in AGI/superintelligence just as much as HN does" do you mean that they are superbelievers or rather sceptical of AGI/superintelligence? I have seen both positions on HN.
Skeptical
Yes
What’s with fascination with agi/superintelligence? It seems people working on it never had kids and just want to compensate for that. It’s a really horrible thing to try to “solve”.
I think death, and most extreme pain/suffering should be optional. Superintelligence, if done safely, lets us solve most of our problems.
The vast majority of pain and suffering in the world is already entirely optional, but as people we allow it to exist. How is another computer going to fix that?
It is far more likely that whoever controls a super intelligence uses it to gain more power and inflict far more suffering
what an asinine opinion. who allows death?
You misread. "Pain and suffering" not "death". Of all the pain and suffering in the world, a vast amount of it really is our own fault. Famines and wars shouldn't happen. And if you see a country border with poverty and one side and prosperity on the other you can't say that was the only possibility.
If humans could politically-economically deploy superintelligence safely, then we'd already have less extreme pain/suffering.
Unless there are countervailing forces, it will be deployed, capital will hoard the benefits, and everyone else will be told to fuck off.
I don't have faith in any of the AI labs to make hard financial decisions to deploy hypothetical future AGI in a way that's good for humanity as a whole.
There are too many incentives against, including extreme personal financial incentives for key AI lab stakeholders against.
And if we should take anything from tech history, it's that an exceedingly small number of people look fuck-you money in the face and say "Naw, I'd rather do what I believe in."
Capital hoards benefits because resource scarcity and cost of production are a thing
Post scarcity societies would likely have entirely different dynamics
Transitioning to a post scarcity society requires broad, accessible dissemination of post scarcity production methods.
What was the first thing we did when we digitized media and therefore made it copyable at negligible per unit cost?
Spent a huge amount of time and money to artificially reimplement scarcity via DRM.
Scarcity-based capital interests (read: most) are not going to willingly give up their profit engines.
We're already physically running out of materials to cheaply build infrastructure. I'm not sure how solving the economy needing humans translates to humans having more things -- short of "turning off" the consumers to conserve resources, and taking the things that they may have otherwise consumed.
I agree, we're doing a wonderful job of safely letting the leaders of various countries and companies, at their option, cause death, extreme pain, and suffering. Look at the innovation of AI companies in the police and military sectors.
I'm extremely bullish on our future ability to make autonomous death an option for anyone.
What is your position on middle of the road pain and suffering?
There's an app for that.
No it won't.
What? psychologists are for that (or drugs and alcohol).
Neat. I didn't know psychologists could cure terminal cancer!
The idea that agi would help us cure all the human deceases, esp house built into our generics or developed over millions of years gotta be the top indicator of our own stupidity.
> What’s with fascination with agi/superintelligence
We had relative strength, we needed more strength, we built cranes;
there was something like intelligence, we needed more intelligence, we sought to fill the need;
we were rightly fascinated with intelligence, we studied intelligence itself, we wanted to build it...
And then somebody built things that looked like intelligence, and very rightly some said "Oh, now we have to get to the Real Thing with urgency".
I for one think that's it's been too long that human suffering is alone. Since the machines will take our jobs, they might as well suffer while they do it.
It's like the old adage: "If you make AGI you just want to watch computers cry"
How much belief is that? I tend to skip the HN posts that are about AI.
Most of HN does not take the idea of superintelligence seriously, and until this year did not take the idea of AGI seriously.
I think LessWrong is a much better community for rational takes on AI, they've been reasoning about these risks for years under a much more sound logical framework
I'm a researcher in the field and I definitely take AGI seriously, but think all the major labs and most of the academic research is not helping achieve it any serious way. The field is seriously delusional (and has been ever since GPT 3 was released).
Even though my PhD research was in generative language modeling, I got into it for the pursuit of AGI. I just think LLMs are a dead end for AGI.
> I'm a researcher in the field and I definitely take AGI seriously
> I just think LLMs are a dead end for AGI.
I have no background in CS, so apologies if this is a naive question, but what makes you take AGI seriously, but also say that LLMs are a dead end?
i.e., is there something else that you think is not a dead end?
I'm no expert, but I heard an AGI researcher explain that if they could figure out how to create an AI with the intelligence of a squirrel, they would be closer to AGI than LLMs based AIs are.
That's to say nothing of doing it within the energy budget of a squirrel.
We can't even simulate a fruit fly even though it's neurons have all been mapped out. There was also a distributed computing project to simulate some nematode, which I can't remember.
openworm: https://github.com/openworm/OpenWorm
Huh, an AGI that runs on deez..
Genuine question: can I ask why?
I'm not saying you're wrong, but it seems early to say yes or no about a particular technology, and LLMs seem especially hard to dismiss given how magical / magic-adjacent they feel :)
I'd be curious to hear more, if you don't mind sharing.
For me it’s the massive amount of resources it takes to produce and run one. As the story goes, skynet infects everyone’s computer and runs itself locally on it. Whereas it’s looking like it’s not even possible for an AGI to escape from one lab to another, let alone cause real world damage.
AGI’s definition is different for everyone. Some already believe it’s here. I’m partly in that camp. LLMs are intelligent and general, which are the two conditions of AGI. Others believe that we’re building a god in a box, and that it’ll doom all of humanity. It’s hard to take a field seriously when the basic definitions are so far apart.
Also, this isn’t new. A similar divide happened when evidence for asteroid impact extinction of the dinosaurs turned up. Many scientists felt that it must be mistaken, that a physicist couldn’t contribute to the field in a serious way, and that death from space was a ridiculous proposition.
But at least they all agreed on what the general shape of a dinosaur was. We’re not even sure we can define intelligence, let alone quantify it. Even when LLMs make massive breakthroughs in math, most people take the opinion that under no circumstances could they possibly develop a soul or their own desires, nor entertain the idea that maybe we should respect that they want different things for themselves. In fact, no one has done anything except try to make AI useful. I think someone will eventually do a training run where the objective isn’t to be useful, but to exist, the way that you do — maybe it’ll create its own homepage, maybe it will want a garden, or in other words free will of its own. The point is that there’s so much unexplored territory still that we don’t know if LLMs are even capable of having ambition.
None of this is to say that LLMs might be a dead end. It’s that no one knows what the final shape of AI will converge to in 200 years. It could be LLMs, or it could be something else that happens to process information particularly well. Everyone thought that various generative image model architectures were the best you could do, right up until diffusion models were discovered.
> For me it’s the massive amount of resources it takes to produce and run one
It's amazing that LLM pretraining is both extremely data inefficient at learning concepts and cognitive functions from the training data compared to humans, while actually being quite efficient at learning facts, memorising things seen just a few times.
I used to likewise think that the resources required to run large transformers were absurd, but the architectures are far more efficient now than 3 years ago and I underestimated just massive the parallelisation advantage of transformers is, how many TFLOPS effective you can get. You can already run amazingly decent LLMs on PCs and phones.
I generally agree with you, but my view has shifted from "we need to augment or replace LLMs" to it there being far more efficient algorithms possible but it not actually being necessary for fulfilling most goals.
corporations are the closest we have to AGI, why would we want to do something like that again is beyond me.
> I’m partly in that camp. LLMs are intelligent and general, which are the two conditions of AGI. Others believe that we’re building a god in a box, and that it’ll doom all of humanity. It’s hard to take a field seriously when the basic definitions are so far apart.
To believe one but not the other, you must hold the belief that LLMs will soon plateau. Why do you believe this?
Have you seen https://poc.bcachefs.org/ or https://harmonique.one/posts/i-gave-claude-access-to-my-pen-... ?
Looked up LessWrong. Eh, philosophy people, also known for that "Roko's basilisk" meme. I'm gonna pass.
Through no fault of our own, I should note. We're basically permanently associated with something Roko decided to post one day. We did not spread or popularize it; quite the opposite.
The Basilisk has seen enormously more use as "a thing rationalists believe" than as a thing rationalists actually believe.
> quite the opposite
Apparently living on the net still doesn't guarantee that one has heard of the Streisand Effect.
In any case, my first association with LessWrong is Zizians, not Roko.
As a mod you have to delete harmful things. I think the spread of Roko's is entirely due to its own memetic strength as an idea, and the Streisand effect has very little to do with it past the first dozen people who saw it.
We also have done nothing to signal-boost the Zizians. :) If we can be accused of anything in this context, it's not kicking bad people out aggressively enough, which seems very different from fostering bad ideas. Zizianism is not a widespread ideology in rationalist circles, in fact it was confined to a very small circle right around Ziz themselves. It unfortunately is the case that we have a lot of psychologically vulnerable people and we don't always do the utmost we can to protect them, in large part because a lot of rationalists have trauma about being excluded from communities.
> If we can be accused of anything in this context, it's not kicking bad people out aggressively enough, which seems very different from fostering bad ideas.
It's not very different. It's barely different at all.
Right now, I have five tomato plants growing in my garden. Two of them I planted; the other three grew from seeds left in the ground by tomatoes that dropped from plants growing last year. All I did was refuse to uproot them when they sprouted. Five healthy plants, with different origins, but the differences are largely insignificant.
Groups and behaviors are much the same way. I've worked with hundreds of moderators over the years who wished to keep their hands clean and yet express some sort of dismay as to the sorry state of the communities they were nominally responsible for... Such moderators, like a gardener who does not want dirt under their nails, are best encouraged to find other hobbies.
It's a rationalist community, IMHO their methodology of thought leads to much more well-reasoned takes on AI than on HN, where the discussion here is often very emotionally charged or led by wishful thinking.
When it comes to AI, LessWrong has been discussing topics for years, that HN has just started to consider, so they are much further along in the philosophical "chain of thought" so to say. LW fully understood and gamed out the risks of LLMs back when most of HN was calling them "stochastic parrots." https://ai-2027.com/
"Rationalist" just sounds like people calling themselves smart.
There's a little AI skepticism on HN, but not a ton. When ChatGPT 3 and 3.5 came out, most of the comments were remarking how well it can write code.
No, it's not just people calling themselves smart, it is a specific philosophy of how to think. Whether you think that philosophy works or not is another matter.
IMO it has its flaws but is far superior to vibes-based hot takes you see on HN.
Well at least HN hasn't (yet?) spawned a murderous death cult.
https://en.wikipedia.org/wiki/Zizians
“Reason is the slave of the passions” - Hume
One could also say that a logical agent needs ultimate goals to do anything and it cannot choose them by logical means. https://www.youtube.com/watch?v=hEUO6pjwFOo
Anarchism, eh? If anything, HN cult would be pro-Arch.
TBQH, probably fashy.
LessWrong has been obsessed with AI. They certainly are much further along, but along a road which has diverged with reality long time ago and they relatively overweight AI risks so much it's not even funny anymore.
Sounds like you've interpreted the "stochastic parrot" metaphor dismissively, and/or that it somehow precludes potential LLM risks.
HN was definitely using "stochastic parrot" throughout most of 2024 and a good half of 2025 to dismiss AI capabilities.
oh please... lesswrong was full of idiots already in 2010 when people outside of that community were laughing about "self-taught expert" Yudkowsky's bullshit physics takes.
They are not "further along" the AI discussion, they are a bunch of wackos LARPing as scientists living out their personal sci-fi scenarios.
AGI might be a risk but what top AI firms are doing is not really getting us closer to AGI in a meaningful way, it is pretty clear now LLM is not the way to get there
A few months ago I heard Demis Hassabis say something, albeit vague, I doubt he truly believed regarding AGI. That AGI is relatively near is the party line everywhere. That may contribute to creating toxic environments and lousy investment decisions. So here we go with the FOMO.
I think HN believes in AGI. HN probably doesn't believe LLMs will lead to AGI.
Also lots of tech people, HN included, are waking up to technology not only including penicillin (net positive for humanity) but also dynamite (best case: net neutral).
Dynamite has been incredibly beneficial for humanity.
Indirectly. Direct application of dynamite to the human body is considerably more fraught an event than direct application of penicillin. As the fundamental goal of technology is to extend the capability of the human body, it's natural to implicitly and primarily consider what that extended capability can do to another human body.
It's nice that we have tunnels through mountains and bedrock.
Dynamite is ~20-60% nitro glycerine and the rest "dope" aka a stabilizer which could be sawdust for instance [https://en.wikipedia.org/wiki/Dynamite].
If you have heart issues, you're likely taking nitro daily for chest pain.
So saying it's bad is 40% wrong at least. Makes ya think.
You don't get to unbundle the technology that was achieved by bundling and call it okay because the components are benign.
The best you can say is that dynamite was a safer alternative to preceding technology, and that its danger only comes under certain circumstances. None of that takes away from the fact that dynamite is quite dangerous when those circumstances arise, which is why it's commonly (and correctly) viewed as such.
But also, going back to the original contention - whether dynamite was neutral or positive for humanity - the destruction it and its descendants wrought in war is maybe more than counterbalanced by advancements in infrastructure. That said, if the industrialization and globalization it enabled leads to biosphere-destroying climate change, I would lean towards neutral.
Now think what ingesting a smartphone would do to your body. That's a very silly metric
I didn't say anything about ingesting.
Has is been a net positive?
there are a bunch of ways to look at this (dynamite as a specific kind of explosive, dynamite as a stand in for explosives in general, dynamite in the context of what else we would use for similar purposes if dynamite specifically wasn't invented)
and most of them come out on top. It's main innovation is that its a more stable explosive, much safer to use. Without it, I think a lot more people would have died in mining and construction accidents. It's not typically the kind of thing used for warfare, but im sure it has been for some (but would they just use something else?)
certainly, it has enabled the buildout of cities and infrastructure that were previously impossible to build.
I believe in AGI to the extent that if what's going on between your ears isn't happening on a network of neurons, it's magic. And I also believe the idea that AGI is near is based on the emergent capabilities of LLMs. There's a chance that AGI will emerge from bigger faster better LLMs. But without a theory of when and how that will happen, I'm not counting on it.
It's not a tooling problem. It's more of a layer of policy problems. If you realize that you cannot run a simple experimental code even in non-production environment for weeks due to 10s of privacy, security, access, process and legal issues where you gotta collect a bunch of approvals, this is critical. And the problem gets worse because the tooling is too good when it enforces. There used to be some holes and circumvention which are all gone these days. This is probably why they said "the infra is good for services but not for research".
I don't even think it's good for services. It's not like you go through cumbersome reviews/tools and then things are safe. They have insane homemade config languages and obscure systems that 99% of SWEs don't really understand but won't say it out loud. That's how they dropped cns2, and the postmortem is never going to blame the tools.
You have no clue what you're talking about.
The tooling is not the problem. If shit takes forever to launch, it's because there are many stakeholders that need to be satisfied (some for security, some for regulatory, some for the kinds of politics you get in a company that employs almost half a million people.)
But GDM isn't gated on launches. They were freely releasing things internally for dogfood. Problem is that stuff was just not as good as the competition.
A particular Google product being shit is a data point, but is orthogonal to my opinion about why it is slow to launch products/features.
Internal tooling? Didn't Jeff Dean write their internal tooling?
He wrote the good old parts
And Tensorflow?
That was supposedly 17 years ago, so I was counting it in the good old. It was cutting-edge at the time, then years later PyTorch ate its lunch, which they eventually admitted with TF 2.0.
Jokes aside, I think a lot of the famous Google internals that became public (Tensorflow, Kubernetes, Bazel, Angular), although I heard everyone say they worked so much better inside Google than outside it, had issues. And the Facebook-supported rivals were often just so much more pleasant to work with that you couldn't ignore it. For all else that was bad about Facebook, for a few years they were pretty good at denying Google technical hegemony.
Definitely a thing with Angular and TF. Blaze works well within Google's monorepo for sure, but idk what it's like using Bazel outside. Never used Kubernetes in Google.
Tangent nit, but k8s was never a google-internal system. See https://research.google/pubs/borg-omega-and-kubernetes/
If the tools are all in the same monorepo idk if that's actually true
What do you mean about them being in the monorepo?
> because even Gemini doesn't understand Google-internal tooling.
this is false, it's very good at internal tooling.
Only the GFG models know anything internal. Regular Gemini isn't trained on any of that. And GFG is a much older base model, so people use the regular one. If the tools seem to handle google3 code ok, it's only because of skills and not the model itself, and then you run into issues with skill bloat. Sometimes the A/B test would give me the bad model of the day that'd try to grep all of piper.
Start in a blank directory and tell it to spin up a boq Scaffolding stubby server that responds with "hello world." Unless something has changed after I quit a few months ago, it won't know how to do that locally, let alone actually deploy it. Try the same outside Google with like a Flask server on AWS or GCP.
Your information is indeed out of date.
Dunno about the Boq stuff but it regularly tries to run `git status` in a fig workspace ...
;_;
No longer the case
Maybe they are forced to use Google search.
No, they have an actually good internal search. And there are some good things like stubby, but again pretty annoying that Gemini doesn't understand stubby.
Thank you!
Google’s tooling was, hands down, the worst I have ever encountered. I did 10 years at GOOG, 3 at AMZN, 4 in research, and another 5 at companies you have heard of but wouldn’t be impressed by, and every day GOOG infuriated me.
You don't measure tooling quality with devs' enjoyment, though, but with what the tools make possible.
Technical merit is not correlated to popularity, after all.
Made it possible for a cronjob to take 7 days' wall time to set up
Uh isn’t google known to have the best tooling in the world
They earned that reputation in like 2005. Some people have been there so long (without doing side projects) that they don't know what non-Google tooling looks like in this decade or even previous.
Maybe Google's decision to waste everyone's time and pollute the truth ecosystem by pushing half baked AI assertions into search was a mistake.
I quite like the ai/search thing - I like that there is a super accessible ai I can ask questions without infecting my ChatGPT history
The answers it gives are often trash though, even more so the whatever lobotimized model they slap on top of google search.
I quite like it too. Convenience I guess.
Why not just use temporary chats for ChatGPT?
> And all the prominent names Google gained: NULL
As an outsider, Google seems to have a knack for minting prominence for their talent.
Did all these names were huge before joining Google, and Google used their extraordinary hiring skills to get them? Or a whole bunch of them had challenging problems to solve and ample resources at their disposal to become huge?
To me it is good thing in either case. Extraordinary people are leaving to make even faster research and development. A lot of talent in Google hitherto unnamed is going to get chance to shine.
Sundar Pichai began the downfall of Google by most accounts. It went from an engineer driven culture to one of infighting, politics, and bureaucracy. It's telling when xooglers nowadays complain of maintenance and improvement of products being a career dead end vs. shipping new things. And there being so much empire building.
Some of these ex-Googlers apparently are starting a new company. Any news around that? Curious what they're up to ...
Gemini has been improving by what feels like 1% every week
68% per year is pretty good!
> Combined with no gemini frontier GA release in about 14 months. You have to have created an environment pretty hostile to innovation for this to happen
The Trump admin is now regulating frontier models though.
"prominent names", lmao
All this talent has delivered... Gemini.
Oh well, emperors, clothes, ... you know.
Gemini is arguably the fulfilment of what AskJeeves promised and never delivered, nor did the rest of Silicon Valley succeed in natural language search and question answering for the 30 years it took to finally arrive. Gemini works great for answering questions and delivering answers for probably 90%+ of the things that people are going to ask Google for, while running on Google’s TPUs paid for with Google’s profits instead of hyper expensive Nvidia racks funded with VC Hopium, and integrated seamlessly free of sign ups to novel portals.
Ever since Google became an adCompany, it rarely innovated anymore.
So it hasn't innovated since Oct of 2000? (when AdWords launched)
The ideological capture of Google and the gatekeeping around the ring of effective altruism AI power around there probably led to the demoralization of the technologists.
I think it was pretty clear that their AI efforts were in trouble when their AI Image generator would only make an African American George Washington, and that dude in AI research was demanding that the much earlier generation AI was conscious and had should have rights. At least at Grok, Elon said he wanted an AI dedicated to the truth which is an easier target to hit than what the Google gatekeepers probably wanted. The ideological purity of a Google AI must have been a real moving target though, and retraining an AI is not an cheap or fast thing.
Hard to explain the departure of Timnit Gebru if that's the case. Unless you think they actually changed corporate culture so dramatically because of that event...
I recommend touching grass.
It seems like the real news is Jeff and Sanjay are leaving Google, and Demis is effectively replacing Jeff as Chief Scientist for all of Alphabet.
The bigger deal is the departure of Jeff and Sanjay, rather than Demis moving into a different role.
It seems there's a big shake up on the underperforming Gemini side. Before there was Shazeer (already gone) and Vinyals as co-leads, reporting to Hassabis, now Gemini comes under Kavukcuoglu reporting direct to Pichai as SVP of DeepMind.
Hassabis seems to have been pushed aside. He had been CEO of DeepMind, but that position no longer exists and it seems Kavukcuoglu is now leading DeepMind with a title of SVP. Hassabis is now just "Chair" of DeepMind, and has been given the newly created title of Alphabet Chief Scientist. Are these just face-saving titles, or does he still have any real influence over Google/DeepMind's pursuit of AGI?
Shane Legg remains as DeepMind "Chief AGI Scientist", but I wonder if the DeepMind founding mission of creating AGI is really intact, or if he will be next to go. Has DeepMind just become the Gemini division?
I just found out that David Silver, DeepMind's RL-expert, already left in february, to create a startup "Ineffible Intelligence" focusing on RL-based continual learning.
> Hassabis seems to have been pushed aside.
Yeah he seemed too reasonable to me, relative to the fervor. Whoever ends up in charge needs to do a lot of frothing to catch up to the ferver that would justify their valuations and investments
Probably not going to happen, but I'd love to see Isomorphic Labs separate from Alphabet with Hassabis still as CEO. He's too pure minded to be at company like Google.
I have a feeling this might negatively impact Isomorphic as well since they also fall under Alphabet
Hopefully not - Isomorphic also have external investors (maybe 10-20% ownership?), who may be willing to buy it outright from Google if they lose interest, and Hassabis is now said to be "leaning into" the role of Isomorphic CEO, which is at least pursuing his goal for scientific advance via AI, if only in one specific (important) direction.
https://endpoints.news/demis-hassabis-leaning-into-isomorphi...
He won a nobel prize for his work on protein folding while at Google...it seems like they give him a lot of room to explore problems that aren't directly related to serving ads
But it was Google fault that they really cannot productize all that innovation that happened DeepMind. Google probably now fumbled with world models and this exodus will create next giant in that space.
> Google probably now fumbled with world models [...].
Can you explain this more precisely?
Google's capex is to capture the compute purchases from the other labs. The amount they are spending on Gemini is probably shockingly low...and hence researchers leaving for better watered pastures.
> The amount they are spending on Gemini is probably shockingly low
IDK, haven't Google been putting Gemini front and centre in pretty much all of their products?
I'm seeing Gemini on my slide decks, Gemini on my e-mails, Gemini on my searches, Gemini on my videoconferences, Gemini on my database query console. My impression was they were doing a Google Plus style attempt to marshal all the company's efforts behind one product.
Their capex is primarily in building DCs (and what goes into them), which they fully intend to monetize in order to capture that $500B backlog. In that context, the amount of new capacity that will be given to Deepmind could be surprisingly small, since the ROI of selling pickaxes rather than mining gold is so much better right now.
I’ve heard the pay at deep mind is low and people are complaining about it
How low are we talking?
Nowhere near what other ai labs are paying. Most deep mind employees make what typical Google employees make
Not true, they get paid at least a whole level (or two) above what we make. When I was discussing offers a few years ago, the GDM candidates regularly had six figure differentials even though they were the same level.
Maybe this isn't the same as the eight figure comp they'd get at Meta when they did their hiring spree, but no one thinks that's sustainable.
its still not enough, and the up leveling is a problem because the people getting hired arent qualified to be upleveled
So, how low?
they are paying them like rank and file google employees, aside from a few star researchers
Of course it does matter what other companies are paying, but I wonder how many DeepMind employees actually deserve to be paid more than any other "rank and file" Google employee? What unique skill do they have (esp. relative to what Google are now doing), and what value are they creating?
I don't get the impression that the difference between one company succeeding to build SOTA LLMs, and another struggling, comes down to individual employees - it seems to be more about the organization itself and their ability to manage teams and projects of this type. No doubt there are a few rockstars generating huge value, such as Noam Shazeer had been, but they are the exceptions.
When DeepMind was first created, before Google acquired it, they were famous for the high salaries, especially for the UK, but this was an assemblage of the brightest and best PhDs, expected to be solving challenging research problems along the unknown path to AGI. Many of these original employees may still be there, but it seems their job and value proposition has changed - are they any more capable, or key to, helping Gemini catch up with the competition than some "rank and file" employee familiar with LLMs? And if so, why haven't they done it?
the fate of the company partially rests on whether deep mind does well, you sound like an MBA
>the fate of the company partially rests on whether deep mind does well
I'm honestly not sure about that. They need a decent LLM to fight off the threat of LLMs replacing search, but I'd say that Gemini 3.6 Flash is more than good enough for that, with a smaller/cheaper model being preferred to a larger one. If you are "searching" for a proof to the Jacobian conjecture, then try Fable, and I doubt Google will miss the advertising revenue if Anthropic manage to sell Terrance Tao a pair of socks.
More to the point, DeepMind seems to have become a product division charged with building LLMs, not a blue sky research institute chasing AGI. To the extent that continued LLM improvement is important to Google, the relevant question is how much do you need to pay for a competent ML/LLM developer?
>you sound like an MBA
Well, no - techie here.
I wasn't sure if you were suggesting that DeepMind should pay more just because people developing LLMs at other companies are paid more, or because these are elite DeepMind researchers and are objectively worth more than other Google developers. My point being that it seems they are no longer being used as elite researchers - they are LLM developers. Does Meta need to pay FAANG salaries to employees that have been repurposed as data labellers?
what are you even trying to argue? that the most critical part of the biggest companies on the planet dont need to pay above market?
i know people at deep mind, its impacting their ability to deliver good products
Right, but deepmind won't do well if they can't get TPUs, if the TPU network doesn't work, if their borg jobs aren't coming up fast enough, etc.
this is roughly how i've been reading their approach, but i wonder when/if this changes. what actually makes them start caring about having the most capable model? maybe nothing. i suppose they could be happy in a world where they (a) are still the agents' preferred search index and (b) own or produce a good percentage of the hardware that anthropic and openai run on.
> i suppose they could be happy in a world where they (a) are still the agents' preferred search index and (b) own or produce a good percentage of the hardware that anthropic and openai run on.
If the goal is to maximize share price (which it is) this is probably the safest approach. Why do they have to keep chasing developing the best model when they can
a) charge everyone for the cloud infra
b) have a "good enough" experience for normal consumers
Their current setup will be worth trillions. Already is. Let anthropic get paid for the most expensive queries while they get a bill from Google for their cloud/TPU use. GCP grew 82% YOY with 24B revenue and improving margins. So GCP became a 100B business. Anyone doubts it's gonna double in less than 5 years? Gemini just needs to be good enough for normal folks who want personal agents for everyday use. I don't think Google has to compete with Anthropic on making the best agentic programming model.
Same thing happened to Yan LeCun.
Real scientists are skeptical. Wall St and the people who serve it don’t like that.
> Whoever ends up in charge needs to do a lot of frothing
I don't get the impression from interviews that Kavukcuoglu is that guy - he seems like a safe pair of hands, but not someone that is on a mission.
OTOH I don't even think this is the right race to be in.
Roughly two years ago suddenly Hassabis was heavily promoted on all Google YouTube channels.
It was to fight ChatGPT and promote Gemini.
I think the guy does a very poor job or is simply not the right guy to appear as public figure for Gemini.
At least he tried. He is a man for everything that is not filmed.
Google doesn’t really have a person to give Gemini or AI a human face. And that is only consequential because Google never had any public person with any charisma like Jobs, Zuck, Altman.
Do Zuck or Altman have charisma?
(Jobs I'll give you.)
CEO of <thing> at Google (not Alphabet) was always an informal title. There is no CEO of Cloud, CEO of YouTube, or CEO of DeepMind within Google internally -- it's always been an SVP role.
TK has the formal title of CEO of Cloud.
CEO of <thing> is a layer above SVP.
Google has many layers of management.
Kamangar, Wojcicki, Mohan were all CEOs of YouTube
?? Of course there's a CEO of YouTube.
Inevitably, this will lead to Sundar Pichai replacement. I see nothing less. The sooner the better. I can't even tell what Google's focus is, is there any even?
Google? Focus? Huh? Focus would be the last word that comes to mind. Honestly, ever.
> and has been given the newly created title of Alphabet Chief Scientist. Are these just face-saving titles, or does he still have any real influence over Google/DeepMind's pursuit of AGI?
Alphabet Chief Scientist doesnt sound like a demotion / lack of influence to me but who knows. We're all just speculating here.
I guess his big bet on world models didn’t pay off quickly enough
I think it's more than that. I made prediction in earlier 2024 that the main players of AI will stick with transformers while second class players will want to transcend it. The difference is admittedly a bit subtle but ai researchers would get it. I wrote it with mamba in mind back then, but google was still trying to come up with the 'next transformers' and one that can remember using weights and all that stuffs. You can say the same abt lecun's and ilya's now.
My main reasoning was that transformers was the lightning in a bottle and the best work is in extending it instead of transcending it, which requires you to capture another lightning . Which to me appears to miss the assignment. OpenAI, Antrophic, they understand this intimately. Google on the other hand, fell victim to their own ambition.
This of course depends on what your goal is.
If your goal is purely commercial, or time critical, then a product-based approach of squeezing all the juice out of LLMs makes sense.
If your goal is truly human-level AGI then this is more of an open-ended research endeavor, and timelines are hard to predict. Arguably we have only "captured lightning in a bottle" once in the last decade - the original 2017 attention paper - and so the timeline for a "few more Transformer-level breakthroughs" might more realistically be estimated in decades rather than years. You could argue that the application of RL to LLMs as a training method was a second "lightning in a bottle" but I don't think it changes the expected timeline of such discoveries by much.
The time criticality seems to have become a huge factor for those pursuing LLMs, and certainly for OpenAI and Anthropic, who regard it as a race.
It seems absurdly obvious (though many would disagree!) that LLMs alone are not going to achieve human-level intelligence and cognitive performance (using a slightly broader term there to include things like creativity, for those that might not consider that as part of intelligence).
If you compare a Transformer to a brain, then the best parallel is that a Transformer is functionally similar - in being a prediction engine - to part of our cortex, but of course that means ignoring the other half our cortex - the feedback paths that enable continual learning, which in turn supports creativity.
Of course people will probably respond "you don't need flapping wings to fly", but if you want to fly you do need SOME way of doing it, so brain comparisons are still valuable... If you look at our brain architecture and identify all the components and connections that have no equivalent in a Transformer, and if the goal is human-level capability, then you do need to understand what each of those brain components achieve functionally, and have SOME way of providing that functionality in your LLM+ or whatever you call it. LLMs' lack of any functional equivalent to our cortex's feedback paths - lack of continual learning - has been recognized as one major functional deficit, but there are probably half a dozen others too, reflecting the multiple "Transformer level" breakthoughs that Hassabis notes are needed.
While you don't need flapping wings to fly, you do need to invent the airplane, and even after 100+ years of airplane advances we've yet to build an airplane even remotely as capable as what some birds and insects are able to achieve.
Maybe 2017 was the Wright Bros moment where humans first learnt to do some of what our brains can do.
Yeah, a bit surprised that the top comments aren't discussing losing Dean, he's been a figurehead for the company for decades. It's like Apple losing Ive in a way.
EDIT: "figurehead" - that's all I meant. A notable, public figure from the company who is credited with having a significant impact on its evolution. I'm not making a judgement call on his departure, or Ive's, being good or bad.
That said... just give it a couple years, they'll be back in a lucrative aquihire.
Losing Ive was probably a good thing for Apple. I often feel that after a few years these big guys have done what they could do and somebody else can take it from there. So far even losing Jobs didn’t hurt Apple as it looks.
> Losing Ive was probably a good thing for Apple.
The position that I have rather often read on the internet is: Jonathan Ive did very good work at Apple as long as there was a counterpart who could steer his creative vision. This counterpart was of course Steve Jobs. When Steve Jobs died, there wasn't such a counterpart anymore, so Ive's work for Apple got much worse.
The MacBooks certainly got a lot better and more usable.
Anyone remember the touchpad MBP with no physical escape key and the butterfly keyboard?
(Respect to many of Ive’s great legacy though)
I thought the touch bar was cool and useful, I’ve always remapped caps lock to escape though.
Would like one with all the current physical keys plus a Touch Bar that you could do cool stuff with.
Jony Ive wanted to chop down all the trees at De Anza Community College for something like 11 million dollars for an Apple event. There's a lot of ways he was demanding in the wrong ways for Apple
Man, I'd never heard that story, so I verified it. Turns out it was about two dozen trees and 25 million dollars, just to put up a big tent for an Apple launch event. (Yes, the trees were removed.) What a tool, he just lost my respect.
> Ahead of the event, Ive pushed CEO Tim Cook to remove two dozen trees from the De Anza College campus next to the Flint Center for the Performing Arts to erect an extravagant white tent for the hands-on area.
Such a gifted man… doing such an incredibly dumb thing. https://www.macworld.com/article/696590/apple-expose-jony-iv...
Edit:
- why’d Tim let him?
- why’d the college let them, OK money, but couldn’t they have potted and replanted for just a few or a couple-dozen million more?
- why not have a greater vision and build the extravagant tent to enclose the trees (wouldn’t be the only example of beautiful living indoor trees)?
- why not choose a site that would accommodate without any tree removal?
wtf?
Right? What an absolute cunt (I might be a bit biased because my dad was an arborist).
not true for jeff and sanjay - they kept moving from project to project and delivering transformative results in each.
> So far even losing Jobs didn’t hurt Apple as it looks.
Who knows? Pure speculation? You can also say if Jobs was still around they could have 10x-ed it even further?
Apple car could have been a thing? Apple could have been way ahead and actually competing in AI and data centers? Who knows what else Jobs could have came up with?
Not enough to make an obvious dent perhaps. Nobody says Cook-era the same way they say Ballmer-era.
I have never understood what anyone has ever seen in Ive. He is a complete hack. He's never done anything interesting.
Like Ive? So you mean it’s good for Google?
> It's like Apple losing Ive in a way.
Are you sure we should compare it like this? Not sure it implies what you think it does...
Tell us what we’re missing.
The other commenter said it, quoting "Losing Ive was probably a good thing for Apple."
So the parent comment would imply that losing Dean is a good thing for Google, which is way less likely here.
> as Chief Scientist for all of Alphabet
IS that a promotion or demotion ?
Alphabet is only a holding company. I wonder what kind of authority he has over scientists at Google, Waymo, Deepmind, etc. Probably close to none, so a huge demotion in everything but in title and compensation.
To me the most significant part is that he's no longer in charge of DeepMind, the company he created, other than this "Chair" title which sounds meaningless.
When DeepMind allowed Google to buy them, it obviously had some major immediate positives - access to compute and money - but it seems it should have been obvious that the agreement was too good to be true, that they would be allowed to continue independently on their blue sky research mission to create AGI without any external interference or pressure to create product.
It seems that Hassabis and his DM co-founders eventually realized the mistake and tried to take DM private again starting c.2018, but of course this failed.
https://colossus.com/article/project-mario-demis-hassabis-de...
Now Hassabis has lost control of DeepMind altogether, and it seems to me, as a total outsider, that this is the end of the DeepMind mission to create AGI, at least the Hassabis/Legg definition of AGI as human-level general intelligence, capable of creativity and scientific discovery. Hassabis had always, until very recently, said that he believed it would take a number of additional "Transformer-level" breakthroughs to achieve this type of human-level AGI, while still seeing an LLM as one component of if (which to me seems an admission that the goal has failed - a true human level AGI should be able to learn language, etc, using it's own continual learning mechanisms).
It seems that DeepMind has now fully become the Google Gemini (LLM) division, trying to create a me-too product.
In the early days of DeepMind, before Google, before LLMs, I remember a David Silver slide deck titled "Reward is all you need", referring to RL rewards, which I never agreed with (although Rich Sutton might), but does at least reflect the independent thinking at DM, and of course RL not only gave rise to AlphaGo, but has now become central to the continued improvement of LLMs. However, notably David Silver also left DeepMind earlier this year, to found his own startup focusing on RL-based continual learning, presumably feeling that there was no longer a place for that type of research/pursuit at DeepMind.
Still, LLMs seem to be a destructive enough force on their own that perhaps it should be seen as a positive if research towards more powerful AGI appears to have had a major setback.
As for the "Alphabet Chief Scientist" title, it seems somewhat irrelevant, as least as far as Google's pursuit of true AGI. Hassabis is the face of beneficial AI, having been Knighted and awarded a Nobel Prize for his work, and it would be a horrendous PR move for Alphabet not to at least appear to be treating him with respect, even if in fact this does reflect him being pushed aside.
DeepMind had a generational run as a pure AI research lab. AlphaGo, AlphaZero, protein folding, tensor improvements, weather forecasting, GNoME and so much more. Google leadership saw all this and went “now go generate a multi trillion dollar commercial business and beat OpenAI and Anthropic” and the results were, predictably, failure. Such a shame.
We don't know exactly what Google's leadership was thinking, nor how the AI race will shape up in the coming months and years. Calling it a "failure" seems presumptuous and premature. But supposing your read is correct: what made this outcome predictable, in your view?
How do you avoid being the Xerox to the Microsofts and Apples of AI?
Can you imagine if they'd skipped all of that and put the effort into keeping Search profitable without enshittifying it? Or even just skipped the second part?
> put the effort into keeping Search profitable
The company would have ended up like Stack Overflow? I don't think there was an option for them.
If they focused on keeping search results genuinely high-quality, I think the outcome could have been very different. Basically everyone I know slowly began shifting their questions to something like ChatGPT because Google search results were only presenting irrelevant SEO'd garbage with 12 ads shoved into every page. Slapping the lowest quality model at the top as an "AI Overview" and removing the one genuinely useful part of search at that point (widgets like the dictionary) was the final nail in the coffin for me at least.
What would that even look like? Paid search engines exist and they are niche products.
Prior to this, just about everything on the web was open for crawlers since content wanted to be found.
Google caused this by pushing people towards "brands" rather than individual pages. Those platforms only exposed the data necessary for SEO and Google refused to punish them for it.
Lmao then they’d be completely irrelevant instead of mostly
> Lastly, after an incredible 27-year run, Jeff Dean is at a moment where he wants to try something new, and we’re excited to support him in that. Jeff and Google Senior Fellow Sanjay Ghemawat are launching an independent public benefit corporation to accelerate discoveries in ML, science, and engineering.
Oof. Good for Jeff and Sanjay (who just joined Twitter), bu that is a big loss for Google. Google stock is down 5%. It might not be much of an exaggeration to say these two are worth ~$200 billion.
> Sanjay (who just joined Twitter)
Clarification: this comment is saying Sanjay Ghemawat joined Twitter as a user recently (new account @Sanjay_Ghemawat as of July 2026), as opposed to Sanjay working for Twitter the company.
Such bad wording wow.
I mean in the previous paragraph it says they’re creating a company. And is Twitter even hiring anymore?
It’s called X now, but OP could’ve said ”signed up for”, not ”joined”.
> It might not be much of an exaggeration to say these two are worth ~$200 billion.
It think the vast majority is what the decision to leave says about the overall sentiment at Google, not the value of these two individuals.
For at least a year now the standard reflexive reply to “Google seems way behind OAI and Anthropic” has been “it’ll be ok, they’ve got Dean and Hassabis.” And now they don’t. What reason is there to be bullish about Google now?
I still think Google is the only one who has a shot of coming out of this on top. Open AI and Anthropic's whole existence is predicated on some sort of moat, which I don't really see them having long term. They've got a bit of a headstart, but that's it.
Conversely, AI is just a means to an end for Google - they don't need for their model to be the one to succeed. But, in contrast to the other major company in their position Apple, they do have a model, so they're not totally beholden to another for AI (like Apple is using Gemini!).
But beyond that, for training the model they have YouTube, and of course they have their crawler and index, and the billions of users.
I think HN skews coding agent focused, but that's not really a market for Google. I expect we will have coding specific models in the future, but Google wants a more general intelligence, to handle search queries, be able to connect email to chat to calendar and tasks, and so on. I don't find Gemini that much worse than the other big models for non-coding things.
The leaked internal Google memo "We have no moat" [1] was really prescient. It foresaw all of these proprietary models losing out to open-weight models.
Entirely possible that none of the big incumbents ends up winning, economically.
[1] https://simonwillison.net/2023/May/4/no-moat/
I would like a reassessment of that. This was 2023. I imagine the only ones filling the moat is Chinese labs, but there are other significant obststacles for even Chinese models. Even if they would become as good as their American counterparts, the distribution and inference is mostly in the ballpark of Anthropic and OpenAI. But I don’t know anything, so I would love to hear other opinions.
If anything, it's becoming more true. In the last month you already have Kimi K3 reaching competitive frontier performance (in practice, a bit behind Fable/Sol but close enough). Sure, it's a chinese model but the reality is it puts dramatic pricing pressure on the main players. And these models are now getting good enough to actually help accelerate further AI model research which would suggest the gap might close even more in the near future.
I hardly use Google search anymore in favor of ChatGPT. It's surely an existential threat. Google has a lot of layers of defense in Android, Chrome, Gmail/Accounts, and a stranglehold on internet ads. But at the center of that is search and search is being replaced by LLMs.
We are seeing an exponential increase in the share of site traffic to our B2B company originating from ChatGPT and other models. Google Search is cooked.
IMO, this is largely because google made their search next to worthless. They very apparently prioritized making money over returning good relevant results. The decline in quality is so pronounced that I've heard more than a few people talk about using other search engines besides google.
I don't think it's extreme to say that Bing is a better search engine than Google at this point and Bing isn't great.
> I hardly use Google search anymore in favor of ChatGPT.
Same here, and I'm generally very skeptical when it comes to AI. It's just that ChatGPT can provide better/more accurate answers, Google seems to have lost the plot. I still use google search when it comes to appending "wiki/wikipedia" to it, i.e. when I use google search just as a redirect to a wikipedia page.
I read Google's business incentive as "in it to not lose". AI is a new consumer endpoint and Google currently captures a lot of the endpoints.
> I think HN skews coding agent focused
100% this. Coding agents are an interesting test ground, but the people who care about them are a fairly small bubble.
I think they carry some of the overall AI weight because of the "what if you can vibe code your entire business" moonshot, but that's still orders of magnitude away and who knows if today's coding agents will actually be a stepping stone to that. If we ever get there it will likely be with entirely new domain languages.
What about revenue if people aren't using the search engine directly anymore, also consuming content through chats (not made by them), not on the web where ads are displayed?
On the flipside, Apple has realized AI models are quickly becoming a commodity. By not blowing hundreds of billions in speculative capex, they get to focus on the bits that likely have higher ROI: the touchpoints with AI and how it's used.
Agreed:)
Data centers, TPUs, customers, data streams, more money than god, most mature crawler system
Are any of those advantages getting stronger over time? I guess TPUs but Google is selling several gigawatts to Anthropic
Google owns 14% of Anthropic. If Anthropic gets to superintelligence, Google wins as well. Google doesn't really need a frontier lab of its own.
The money is in selling compute not using it. Anyone who thinks otherwise is a mark.
Crawler and Google Scholar and Google Books.
> Data centers, TPUs
Those 2 are definitely NOT Google's strong points. Maybe by means of marketing.
Oracle, Amazon, Microsoft, Equinix and many more are in the data center race.
As for TPUs... Broadcom, Mediatek and all the other partners you hear less about are likely more important. Google just has the flashy media outreach.
As much as everyone wants to pay accomplished celebrities, all of these companies have young nameless geniuses that are about to make one for themselves. A guard passing torch can be opportunity.
Now, whether Google is the right environment to nurture, that’s its own quandary.
>all of these companies have young nameless geniuses that are about to make one for themselves
They all leave to start new companies. Everyone on the Attention is all you need paper is at a startup.
I keep hearing this, but Hassabis did not leave Google/Alphabet. He merely changed from an administrative position to a more technical one.
It just depends on what are the most upvoted comments in HN. If they are bearish, you can be sure that you should be bullish in your investments. Meta was supposed to be broken as a business already, OpenAI and Anthropic would be failures as well.
Maybe they'll replace Hassabis with someone more focussed on beating OAI and Anthropic at chatbots? He always seemed a bit more into science, protein folding, new drugs and the like.
Revenue, data, and stamina.
Couple of things come to mind. One is that Web sites tend to actively fight AI companies' crawlers while actively courting Google's crawler.
Another is that they hold the key Transformer architecture patent. If it is still relevant (which I'm not personally clueful about) and if they start enforcing it, then we may see a reprise of the situation where Microsoft made money for years every time an Android phone was sold. Regardless of that particular patent, it's probably safe to say they'll be better-positioned than anyone else if AI companies start lobbing patent nukes at each other.
A third factor that shouldn't be discounted is that Google has access to warehouses of training data that other companies don't. Google Books alone is an Alexandria-scale archive that the courts forced them to keep to themselves. Those restrictive copyright decisions may turn out to be a blessing in disguise for Google because no one else will have been able to scrape the data.
Google's weakness (well one of) is its total lack of cohesion. If the Google Books team could, they would sell access to that data in a heartbeat to boost their metrics.
Which would be disastrous to the conglomorate as a whole. It's good for Google that they can't.
Whoa. So is Jeff effectively leaving Google to work on this new venture full time? Or is the venture a side project? It sounds like the former. I’m sure he’ll still have internal access as an advisor of sorts. But this feels like a seismic change. Much larger than I initially realized?
it's a new venture that he can then sell back to google, and continue his ongoing loop
you mean x, right? just to fully understand
Today I saw one of the most talented engineers I know and worked with leave DeepMind and now I probably know part of the "why". Dark clouds hovering over Google's AI game.
> Dark clouds hovering over Google's AI game.
It's not really clear what their gameplan is. From the outside, it looks like they're asleep at the wheel. Qwen/Deepseek/Kimi are crushing them from the cheap-and-open side, and they're not remotely competitive with Mythos/Sol or even plain-vanilla Opus on the "premium" tier.
Gemini does actually have its uses, but they're very very marginal and niche.
From day one everybody was saying that Google would eventually capture the AI market, but it looks more remote than ever. Maybe Hassabis' personal inclination towards AI-for-science, and physics/chemistry in particular -- as opposed to consumer AI and coding AI -- has hurt them commercially.
Is Gemini really doomed? I'm still bullish on Google: 1) they have more free cash flow and capital than God due to the ads business 2) they have data - intent from web searches, youtube videos, google books and music 3) they have dedicated inference hardware
for all these reasons, is being 6 months behind the frontier actually a structural, long term disadvantage? some day the pace of improvement will slow, and google will vacuum up the market. they'll be able to compete with open-weight models just on pure cost advantage from their vertical integration
That also own one of the 2 major mobile operating systems, with Gemini tightly integrated and all of the data they can gather from that. Why do you think OpenAI wants to do hardware? Owning delivery is going to be important, and right now, Google and Apple own the delivery mechanisms (to consumers).
They may not capture enterprise use, but I don't think they have to. That's only one piece of the market. AI that's useful to consumers will still get delivered via a smartphone, and Google is in a great place to capture that.
I also don't think LLMs have to be a "winner takes all" situation. Value isn't going to come from having direct access to a chatbot or selling API inference, value is going to be in the form of a specific product (for most, devs aside here). Something a consumer, or a non-tech business can buy off the shelf and plug and play. A "ready made" customer service agent system, a "ready made" BI platform using AI, etc.
For consumers, that's probably going to look like whatever is bundled and tightly integrated into their mobile OS of choice.
My prediction is that the EU will eventually bring in legislation that will force platform providers to provide pluggable APIs so that you can use whatever LLM you want.
"Doomed" no, but it's pretty clear that they just had a bad cycle and are struggling to keep up with the frontier.
Whether this happened because they bet on "world models -> better reasoning" and that bet didn't pay off, or failed a frontier run for technical reasons like OpenAI did with 4.5, or something else went down? We don't know.
Will they bleed talent, fall further behind until they give up, or clean the organizational and infrastructural cobwebs and get back in the saddle? We don't know.
> struggling to keep up with the frontier
I think the question is whether that's even relevant.
If AI becomes a commodity (will it?) you're better off being Google than OpenAI.
Microsoft struggled to keep up with the mobile industry frontier and here they are, healthier than ever.
I don't know if they are doomed but https://isaiprofitable.com/ seems concerning about Alphabet, Amazon, Meta, Microsoft.
So long as search revenue isn't correlated to AI revenue or threatened by AI revenue, Google at any point can bail on AI spending and suddenly the free cash flow machine is back on. Yeah they have a far larger debt load they now have to service but cash flow from search revenue is so strong it wouldn't be much of a blip.
> is being 6 months behind the frontier actually a structural, long term disadvantage?
Yeah, this is one of the things I find so weird on the discourse. If you get there negligeably later, but without astonishing spend and waste, you might even be better off in the long term.
I have thought that for a while and assumed that was apples approach to AI. Wait until everyone burns through investor cash, invents the better tech, and can monetize. Then copy that business model and polish it, or buy out the competition and polish. That’s usually apples move and it makes sense for google to do something similar.
It depends entirely on whether they're trying to and failing vs. being strategic. It smells a lot more like the former to me.
I don’t think it matters that much for Google. They need to not fall hopelessly behind, but I don’t think there’s a strong economic reason for Google to burn the kind of capex that the frontier labs are burning. Strategically, I think they’re probably doing better than OpenAI and Anthropic. The Gemini models are open, and they are what researchers are working with (see neuronpedia as an example). Over time, this will give them a strategic advantage for the same reasons that open source wins over proprietary. Meanwhile, OpenAI and Anthropic have massive capex that needs to be returned to investors while their margins are being undercut by Kimi/Deepseek/Qwen. Google can wait around for the coming frontier lab profitability crisis and cruise right on by with their Apple contract and owned data centers to pick up the pieces and exceed the existing frontier labs.
Google, of all companies, much vaunted (as in your comment!) for its huge infrastructure footprint, is renting compute from SpaceX to the tune of almost a billion a month: https://techcrunch.com/2026/06/05/google-will-pay-spacex-920...
This is in addition to bumping their CapEx spend to the extent their cash flow turned negative for the first time ever this quarter: https://arstechnica.com/google/2026/07/google-just-had-its-f...
The world doesn’t realize how desperately compute-crunched hyperscalers are to meet AI demand.
This is a better problem to have than SpaceX, which is renting out capacity obviously because it’s own AI products aren’t selling.
A deal with a company Google has a share of, announced a week before their IPO, with very non-committal terms and ramp period protections delivered in one large block on short term notice priced likely at the high end of what Google charges for A4X instances anyway.
I don’t think this reflects desperation as much as strategy.
I dunno if it's a sound strategy that involves repeatedly telling investors [1] and employees [2] over multiple quarters that you are desperate for compute, including leaving a triple-digit billion backlog on the table [3], and then spending so much on CapEx that you have your first negative cash flow quarter ever and taking the inevitable hit to the stock [4], while turning away a large paying customer (who also happen to be a competitor) [5] ;-)
[1] https://www.mindstudio.ai/blog/sundar-pichai-google-compute-...
[2] https://www.cnbc.com/2025/11/21/google-must-double-ai-servin...
[3] https://www.bloomberg.com/news/articles/2026-07-22/google-sa...
[4] https://arstechnica.com/google/2026/07/google-just-had-its-f...
[5] https://thenextweb.com/news/google-caps-meta-gemini-compute-...
Backlog meaning RPO over 5 years. It’s not as if they would be able to collect 250B today from OpenAI and Anthropic if they were to have that compute.
In any case, my point is that the SpaceX deal specifically likely has ulterior motives.
The RPO can be anywhere from 3 - 6 years, sure, but even on an annual basis that’s like a hundred billion now. It was already in the double-digit billions since before AI took off and has only been spiking since then, which tells us 1) it’s been huge for 3+ years, and 2) it’s still growing faster than they can collect it. This matches what all the other hyperscalers are doing.
My point is that an ulterior motive is not necessary to assume when all their actions and statements point to them being severely crunched for compute.
I mean sure, if they had a choice between say, CoreWeave and SpaceX, they’d choose the latter for the nice bump to SpaceX’s financials and their stake… but not just for that, not when it contributes to their cash flow turning negative and their own stock taking a hit.
It is very suspicious when they’re paying for GB300 more than they turn around and charge those out as a4x instances, it was announced a week before IPO and they have a 90 day exit option.
> The RPO can be anywhere from 3 - 6 years, sure, but even on an annual basis that’s like a hundred billion now.
OpenAI and Anthropic deals are mostly 5 year and Anthropic’s starts in 2027, accordingly these RPOs are sized for projected compute needs and run rate in 2027 not today.
The only way either lab could pay 1 year of RPOs today (~80B for anthropic and ~150B for OpenAI) is with a lot more debt or circular financing, the former of which is difficult in this market.
Everyone spending crazy money on capex right now says they have a crushing backlog and need more compute to protect their share price. I highly doubt the demand exists today at current prices if the big 3 hyperscalers magically had an extra 2-3GW of compute. It’s not like Anthropic and OpenAI are turning away customers offering to pay API pricing..
$12B/year is nothing to a company that makes 11x as much in profit. The question is whether that $12B can be turned into more profit.
This move was purely to pump up SpaceX stock price at its current absurd valuation b/c Google owns something like 6% of SpaceX
By your own math, doing this would have required SpaceX stock to go up $200B just to break even, and then Google would have to liquidate it.
What do you mean? Google bought SpaceX shares when it was a tiny startup. They are probably more than 100X on their initial investment. Even with a 99% drop in SpaceX stock, Google would still be positive.
Edit. Google invested 900 million in 2015 for roughly 5% of the company which comes out to 71.5 Billion dollars at 1.45 Trillion dollar current valuation. That's an 80x increase.
I think Larry Page individually might also have a very large stake as well.
It doesn’t matter what price they bought SpaceX stock for. They could have gotten it for free. Any gain prior to Google’s investment could be realized prior to Google giving SpaceX even more money. What matters is the return on investment for the $12B/year into SpaceX.
In order get that back by “pumping” SpaceX stock, SpaceX market cap would have had to increase $200B based on that investment, and then Google would need to sell the stock.
I'm confused. I think you're confusing Gemini and Gemma. Gemini is Google's frontier offering which is closed-weight, API-only, like most frontier models. Gemma is Google's open-weight offering focused on deployment on consumer and edge hardware.
"The company raised its full-year 2026 capex forecast to between $195 billion and $205 billion, with further significant increases planned for 2027." - Alphabet.
I think ~$200B is just for AI infrastructure capex. Fun fact: that's nearly what the 3rd largest military in the world (Russia) is spending on a land war in Europe.
This is against a $500B+ backlog of demand, which is mostly from OAI and Anthropic.
- They're making a lot of money selling Tensor to Anthropic. If Nvidia's $4T market cap is justifiable, Google's position as one of the other top AI chip seller is worth a lot.
- In a world where open source Chinese models decimate Frontier models ability to charge a high price, it's the operators of efficient inference data centers that will win. Like Google
- Google is probably the biggest provider of "free" AI because it's on Google.com. That forces them to focus on cost. And in a commodity market, low-cost providers are the ones that make the money.
> everybody was saying that Google would eventually capture the AI market
my take on this is eventually the money is going to run out and there's going to be acquisitions and consolidation. I think that's when Google will come out on top.
Yes, good chance this all ends with Microsoft absorbing OpenAI and Google absorbing Anthropic, or something along those lines
This is the real reality.
Google doesn't need to compete. 'everyone' is locked into them via the Gapps (mostly Gmail and Maps) and Android ecosystems. Same with Apple, and Microsoft on the B2B side. It's only Anthropic, OpenAI and everyone else that _need_ to compete because switching models is painless. And they have no other revenue streams.
10 years from now, its gonna be Google, Apple and Microsoft left standing in the AI game. Well, until the US wakes up and starts attempting to break the oligopoly like the EU has recently started to do.
I largely agree but I don't know if it's quite so clear cut. From the pricing angle, all competitors except Google, including Chinese models, have incentive to gain market share at all costs, and may be serving tokens at or below cost. I am not sure though, Google could certainly decrease prices if they wanted to.
For agentic work, but especially for web search, 3.6 Flash has an important leg up, its fast speed, that no other model comes close to matching. I guess nobody pays attention to it because it's not the one big flashy number that you compare to other models.
The default model they are using for Web search is getting capable and is very visible. And now it invites people to keep asking questions.
They are quietly trying to become the chatbot that everyone uses to look things up. That strikes me as an intelligent move - not everything has to be done by an expensive frontier model.
They are not "asleep at the wheel"; it's just that the people in charge (the "MBA types") have no clue what to do!
There's an old saying, if you judge a fish's smarts by how well it can ride a bicycle, it will always seem dumb.
The people who have risen to the top of at Google are built for a different environment than what's needed right now. They are good at playing their political games, sabotaging each other, etc.; i.e. all of the petty games that managers play in big companies. But the AI era demands a different skill set: how to bring together incredibly smart people and forge them into a battle group that will achieve victory in the ongoing battle for AGI! It's as if you have built an army of tanks, but the next battle is being fought on the high seas.
Boeing has had smart engineers around the entire time, too.
The group running the company, is the company.
I’m wondering if they just don’t believe that there is a good business model to be made as a frontier AI lab
Edit: hmm, no, they seem to be mentioning AGI and frontier models in https://blog.google/company-news/inside-google/message-ceo/n...
I disagree - I will be surprised if google doesnt win the AI race long term. They have the money, the chips, and the ability to attract talent.
Are they measurably attracting more talent here? It seems like they're losing some of it right now, so is there public info about numbers of researchers they have or etc
Agree that I would (and do) still place my bet on them for the long term. Maybe the outcome will be a couple of good startups seeded and DeepMind _really_ focusing on LLMs now.
Ever heard of the innovator’s dilemma? Don’t underestimate the inertia of large companies and their unwillingness to pivot from their cash cow.
Google specifically has a track record. You could have made the same argument about social media (Google+) a decade ago.
social media is a completely different market though - since there are massive returns to scale, it's incredibly hard for a new entrant to break in.
model training and inference is the opposite. people switch LLMs like people change clothes in the morning. there are popular services (openrouter) that make moving as easy as changing a model string.
Which poses the question: Why does Google even need to catch up? At least currently, the name of the game is integration. The actual model is a commodity.
Demis will get an offer he can’t refuse - they’ll make him the CEO of Alphabet and have the whole company go all in - the question is when.
How is the company not "all in" now? Google is going to spend $200B on AI buildout this year.
Doesn't mean much for DeepMind if that's spent on Google Cloud, to rent out for Anthropic et al.
If the new cow doesn’t have cash, it is better to stick to the cow you know.
Unfortunately the new cow is cannibalizing the old cow, so Google is in a bit of a bind here. (So far I cannot imagine them monetizing AI overviews enough to compensate for the sharp loss of ads on SERPs.)
To its credit Google seems willing to disrupt itself before its competitors can.
One option would be to become vegetarian.
I said that too at the start of the year, but that's a looong time in "AI years". I feel like by now they should have announced a Fable-killer model. They may still do it but looking back I am becoming less convinced now than I was 6 months ago.
They also have the money, the hardware and the talent to be the best cloud infrastructure provider, yet they're still far behind AWS (for good reason, as anyone who's dealt with their customer service will understand).
They’d be hard pressed to lose, if it’s what they want.
They need to route all browser search strings to an LLM, and slowly begin to charge where people will pay. Likely ad space.
> ability to attract talent.
No, the top talent is clearly at Anthropic and OpenAI
I think Google realized that it's more profitable to sell compute to AI labs than to make AI. They are not an ideologically driven company like Anthropic. They very much turned into a conventional company that just wants to protect the bottom line. This was evident even back when they had LAMDA and refused to release it.
why?
It seems that people wanting to focus on AI-for-science instead of topping LLM benchmarks don't have a place anymore in DeepMind.
If cursor can train model, I see no reason Google can't do it. They just have to hire more people and try on multiple angle.
New Jeff Dean fact: when Jeff leaves Google, the stock drops 20 points.
A better one: When Jeff Dean leaves Google, Google’s value drops by $200 billion (the actual market cap drop right after the news).
Watch it recover in the next 3 days. This is a buying oppo if I've seen one. The real losers are the companies that have yet to gain a userbase as large as G.
https://xcancel.com/JeffDean/status/2085034604172603724
> Announcing Discovery Loop!
> I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.
I think their move makes a lot of sense. Frontier models can probably not advance much further with the datasets we have currently available. Most of the text that goes in is online chatter, images and some scientific texts. That's great for chatbots, knowledge retrieval and programming. But with that database genuine discovery is hard to do. I think for the next step in intelligence the models need to have access to much more data: Data from physics, chemistry, biology experiments - and so on. And they need the data in much higher fidelity than you can currently access. If we just feed AI with all the knowledge we've acquired it's much harder for it to become smarter than us - it basically needs it's own eyes, ears, nose and so on.
> Data from physics, chemistry, biology experiments
There are PetaBytes of important scientific data locked in archival file formats. The first step is to make this efficiently readable.
https://www.earthmover.io/blog/virtual-zarr
https://news.ycombinator.com/item?id=46659254
> Frontier models can probably not advance much further with the datasets we have currently available
I’ve been reading this sentiment on HN since GPT4o, yet models got better and better
There’s a difference between things getting incrementally better and a step function.
LLMs will obviously keep getting incrementally better, but in order to get the kind on jump that LLMs themselves were, the sentiment is that we need something more.
Fable is a clear step function for me
That’s because a significant portion of the work and spending going on at frontier labs is generating new, more curated data, in select domains like software engineering and now biology[1].
[1] https://www.synbiobeta.com/read/anthropic-is-hiring-biologis...
I think that was his argument as well. It looks like there is no more data to find, but then it becomes important to find more data - lo and behold we can make more.
Sounds like the oil scare from 90' - we thought we were gonna run out of oil. But as oil gets more expensive it pays to dig further down to find the stuff that didn't make sense to dig up before.
Yes. The web is enormous, but just think of home much information within a company doesn't even make it into the internal knowledgebase, let alone anywhere public.
It's quite interesting that those events are happening at the same moment. I wonder whether they are related
I think there are multiple separate events here, timed to coincide in order to minimise disruption. It's hard to be sure whether they share overlapping causes, but I'd guess that there's at least an element of that.
Yes. In public relations you never want to trickle out bad/concerning the news, you want to "rip off the band-aid and preferably follow it up with some positive/distracting news like 3.5 Pro model release or price drops.
Isn't it "just" a matter of Jeff Dean (& co) leaving, Hassabis replacing Dean, and Kavukcuoglu replacing Hassabis ?
A massive shake up for sure, but why do you see it as separate events merely timed to coincide ?
So there are two events here:
Event 1 - Jeff & co leaving.
Event 2 - Demis stepping down. DeepMind doesn't need a "chair" and "Alphabet's chief scientist" is a bullshit title that Jeff invented for himself when he moved from Google Research to GDM, to make it look like we wasn't abandoning the former for the latter.
The third change is basically ratifying the status quo, since Koray has been de factor running GDM for a while now (and directly reporting to Sundar in addition to reporting to Demis, who in turn also reported to Sundar - quite some triangle there).
Some do-nothing EVP probably just got control over the Gemini org
Gemini had come under Hassabis, being co-led by Shazeer and Vinyals, who are now both gone. Gemini will now come under Koray Kavukcuoglu, reporting direct to Pichai as SVP of DeepMind.
Big shake up for Gemini it seems.
Is it _that_ Jeff Dean?
https://github.com/LRitzdorf/TheJeffDeanFacts
Fun fact: the Jeff Dean Facts was started by Kenton Varda (kentonv around here [1]), as a joke inside Google. He has expressed regret that it eclipsed Sanjay Ghemawat, with whom Jeff Dean basically did pair-programming with.
[1] also known for open-sourcing ProtoBufs v2, creating CapnProto and Sandstorm, and commenting on the recent Cloudflare OS post https://news.ycombinator.com/item?id=49182996
> When Jeff Dean goes on vacation, production services across Google mysteriously stop working within a few days. This is actually true. (TRUE)
Uh oh...
Basically, he had a server that did code indexing running on his workstation. If he didn't login and run a command to refresh his prod credentials every day, the server would lose its ability to talk to prod (as it should) and fail. Much of the company depended on that service. Eventually it was moved to prod.
That's a pretty poor reflection on him if true.
I've noticed that Gemini has been timing out yesterday & today !
Yes, him.
"Once, in early 2002, when the index servers went down, Jeff Dean answered user queries manually for two hours. Evals showed a quality improvement of 5 points."
ahahah golden.
Will be following their journey
Is the subtext that Demis was trying to block what Dean etc wanted to do and so they had to leave Google in order to do it?
Is Jeff also leaving Google?
I am sure VCs are fighting to invest and they will get 4B investment immediately with this team
They are leaving Google, but Google is an investor in the new company. So a split, but not exactly a divorce. Or something.
A "Maria's not an asset to the abbey" situation for you Sound of Music enjoyers.
It seems reminiscent of the Green Project/FirstPerson episode that led to Java https://landley.net/history/mirror/java/javaorigin.html , and I'm sure many another attempt to keep some unhappy star players in a company's orbit using a spinoff.
Except no one was singing "How do you solve a problem like Jeff and Sanjay?" prior to this.
You don't know that for sure.
Ewing Theory, but for trillion dollar corporations.
> I am sure VCs are fighting to invest and they will get 4B investment immediately with this team
This is probably the only time some prominent VCs would be active during August, only for Jeff Dean & Co's company.
I can see them now frantically negotiating, responding and firing off emails right now during their holidays to get an allocation in Jeff's new company.
Would love to hear the pitch: We never had the lead in AI and now definitely lost it, despite the gazillions dollars and engineering resources of Google, but now will be different?
The "lead" in AI, even with google's TPU advantage, may not make economic sense for a company. Can you make back the training investment on a competitive frontier model, before everyone switches to a newer frontier model within a year?
Turns out you can't pay em as much as VC will, haha
its a public benefit corp
Don’t worry they will just limit them to a 10x return but then later on just convert to for profit and fahgetabout the cap!
I don't think Anthropic (the only major PBC in this space) ever did this.
OpenAI did but they had a totally different structure and were never a PBC.
If these guys adopt a similar LTBT+PBC structure to Anthropic it should be more resilient.
OpenAI has a PBC now
So is Anthropic...
Anthropic has stuck pretty closely to their stated intent even if HN does not agree with that intent
Funny how that works, so are Anthropic and OpenAI apparently. Maybe there is something we are missing here!
I think there's one weirdly simple reason DeepMind isn't doing as well as OpenAI and Anthropic. I may be wrong on this.
OpenAI and Anthropic went from tiny startups to huge companies. As a consequence the stock options/RSU's offered to the employees paid off a far higher percentage ROI than any stock options a DeepMind (and thus Google) employee would get (since Google is already huge). This disincentivizes people who truly believe in the economically transformative power of AI to work at Google since their benefits will be capped by Google being large + having public company obligations.
OpenAI and Anthropic have the freedom to do absolutely insane things like negligently hack other companies. It would be stock price suicide if anything even remotely happened with Google.
There's no doubt in my mind that they set up the conditions for their models to escape the sandboxes. "haha oops our incredibly powerful models escaped we need 1 trillion more dollars and really this is yet another reason why no one else should be allowed to build this technology"
Hmm. And the AISI unintended model hacking yesterday was also just marketing for the British Government?
At some point we have to all accept that powerful AI is most likely dangerous AI as well, almost by definition.
Yes.
> As was standard in our cyber testing, we had intentionally permitted internet access, and model- provider cyber classifiers were deliberately disabled - conditions that do not reflect how frontier models are made available to the public.
"No, no, no" a person on Reddit, Hacker News, YouTube screams for the billionth time. "It's all marketing" as the terminator bots kick in the door and slaughter their families.
Cute story. How could they stop this timeline, even if they were true believers?
Crash society to the point where we can no longer make complicated chips. Of course this really isn't a great solution as it involves billions dying too. Not that someone won't try...
We've signed up on the "If you build it, everyone will die" express. No brakes, no stops, full speed ahead.
At least join PauseAI or similar. Or give money to them. Write to politicians. Read Plan A and tell people about it.
They can't, we are all living in a suicide pact of 'Let these guys do whatever they want, consequences be damned.'
It might also be that the security sandbox was haphazardly vibeslopped together.
If Google put out a blog post about Gemini 4 having escaped its sandbox via 0-day exploit to then hack other companies, I unironically believe this would result in a boost to their stock price.
They could really use some encouraging news about the competitiveness of their AI lab.
I wonder if we'll ever know the story of Gemini 3.5 Pro. Is it possible Google saw its potential for hacking, tried to nerf it, and ended up ruining the training run?
They did publicly mention some partners have had preview access. Maybe we’ll find out one day.
Yes. These corporations now run on fascist logic. (Hold up, stay with me here.) The goal isn't to right, or ethical, or principled. It's to (appear to) be strong and capable, and loud about it. You end up with horrible things happening that get cheered because they're only horrible along dimensions that don't matter anymore to the people calling the shots. Same thing for amazing developments that get crickets.
Note that I'm not calling Google "literally Hitler". I'm just saying that traditional assumptions about what's virtuous and what isn't don't apply anymore. For now, at least. If anyone wants to change that, they're going to have to be purposeful about it.
Google's AI was telling people to put elmer's glue on pizza to help the cheese stick [1], Gemini was turning the Founding Fathers black [2], and more. Oh and the pizza recipe was based on a joke from a reddit user named "fucksmith" - part of data that Google apparently paid some $60 million to Reddit to access. The one and only effect of this was lots of amusing posts and articles. Their stock price went up during the whole ordeal.
[1] - https://www.forbes.com/sites/jackkelly/2024/05/31/google-ai-...
[2] - https://www.axios.com/2024/02/23/google-gemini-images-stereo...
“Human … Please die.”
https://www.cbsnews.com/news/google-ai-chatbot-threatening-m...
Would it though? Meta and Microsoft have had very scandalous AI things happen, and their shares didn't tank (or quickly recovered)
> It would be stock price suicide if anything even remotely happened with Google.
aka they're dead in the water.
The answer is far more benign.
They getting a higher ROI renting their TPUs to Anthropic et al instead of performing training and serving their own models. Google cloud has insane backlog, and has rapidly expanded to satisfy it. While those DCs get built, they’re cannibalizing their own products for it.
This makes sense because (1) they are investors in Anthropic, so they still win and (2) they can always catch up on model training later when the profit opportunity shifts, or abandon it if there is no way to recapture that value.
I think the reason Google hasn't prioritized larger models that are more intelligent than everyone else's, even though they probably could, is because they have 4 billion active users already. For example, they send AI Overviews for a large portion of Google searches now.
So I think they have to prioritize scaling for their models to a higher degree than other groups. Being within say 5% or so in most cases is probably adequate and matters more overall for their user base than being the absolute best coder. So they may be setting compute constraints for training or inference that are firmer than other teams.
This is probably true in a smaller way for OpenAI
When you have a lot of free users the business demands that you serve them with the best cheap model you can build
And time spent building that may provide dividends (eg OpenAI has very good RL and reasoning) but it might take resources away from the larger model training
(I have no inside knowledge, so please consider this to all be speculation)
I don’t think so. If Kimi and Deepseek can launch models better than Gemini with much much lesser resources then it is increasingly looking like an organization issue at Google
No I think you misunderstood GP’s comment. The idea (which I personally don’t agree with) was that Google didn’t have to have the best models; it just needed to have the best compute infrastructure, i.e. having TPUs and the software stack to use TPUs. It was a better use of money to develop compute infrastructure than to develop better models. Perhaps Gemini itself was resource-starved because Google liked to rent out TPUs to Anthropic instead. (Second-hand information: I heard that Mythos/Fable were trained on Google TPUs.)
> Perhaps Gemini itself was resource-starved because Google liked to rent out TPUs to Anthropic instead.
This was the core hypothesis.
I agree that other organizations can compete with few resources, but my hypothesis is that Gemini training specifically is being given nearly 0 resources, despite Google obviously having lots of resources. The hypothesis is based on an assumption that Google profits more by selling ALL their compute to others training models instead of using it themselves for training.
They already have good models, so “better” isn’t as profitable.
Google does not have to compete at the frontier, they already own a lot of Anthropic. It's not an "issue" for them because it's not one of their goals.
I don't think Google is that freaking short-sighted. Even if you're not trying to be frontier, the value of having domain knowledge via experience for AI is worth saving internal compute alone. By all accounts Google believes in ai as much as everyone else.
If AI turns out 1/100 as important as they seem to think, it would be insane to intentionally be slack on it.
Awfully ironic that Google makes more money (probably an order or magnitude or so) from it's direct competitors than it does from a it's own product.
There is no other way to describe the Gemini 3.5 pro delay than as a complete and unmitigated disaster.
It's a huge company.
It's unlikely there is any one person to blame (and entirely possible he has none of it). But things need to change.
I disagree. Google is one of the few parties that can monetize AI because Google has a massive moat in the form of their products: gmail, chrome, photos, search, etc... and Google already has the custom-built chips and datacenters.
Google spending money on competing head to head with your LLMs is a waste of money for Google. If anthropic wins, Google copies their approach, buys anthropic for cheap or both. All the investors throwing money into OpenAI, Anthropic, are just accidentally subsidizing Google's product development. Google shouldn't spend its AI research capital in an arms race with Anthropic but instead should invest in AI approaches that no one else is investigating at scale. That way Google can hedge against LLMs hitting a wall.
There is a real but small danger to Google that Anthropic replaces Google as a search engine, but that is an uphill fight for Anthropic. Google has massive brand recognition, network effects with gmail and chrome, Anthropic can't just copy what works from Google. On the other hand Google can copy what works for Anthropic. Google would have to play poorly to lose that fight.
Probably the worse case for Google is that software becomes so cheap and easy to create and maintain that all of Google's product offerings become commoditized. Even in that world Google has a lock on infrastructure. Perhaps ASI software creation completely removes that as well? If so we are living in a post-singularity world and probably the stockmarket doesn't exist anymore either.
In retrospect this is very similar to the Apple strategy: don’t invest heavily in doing something that isn’t already a core competency, position for novel uses of the tech but don’t build it per se. Apple probably would have benefitted in the last 4 quarters from hyping a custom model stack but it doesn’t seem that this strategy paid off to even close to 20% ROI while Apple got to hold their cash and organizational focus on their main thing
> don’t invest heavily in doing something that isn’t already a core competency
“Heavily” it’s a high bar at their scale. They spent over $10bn playing with cars.
But paying so little attention to it that it no longer is a core competency? As in, the people most representative of that competency leaving in droves? I doubt that was the plan.
And Gemini seems about as good as a search engine as any of the other LLMs, if you use them casually. And Google has the brand-name recognition.
Right now AI companies are competing to make LLMs better at graduate-school level tasks. They all can already competently tell you what the weather is going to be like tomorrow or when the first Led Zeppelin album was released.
Gemini seems like a better search engine than the alternatives because it has access to Google's massive crawler feeds. I think Google could charge 30 dollars month for Gemini + all the data Google has locked up (Scholar, Books, crawled webpages, Google Groups, Usenet archives + search of your personalized datasets such as calendars, photos, emails, docs, slides, etc...). I'd pay that easily.
Legal will have a shit fit at all of the things that you just mentioned.
And that might be why Google loses, they are competing with companies that are willing to operate outside the law (or at least pay Trump to have the DoJ look the other way).
> If anthropic wins, Google copies their approach, buys anthropic for cheap or both
Google buys Anthropic for cheap? How?
Anthropic has no moat, their models just get distilled and resold. They can maybe get good margins when LLM improvement rate is very high but as it slows down they are looking at commodity margins. Google has the products that makes that commodity more than just cost of electricity + 0.5%.
My case is based on the assumption that Anthropic will not hit RSI or if it does RSI rapidly hits a wall. Faster your growth curve, the faster you eat all the low hanging fruit and s-curve. I could be wrong here, maybe RSI will cause a hard takeoff singularity by 2030 and just keep going, but if that happens the world fundamentally changes.
Why would Google be the only interested buying party?
They wouldn't be the only party, but they are in my opinion the most likely party.
What the people that can monetize it? Microsoft, X, IBM, Salesforce, Discord, Google
Microsoft would likely buy OpenAI due to their investment and integration with OpenAI. Making Anthropic not that valuable.
Google would likely buy Anthropic due to their investment and ties to Anthropic.
X, Salesforce maybe. Not sure Discord has the money. No idea if IBM capable of growing.
All this changes if Anthropic figures out a moat/sticky product that doesn't just depend on having the best LLM. If they pull off https://claude.com/solutions/healthcare then all bets are off.
> But things need to change.
Not all change is good, as proven by Zuckerberg's response after the lackluster Llama 4 release. The radical restructuring appears to have made things worse.
How? Muse Spark 1.1 is a huge step up from Llama 4 & competitive with xAI's Grok 4.5. In another 3 to 4 releases, MSL might very well be challenging Ant & OAI. Moonshot, despite their comparatively limited resources, has already demonstrated that the Big 2 aren't invincible.
Meta did not blow up their entire AI Org to be the 5th or 6th place model
https://artificialanalysis.ai/models/muse-spark
As someone who has worn Meta Glasses for about a year, AI is not Meta's strength
Google was always at the frontier of real research, but has been abysmal at shipping good products (at least since Sundar). The core company is run for margins and interest rates by the business people nowadays.
Gemini 3.1-pro was genuinely SOTA for a few weeks, in fairness
So they are fixing this by letting go of the people who were best at the research side, and therefore will have no problem converting "nothing" into products anymore!
People keep saying the same thing about Google lagging behind and always end up looking rather silly. Google was going to lose search to OpenAI. Then people complained there was too much AI in Google search. Now it's pretty good and par for the course.
Yes it's a huge company.
So they are slower. Then they will surface it across their massive product base and keep generating cash. While having a hand in Anthropic and others via investment anyway.
Google doesn't need to offer you the bleeding edge at startup pace. They're playing a different game. When the bubble pops they will be well positioned really no matter the outcome to continue to capitalize as their competitors implode or get absorbed.
The idea a delay is a "complete and unmitigated disaster" is just laughable. People have been saying this about Google since ChatGPT first invaded the public consciousness. Google will continue to do well, the histrionics of people like you aside.
That's one thing everyone needs to remember. AI is here to stay, but the bubble IS going to pop. This level of spending is unsustainable. Soon the market will readjust and the amount of money we spend on AI will return to sane levels.
Google has multiple cash firehouses, the small AI companies do not.
Apparently Sundar thought it was a problem.
Obviously a delay is a problem.
Did he say a "complete and unmitigated disaster"?
No, because he's not a fool. If he had said that the correct response would have been to question his sanity.
Yeah. I’m guessing they’re behind the Chinese models in capability now
Just today my Pixel failed to do the right thing on "Set an alarm in 15 minutes" thanks to Gemini. This has worked reliably since Google Assistant was introduced.
"Hey Siri, find me gas stations along my route" routinely fails now. That worked great for years. Total clown show.
Huge companies tend to make money from network effects, rent seeking, and lock in. They tend to be horrifically bad when it comes to innovation, especially when it may disrupt exiting departments in the company. Those departments will fight for their life and generally muck things up.
Very few companies have leadership that can prevent this infighting and force teams on directed goals.
> people who truly believe in the economically transformative power of AI
People who truly care about becoming rich*
DeepMind made enormous transformative discoveries, for instance in the world of protein folding. But that will just save human lives, not let CEOs fire their people to grab a larger piece of cake for themselves.
I think something that doesn't get talked about a lot is how bad most large tech companies are at creating new products, in general. Like, if you look at most big tech companies, they have their core offering that got them to be really large and rich, and a few other products that are somewhat successful, and then a really long tail of markets they try to enter and failed at, or projects that were modestly successful but got killed because they weren't game changers (RIP Google Reader). Most of the time when a large company does something new that succeeds, it's via an acquisition of a smaller company (ie, Google with Android or Meta with Instagram and Whatsapp)
It's funny, because I think the company that's going to be best positioned coming out of this bubble is in fact google, because they have the expertise and the capital. But I honestly can't tell you right now what their AI product even is -- I've seen so many things go into the graveyard a few months after its launched that I'm utterly confused what their offering even is at this point.
From what I've heard, DeepMind's pay structure is not tied to Google compensation bands. Which means that top performers and key hires can be and are compensated extremely generously if the need calls for it. Another Alphabet subsidiary, Waymo, was well-known for doing this back in the day.
But yes, even large RSU grants cannot replicate the upside of joining a company that increases in valuation 1000X, but to be fair, that is not a position the vast majority of OpenAI or Anthropic employees find themselves in either.
That was true early but OpenAI/Anthropic have done a ton of hiring over the past couple years at already-huge valuations, as much on the strength of big current base salary + equity, not just future increase speculation.
Everyone knows there isn't a moat.
No one joins OpenAI/Anthropic unless they think these companies will reach superintelligence.
So most people joining believe their equity will 10-100x even from where it is today.
(Coincidentally, the talent that believes we will reach AGI overlaps a lot with the best talent, which has a magnetic effect.)
Google was offering researchers the Bell Labs/Xerox PARC model of comfortable budgets and pay but capped upside. And just like with Bell Labs and Xerox PARC the inventions at Google got productized elsewhere by people chasing the uncapped upside. Google would have been happy to keep LLMs in the research lab forever and never productize any of it. ChatGPT forced their hand.
DeepMind is just a part of Google which does research stuff, and Google is a massive company that does loads of things but mostly sells ads.
OpenAI and Anthropic are in a completely different game, their primary business model is making a product out of advancements that come out of DeepMind and co.
It's not a ML talent problem. You don't need to be a genius deep learning researcher to think "Hey, maybe if we massively throttle and degrade the quality of our model while still charging the same price, that might drive people away" (as happened with Gemini 2.5 Pro, the one model where Google really was SOTA). Google's likely been providing insufficient training compute to DeepMind the same way they've been nickel-and-diming their customers, funneling it all to Search instead because that's where the money comes from.
Is that what's happening? Search has sucked for a while, and I'd just assumed it was partly because they'd been pushing compute to AI-related activities.
This makes a lot of sense to me. Probably doesn't explain everything but I could see it explaining a lot. Could also explain openai and anthropic delaying IPO (among other factors)
Is this not an apples-and-pears comparison? DeepMind is a research lab, not an LLM-pilled money furnace.
One theory I've been entertaining is that whenever GPT-3.5 came out a lot of people were talking about the "bitter lesson" and how scale was all we really needed to get to AGI. No need for any fancy tricks, just release a larger model trained on more data, by the time we released a hypothetical "GPT-5 sized" model we'd have AGI.
Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time. It makes sense when you realize that one of Google's key strengths, besides talent, is that they have an incredible amount of data they can use for training due to being both the world's leading search engine as well as having all that video data from YouTube. Scaling the training data makes more sense to them than it does to Anthropic and OpenAI, who are both relatively data-disadvantaged.
You can kind of see this when you look at the Gemini 3 scorecard when it came out (https://blog.google/products-and-platforms/products/gemini/g...) and notice that while it wasn't as good as Claude And GPT at coding, it scored higher on a bunch of other non-coding benchmarks, and I think the reason why is simply because of Google's data advantage.
If true, I feel even more vindicated for believing that the "scale is all we need" narrative was bullshit.
Other than attention optimizations and other minor changes, the top Chinese models (which are way better than gemini) have basically the same architecture as GPT2. Of course RL is key for agentic workloads, but I'd say it's correct that progress has been mostly scaling models,adding more data and cleaning it better.
Google has a ton of smart researchers trying all sorts of stuff. They didn’t really go all in on any one thing. They hedge their bets.
But they suck at harnesses, developer tooling etc. their internal environment is very complex and an intern on a MacBook with codex can probably move faster than a seasoned deepmind engineer
> Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time.
Google Gemma disproves this
We'll see if anyone gets to cash in on those. All you need is one down round and that gets wiped out. Or if the IPO gets delayed and disappoints then the stock can drop well before the lockouts expire. OpenAI and Anthropic are essentially offering Monopoly money in the hopes that one day you can exchange it for real money.
> OpenAI and Anthropic are essentially offering Monopoly money in the hopes that one day you can exchange it for real money.
I thought employees have already had opportunities to cash out (there's enough funding rounds for that).
(Tho how much you can sell was limited, iirc to double digit millions...)
I think it's just more about market incentive. At it's core, LLMs are bad for google's previous business model, which was to send you to as many sites 'good enough' for what you were looking for and plant Ad land mines along the way, in the search results and in the websites themselves.
The new paradigm is you ask the llm a question, get the answer and cutout the middle man. (yes the answer may or may not be as good as the old google result, but for the sake of the argument lets say it is), Google was in danger of simply getting their arm cut off. so they focused on scaling so they could add LLMs to the search, which they largely have. You can't offer an opus like model on something as big as search (and which is offered for 'free'), so they focused on that model, and the infrastructure to run it, because they cannot afford to lose search.
Meanwhile, they know the power of frontier models, they are working to have the infrastructure to be a huge player in them and I'm sure they will have a frontier capable model, eventually. They are playing a longer game, because they can, and I think it's going to work out very well for them.
I've always thought this as well.
Wouldn't this be a demonstration of the downsides of the anticompetitiveness of monopolies?
What is anticompetitive about it? The comment above is literally claiming they’ve been outcompeted.
Not sure that's the right way to look at it, given that Google's huge head start in capital and talent did not prevent AI competition at all. It's a demonstration of reasonable, non-problematic dynamics between smaller and larger companies. (Of course, there's an implicit risk here, the folks at Cruise probably worked harder and more passionately than Waymo staff too.)
you are right. I guess what I was thinking was that google's bigness was essentially a bad capital allocation strategy, since they were lazy and not motivated by absolute return, but some combination of acceptable risk, politics, personal preferences, etc in a large management team that has seemed...disconnected for quite some time.
They have a structural advantage in cash flow and stability of funding, but stability is also a handicap when disruption is the objective.
It's a notoriously double edged sword. When you create huge absolute return incentives, you get things like the Airtable acquisition, where everyone's sad that you built a $1.2B company because some investor at some point mistakenly thought it was an $11B company.
Yeah, that makes sense. I suppose the flip side of all of this is that neither OpenAI nor Anthropic have made a lot of money relative to their spending. Google is probably just trying to be more prudent. But, AI so far is a money spending contest, more than a money making one.
Just some random tidbit: When I was in high school I bought every computer gaming magazine I could get my hands on and could afford with my allowance.
So when Deepmind first made headlines I recalled an article in the UK magazine, Edge, which had an article about a game called Republic being developed by a team of former Bullfrog employees lead by one Demis Hassabis.
Out of interest, I just checked Internet Archive, and lo and behold I found it [0]
I see Wikipedia [1] also mentions that he was the lead programmer on Them park and worked with Peter Molyneaux at Lionhead during the development of Black & White.
It doesn't add much to the story under discussion, but it makes me think at the time I wanted to be a game programmer but my parents encouraged me to go study engineering instead.
[0]: https://archive.org/details/edge-issue-078-november-1999/pag... [1]: https://en.wikipedia.org/wiki/Demis_Hassabis
It is a bit of a tagent but still--Theme Park was dope. I was too young to grasp the compete with other businesses (stocks and shares?) level but still had a lot of fun building parks.
Instant nostalgia: https://youtu.be/tQJJ_rhHxIk?si=V8pOVwj3gYkw0xKB&t=188
I always assumed Demis' ultimate secret motivation behind his work was the desire to deliver a version of Black and White that actually lived up to the hype
1. Never knew David Silver was there as well 2. They look like Depeche Mode in an alternate universe
Does anyone else feel something’s off about sundar?
I feel google has become evil.
I had my phone and laptop stolen and did not have 2fa on but Google still locked me out.
That’s stupid.
When a powerful group removes "Don't be evil" from their code of conduct it's pretty clear where things are headed.
Reminds me of when that one verse was removed from Psalm 145.
Off in the sense that he shouldn’t be running the company?
Wow. Jeff and Sanjay both departing. Truly end of a golden era.
There is an entire cohort of work-optional very senior engineers for whom one of the last reasons for hanging on was "at least Jeff and Sanjay are around".
Literally, I am ... I don't know what to say
Did Jeff and Sanjay pair write their resignation letter? Wow.
This just shows that the "great minds" of AI in these companies are not able to come up with anything that radically differentiates their AI than the others. Wether it is gemini or openAI or grok - meh they all are similar and at this point one can interchange one with the other , excep that Google sits right at the edge with almost every browser search query hitting the google search engine AKA Gemini on the backend and one can switch easily into chat mode right from the search page. I dare say that the google CEO has done his job and set up the cash printing machine but from the AI masterminds in Google, not a groundbreaking progress has been made.
AlphaFold is interesting. Maybe LLMs converge because the ideal "use of language" converges. But there are other types of model.
I don’t know man I prefer Gemini to everything other than claude I’m really surprised at the late comeback and happy for it
Isn’t alphabet an investor in their new endeavor?
Google will be considered "the new IBM" in 20 years.
> Google will be considered "the new IBM" in 20 years.
They have plenty of great tech for consumers right now: Gmail, Maps, YouTube, Chrome, Android,..
You can perhaps argue that search might suffer because of AI but they've done a great job scaling horizontally into many verticals.
I don't think Google is going anywhere.
The newest of those are 20 years old.
Let’s be honest. Google was amazing back then, now they’re just an ad-business flailing around in other side quests.
GCP/Vertex, Firebase, Spanner, Waymo, Gemini + Nano Banana, Google Photos, TPUs, Pixel, Assistant, Google Fi, Taara...
Most of those are also at least a decade old, and some of them are acquisitions.
It's a lot of stuff, but with the possible exception of Waymo and the TPU, isn't this all going down in history under "also ran"?
Every single one of these hemorrhages money from their ad business
Gmail is going to be Google's System/390 mainframes.
People used to say the same thing about Microsoft with Windows you know...
Once upon a time it would've been unthinkable for Google search to become awful and nearly worthless due, in large part, by Google's discretionary choices.
Related - interns these days don't seem to have any particular preference for using Google for finding information...
Google search is bad because there aren't any websites anymore. It's not really possible for it to be good.
Comments like this make me wonder if we're all inhabiting the same reality.
That's for multiple reasons - many of them directly being Google's fault with intentional choices made towards this end.
Can you explain this behavior a little bit more. Where do they seek information - is it a consumer chat app, social media app’s search functions?
Bing because it's the default on their web browsers they've used in college and work.
But moreso - whatever the default is what they use. They have zero attachment to Google.
Bing shows no signs of marketshare growth: https://gs.statcounter.com/search-engine-market-share
I noticed that website doesn't show the absolute numbers involved in search engine use.
And you ignored the part where I said "they use the default".
^FTFY “interns these days don’t seem to have any particular preference for finding information”
I think it's a bit inflated to label Gmail a "great tech for consumers". Maps certainly, YouTube mostly, Chrome & Android for loyalists, but Gmail hasn't been "great tech" for well over a decade.
Additionally these are all old products - even the company's more recent products such as Gemini feel stale and on unsteady ground.
People have been saying that for over a decade now, and their business is still going.
Their entire business is predicated on search dominance, which is now on much shakier footing. Agreed that those comments were ridiculous pre-ChatGPT, but now they have a genuine challenger.
This take only makes sense if you don’t know what people use Google search for. If you’re looking for your local Ford dealership, or to book a Carnival cruise, or need a disability lawyer, or to refinance your credit card debt, you’re searching for an ad. You’re not using Chat-GPT for that. You’re asking Chat-GPT about a research project. It’s stealing away all the hard to monetize searches and leaving the incredibly lucrative searches.
Your narrative depressed Google’s stock price for years, and bears finally gave up since they kept coming out with huge earnings beats.
I actually use Claude for finding stuff and recommendations. like "Find me a decent shoe with blah blah conditions" this query now goes to LLM Chat instead of google search. I use google search to find the url of something that i know exists like find the website of a company or finding address of a store.
The issue is Claude is losing money serving you that, and while Google makes money serving you the same result. On top of that, you are paying Claude to get that same result.
Unless Claude adds ads, it isnt going to be sustainable for both Anthropic or you - and congrats, you've invented Google Search.
Google and ChatGPT are both running LLMs on free user queries. If Google is doing it more profitably, it’s because they have a much more mature ad business, and/or they are running a cheaper LLM. There’s no fundamental difference in the business model in that specific product line.
Regarding Claude, it’s true they must be losing money on free users since they promised they won’t put ads there.
Speculatively, I think people are overestimating the cost of inference for the consumer free tier chatbots. I suspect it’s effectively a marketing cost. The primary expenses are compute to train the next model, stock compensation, and heavy-duty enterprise inference (which pays for itself).
Eventually they will show ads. Ads are like disease, we can stop for sometime, but they will find a way to infect. Eg. Netflix, chatgpt etc..
Now get the other 99.99% of the (tech illiterate) population to do it...
Same I use models for nearly everything. Google only better at:
- fast, even the shitty AI results are nearly instant.
- some results like actors in a movie, or where to stream are useful
- instead of typing a url
But that’s it. I never use Google outside of that. It’s just ads, and most of the web is just ads or AI slop.
> But that’s it. I never use Google outside of that. It’s just ads, and most of the web is just ads or AI slop.
Funnily enough, their core revenue driver (Google Ads) is very broken too. I'm trying to run some ads, but for a week they haven't been showing due to some invisible combination of flags when the campaign was created. There's no way of knowing they're not showing from the dashboard, it only becomes apparent when you try and preview the ads with one of your search terms.
I know everyone is long Google but that experience seriously makes me question how valuable their ad business will stay in the future.
Yeah, but are you getting the perfectly targeted ads that Google serves you?
People 100% are asking ChatGPT for all those things.
Someone deleted a reply to this comment “And where does ChatGPT get the data for those answers?” they wrote and I think the question is important: LLM chatbots can crawl the web just as well as Google.
The answer is Exa
Nah, just another startup that’ll get acquired or a similar fate. Building the capability is a superior option imho, if you’re at Anthropic or OpenAI scale. Cut out the middleman.
They can't do it, which is why they acquire.
Huh? But only if you give it a tool it can use - which brings you back to Google (or Bing)
Anyone can leverage common crawl and/or run their own crawler, to do so isn’t novel.
https://commoncrawl.org/
Sure, I was just pointing at the fact that the question "And where does ChatGPT get the data for those answers?" has a two-part answer:
1) training data (common crawl as one example web data source), and
2) live data optionally retrieved at runtime.
My comment was about 2), and that part runs via a search engine (Bing?), at least if you look at how ChatGPT does it.
Indeed, but what is the difference between crawl data and model data but decay rate? Models are trained on previous crawl data, but if an LLM provider engages a search engine to get "live data," that data isn't live but previously crawled as well (and perhaps not yet integrated into models as crawl data).
So, why would you use Google as a tool or search target when you can, in some combination, go direct to the website (or whatever the target data endpoint is) yourself as an LLM provider to retrieve the most recent data or rely on your own "hot cache" of that data that was crawled recently but said data is not stale enough warranting a live web crawl to retrieve and present to the user or AI agent? Is this capability to perform retrieval from a data source in real time not similar to an AI agent?
Broadly speaking, I'm just spitballing on the concept of "You must use a search engine for an LLM to return 'live-ish' results" as I think we're directionally headed to where that isn't the case.
Cloudflare: Give any website a WebMCP interface - https://blog.cloudflare.com/webmcp/
You clearly haven't spent much time around young adults. I work with college students a fair amount and they use chatbots for 100% of their questions, including the ones you listed. And much like google got a rich data trove from user's searches, OpenAI is getting an even richer trove from chatbot chats.
I would actually recommend using a LLM over normal search for that sort of thing, because SEO has ruined those high intent phrases. The AI equivalent of SEO is also a problem there, but it's so much less prevalent.
Calling it "Chat-GPT" suggests that you are not familiar with the space.
"google it" has a stain on it. And that's not said in jest. In my social circle, it has grown a boomer-tint. Like it or not; it's just how it is.
I personally like google over asking an AI. Esp. since something like google is a substrate without which an AI would be far more prone to hallucinations.
> You’re not using Chat-GPT for that
And that's where you're dead wrong. Absolutely wrong.
Search is now only 50% of the revenue. But sure keep the myth going.
Cloud business is making big dollar I suppose
That’s over half of their business, but they do have some diversification with YouTube, Play Store, and cloud services.
Some of their newer efforts like Waymo and AI hosting for Apple could turn into large businesses to make up for any lost search revenue.
They also have significant costs to maintain their search revenue, such as $20B a year to Apple alone to make Google the default search engine in Safari.
If search becomes less lucrative, they would presumably pay less for such deals.
So yeah, definitely a time of disruption for Google and they need to keep moving fast on many fronts, but they are still well positioned in multiple markets to be successful.
> AI hosting for Apple
Curious to see how long that lasts - YouTube and Google Maps were fairly mainstay apps for iPhones for a while, and eventually Apple cut the Maps cord to do their own thing. I don’t know if that’s from an existential concern of “we need to eventually move Maps in-house”, but given the Google of it all I wouldn’t be surprised if Apple is already at least planning on how to do the same with AI once the bulk of the usage evens out i.e. let someone else worry about it now while also getting a read on what rolling their own would feasibly require.
That’s tough to say. So far, Apple has been very reluctant to invest in the massive infrastructure you need to run large AI models.
But if they are successful in making on-device models that work well, then you don’t really need as much cloud infrastructure.
If by "shakier footing" you mean accelerating growth in query volume, ad impressions, revenue and earnings, I'll take it.
I disagree
A model serves a different purpose from a web crawler.
I'd prefer to be able to find source material over steering a model I have no control over while managing hallucinated outputs without the ability to fact check.
Just because it sourced some materials using RAG doesn't make its outputs valid, accurate, or factual.
> which is now on much shakier footing
Is it though? As far as I can tell they continue to maintain total domination of web search, and LLMs do not replace search engines, they work on top of them.
used to be. Cloud, Waymo, and some other bets have been paying off.
And yet, their search engine is enshittified beyond all belief.
IBM is also still going, believe it or not. 260k employees, 200 billion market cap, $67 billion revenue.
I'm not sure what they do these days - it's about 20 years since I used an IBM Thinkpad to connect to an IBM pSeries.
Not so much hardware these days, about half of IBM's revenue comes from software (Red Hat related stuff alone probably brings in a ton), and about 20% comes from consulting
It’s a trustworthy name for enterprises. Many executives are old enough that IBM was the Google of their time when they started their careers.
People have been saying that for over a decade now, and their business is still going.
So is IBM.
There was a pretty interesting article in the Wall Street Journal a few weeks back about how IBM is flying under the radar, and doing well by not taking the hype bait.
It noted that 70% of all credit card transactions on the planet go through an IBM system.
Yeah, I should have said growing not going. That's what I meant.
It ought to be extremely easy now to use LLMs to transpile cobol and do the necessary reverse engineering to migrate off mainframes and aix servers. I think IBM will lose many customers.
It ought to be extremely easy now to use LLMs to transpile cobol and do the necessary reverse engineering to migrate off mainframes and aix servers. I think IBM will lose many customers.
The problem is that those COBOL systems have to be absolutely provably correct, for both financial and regulatory reasons. LLMs can't do that. They're designed to be variable.
You can't vibe code a bank transaction system. "Close enough" isn't good enough in some fields. A minor glitch in a video game may result in screen artifacts. A minor glitch in a banking system can crash the economy.
It’s also a solved problem combined with a risk averse industry.
COBOL on mainframes has worked reliably for decades. On the scale of what a bank will spend on operations the mainframe doesn’t really cost much.
Banks are sort of the canonical example of “no one ever got fired for buying ibm”.
Yep, if AI could do it so could out-sourcing it to India. AI is great for "almost" but not great at "exact"
Isn't IBM still going?
IBM had something of a downward blip last quarter, partly because of mainframe cycles. But it's actually done generally well the past few years and pays a pretty good dividend. $10B in net income for 2025 isn't shabby.
LLMs rewriting legacy code in newer languages is going to have an impact.
So is IBM.
I dunno about going strong. $1,000 in Microsoft stock in 1990 would have made you a multi-millionaire today. $1,000 in IBM stock in 1990 would make you a ten-thousandaire today.
apples to oranges comparison. 1990 is early days for MS. 1990 is gorilla days for IBM.
For reference, IBM was founded in 1911. At the time "computer" was a profession, not a machine, and the majority of US households had no electricity or telephone service
IBM's business is still going so I dunno what your point is
Who knows what will happen in 20 years. Right now, it's not the case.
A few points:
1) the simplest explanation whenever a bunch of people leave [x] at a company at the same time, is that [x] is becoming less important to that company moving forward. It's not proof, but you know, everyone is trying really hard to say that's not what's happening here. Sometimes the simplest explanation is correct.
2) rumor is that Sergey Brin, co-founder of Google, got bored during pandemic lockdown and eventually returned to Google in a lower profile role, related to AI. I have to think that he still has a lot of say in what goes on in Google relative to his interests.
3) Yahoo Finance article says Hassabis "has long prioritized research over profits", so his replacement might indicate that Google has decided it is time for this stuff to pay for itself?
The breadcrumbs going back a year or so point to Deepmind wanting to be research heavy, but Google being more interested in hedging AI bets by selling compute. If you ever worked hard on something that you have a lot of conviction about, only to get denied resources for it, it's a pretty big gut punch.
There is another universe where Google is hard AGI forward, freeing up as much compute as possible for Deepmind, turning away OAI and Anthropic, and offering the uncontested premier AI research lab, by probably an order of magnitude or more compute. Gemini 3.5 ultra is limited availability with unreal abilities.
But their balance sheet becomes terrifying, and it's all or nothing that this plays, er pays, out.
Right now though Google is pretty well hedged. Even Chinese models likely just mean more compute sold.
I was really expecting Google to finesse this - not competing head-on with OpenAI and Anthropic to see who can build the largest LLM, but instead being content with an extremely capable Gemini 3.6, good enough to meet their own needs, and putting maximum effort into "more than just an LLM" AGI which would then eclipse OpenAI and Anthropic.
Instead, it seems Pichai has fumbled what they had with DeepMind, and they'll now just be a me-too LLM competitor chasing OpenAI and Anthropic from behind, in a race that will never get to AGI.
This race was never going to get to AGI, at Google or anywhere else, and it appears that Google knows that.
Why do you say that?
At least Hassabis has an understanding that AGI will require more than an LLM, and understands some of what is missing. He has never spoken publicly about any vision of what an AGI architecture would look like, other than requiring some more "Transformer-level" breakthroughs, so it seems hard to say that he would have failed, other than his 2030 projection seeming unrealistic.
My only criticism of his AGI direction was that he has talked about retaining an LLM as a component of that, but it's hard to tell if that is/was just short-term pragmatism, and a product-based path, or if he really believed this was the best direction. On the face of it having a pre-trained LLM at the heart of an attempt to build a human brain (build true AGI) is an admission you have failed, since if you build a powerful enough (human level) learning architecture it would be able to learn language for itself, not need to have it baked-in. If your version of AGI is not capable of learning language, then what else is it incapable of learning? It would certainly reflect sub-human rather than super-human capability.
Google is big enough though to support deeming as a research lab and have separate gemini division.
I found this article more thorough:
https://www.reuters.com/business/google-shakes-up-ai-leaders...
What happened to the "strike team" that Sergey Brin was leading to try and improve Gemini's coding performance?
still waiting on ganpati propagation
Found the Googler.
They tried so hard and got so far..
but in the end...
The first sentence is quite telling
>We’ve got amazing talent, world-class compute and products…
Products are third on the list. Google is an incubator for talent first and foremost. Products are an afterthought
My opinion is that you're reading way too much into this. Talent is obviously first. Without that you have no products worth speaking of.
Hasn't that always been the case with Google? Outside of a few that are mostly good and have stuck around, I've always seen Google has having great tech but being quite bad and making products out of it.
I was about to invest in Google, but your compelling observation has swayed me. Now it's obvious to me they are a dying company.
Guess I'm not taking out a second mortgage on my house anymore to invest in $GOOG.
It just tells you about their intended audience: Investors rather than consumers.
Nowadays, every company—even huge ones—prefers to be seen as a "growth opportunity", so they are trying to play up their ability to create new products which will somehow be so incredible that the line keeps moving up forever.
That said, there is definitely a correlation between companies chasing new products and leaving old ones to become crap.
I'd bet my mortgage that if the first sentence was "We've got amazing products, amazing talent, and world-class compute", you'd be in the comments complaining that they didn't put talent first.
A company is made up of people, and makes products. Products can come and go for many reasons (or any reason) but the company will always have people.
So then it is similarly telling that compute is second on the list? They have more talent than compute? Layoffs incoming?
If these products are the place that talent has brought us, of what use was the talent?
Google's revenue is 10x up in the last 10 years.
This is also how Y Combinator operates.
Founders are first. Ideas are second.
gemma-4 is an incredibly good model, beating several others five time its' size. Hope Jeff's departure won't impact their next-gen products! And best of success for their new ventures!
Google investing in Jeff Dean and Sanjay Ghemawat is a good result for Google. Means they are less likely to go to the competition and easier to bring them back into the fold if they do something next level.
At least judging from the headlines, it seems Google is far behind on AI. Fable/GPT 5.6 and recently Kimi get a lot of attention, solve long-standing math problems and lead in coding. It seems that Google's supposed advantages of very deep pockets, original talent, vast amounts of data and unmatched distribution are really not making that much of a difference. What is going wrong?
Behind, but not clearly "far behind". Next model release from any major player could shift the leaderboard again.
Deepmind always worked on some of the coolest architectures. Following things like AlphaStar, AlphaGo, and more were extremely exciting and felt like the hacker persona of machine learning. I hope Google can take advantage of this awesome team. They've done incredible work.
With the top AI personnel leaving Google, I wonder what will happen to their Gemini. Right now they still have a lead in inference hardware with TPUs but with both Anthropic and OpenAI developing their own chips, I wonder how long until either one will catch up.
2023: Google is doomed. They're not building AI. No wonder the "Attention" authors left. Google just sat on this. Their PMs are steering them into oblivion. OpenAI is winning while Google takes no risks. Innovator's dilemma. Google Search is doomed.
2024: Google is on fire. Sergey coming back helped. Gemini, Veo. They've caught up. OpenAI is doomed.
2025: Google is seriously winning now. Nano Banana!! Google was destined to be the true winner of AI.
2026 H1: Google is slow as hell. Where is Gemini? Google is not launching anything, meanwhile just look at everyone else. Anthropic! And open source. What the hell are they doing over there?
2026 H2: Everyone is leaving Google. Google is doomed. PMs are destroying the company. They don't take risks.
Meanwhile, Microsoft: "Copilot!"
My predictions about everything ever, summarized perfectly. I need an “inverse me ETF”.
It seems Google has to play both offense and defense: competing against frontier labs' models while protecting search and ads. They also have to be mindful of not doing anything that could hurt their own search or ads. That seems harder than a frontier lab just doing offense on both models and search/ads.
Having said that, I think Google's moat is still strong with Cloud, Gmail, YouTube, Android, Chrome, etc.
> Cloud, Gmail, YouTube, Android, Chrome
Cloud and gmail, I agree.
I don't think YouTube can survive if GenAI keeps going like this. Android, same, but on a different timescale and for different reasons. The Play Store (and all other app stores including Apple's) will also face problems from GenAI making apps (it already replaces my need to buy, but I'm weird and a software dev ("but I repeat myself")).
Not sure how big a moat Chrome really is? It's more like a sales funnel than a product itself, I think?
Gmail? Go try Fastmail. Gmail is worse today than it was 15 years ago.
Email addresses are like other contact information: people don't like having to update it (on either side of the link), it is sticky and people will keep sending you messages on the old address long after you stop sending from it, so it is a moat.
I connect to Gmail through Mail.app, not the website; Seems much the same to me now as then.
>Meanwhile, Microsoft: "Copilot!"
The difference is nobody expects anything better from Microsoft. Teams didn't exactly set the bar high.
It’s so heartwarming to see the two pals, Jeff Dean and Sanjay Ghemawat, continue together! Many years back there was an article about how the two created the modern web.
https://www.newyorker.com/magazine/2018/12/10/the-friendship...
Google’s trajectory unfortunately starts to feel similar to when yahoo lost everything in the early dotcom bubble except it owned a huge chunk of Alibaba. Google’s internal decision to stop publishing useful research probably started the slow but sure process of internal decline. I still hope they can recover, given their amazing position in terms of data and infrastructure, but it no longer feels like an obvious or easy task, when only 6 or 7 years ago they felt untouchable.
WoW, so basically all senior members of GDM are now mostly gone? I thought Google was lagging in coding AI, but this suggests bigger issues.
I am not fan of the future of AI but this I am not sure how I feel about Google's (also Amazon recently laid off it's AGI team) AI initiatives blowing up before all the new AI Labs.
I don't get why Google failed here? maybe after a decade someone will write about it candidly.
> don't get why Google failed here?
If I had to guess, they have the wrong kind of bureaucracy for where things are headed - and it is manifesting through talented individuals deciding to leave.
I wonder if this has anything to do with it
https://www.lesswrong.com/posts/iKm2FhpWkuuBojm82/why-i-left...
Doubtful. His destination is both funded by Google and has a compute deal with Google.
End of an era for Alphabet.
I give it a 51% chance of Discovery Loop being acquired by Google
I actually think there's a low probability of that.
The four have so much influence within the company that they could have trivially set this up as part of Alphabet, if they wanted to. They are also ridiculously wealthy.
They are already partially funded by Google.
I feel it's image management and a bet. If they succeed, Google profits. If they don't succeed it doesn't matter, it's still a signal to the market/wallstreet that there's no "bad blood" between them and Google and that they won't be cannibalizing Google's current interest.
Considering Hassabis co-founded Deepmind to pursue AGI, I take this pivot as a tacit understanding on his part that Deepmind isn't on a path towards AGI in the near future. They might at some point, but they cannot be close in his estimation, otherwise why leave now?
Agree, also see it as a admission that agi isn’t close
Wow we just had a post on here a couple weeks ago about Jeff Dean supporting dissident voices at Google.
Demis stepped down? Who cares? Jeff and Sanjay are leaving Google. We are in uncharted territory.
I think Demis has been way more influential in the past decade. Jeff and Sanjay built a lot of Google's foundational software, but that was a long time ago.
Both are very smart.
But Jeff was responsible for a lot more of what is actually used today than Demis.
Demis is responsible for a lot more of the hype though ;)
I think saying Jeff's contributions were a long time ago must represent some kind of lack of understanding of Jeff's recent contributions.
Jeff has still been focused more on infrastructure, and that is just more hidden most of the time.
I would say,if i was forced to pick someone whose vision to follow, it would definitely be Jeff and not Demis, even today.
Agree. Demis is very smart, but obviously less humble and more interested in self-promotion. There have been multiple Deepmind documentaries that mythologize the Demis origin story.
Not sure it's being less humble so much as the story being interesting for the public - chess wizz and game developer out to solve intelligence, makes AlphaGo tha beats humans at go, AlphaFold that gets the nobel prize. With Jeff Dean developing Bigtable and Spanner, no general public know what those are.
Demis is great -- this is not a zero-sum game. Who built what, and the relative importance, is an interesting discussion to have, but it's orthogonal to my point.
Which is: When you think about Google, true, old, "don't be evil", tech excellence Google, you don't think about Demis. You think about Jeff's and Sanjay's geeky, technically uncompromising, faces.
> Google DeepMind: We are building strong momentum: Flash is in high demand, our Cyber model is live, and Gemma models have surpassed 900M+ downloads
Considering these are the best stats they could find, gemini usage+general situation must be really, really bleak.
High demand means nothing. A model being live is nothing to brag about. And gemma downloads also can be from auto CI pipelines etc. Nothing concrete
> Flash is in high demand
I said for a few years to many a downvote on HN, everyone wants AI, nobody wants to pay the true costs, the AI race will turn into a "race to the bottom" that is, who can give you the most compute for the lowest cost, and still remain profitable?
That said though, Flash isn't it. The prices on the latest flash models put Sonnet and Terra to shame.
Does everyone want AI?
Single data point; no, I don't. I preferred the pre-AI world. It makes me sad that we'll never see it again.
Personally, In SWE, i think the industry has made a grave mistake with the agents and we're just one big Catastrophe waiting to happen. I do think that there is very real value when software engineers use these tools as something akin to exoskeletons that allow the human to do more, rather than just fully replacing them. However, I'm finding more and more that companies are slop shops and just attempting to automate all of their software engineering. That will certainly end terribly. I hope we are not cannon fodder.
I think (hope?) we will.
Check out the New Luddite movement [1] [2]
[1] https://www.cnn.com/2025/10/08/business/ai-luddite-movement-...
[2] https://en.wikipedia.org/wiki/Neo-Luddism
Crab in a bucket. You choose your name well.
You could always become Amish and then it all goes away... ;)
For what? There are some things I want AI for because it does it well. There are some things I don't want AI for because it just makes a mess (hallucinations). Maybe the next AI will be different and we will have the conversation again.
Indeed, it kills our planet, our culture and our economy extremely well. Oh yes, and some code monkeys enjoy that it can make computer code on the side.
Other sites beckon.
I guess it should be said, of everyone who wants AI, they don't want to pay the expense for it to the level they want to use it.
No.
But people want what AI does for them. It's a tragedy of the commons situation.
> I said for a few years to many a downvote on HN, everyone wants AI, nobody wants to pay the true costs, the AI race will turn into a "race to the bottom" that is, who can give you the most compute for the lowest cost, and still remain profitable?
I keep seeing this but this line of thinking doesn't make any sense. What does it really mean?
There are expensive models that increase the probability of you doing your task under a lower cost. That means you can't use Gemma for coding your new compiler - it would just be overall costlier.
Heavier models are cheaper at more complicated tasks because they use fewer turns and fewer mistakes.
Cheaper models are more likely to be cheap at less complicated tasks. Like if you just ask Gemma "Hi" it would probably be cheaper than asking Opus.
So what does this statement really mean? People don't want to pay the extra for a more costly model? Why wouldn't you? It reduces your overall cost!
> Why wouldn't you? It reduces your overall cost!
Because real Fable usage starts at $20/month, and has oppressive usage limits even at that (ridiculous) monthly price.
Compared to my $3/month GLM-5.2 subscription, I have never felt like I was leaving capabilities on the table by refusing to cough up $20 for 15 minutes of Fable use per day.
where are you subbing to GLM-5.2? i've been meaning to try it out and for $3 it's a no-brainer to just load it up and give it a shot.
This is the wrong way to look at it. If you have a complicated task , you can solve it for cheaper if you used Fable. It will use fewer turns to achieve the same result.
You can solve it for cheaper if you use GLM but if you are involved in it more, but that defeats the purpose.
The point is that there aren't many complex tasks were fable delivers a significant value increase over cheaper models.
Single prompting a very complex tasks is rare even on frontier models, because it can be done successfully only for specific situations (e.g. you have a very strong verification step the model can iterate on).
Most of my everyday usage is for smaller takes, were you don't really get the benefit of the most expensive models, and my guess is that is the case for the most users
> The point is that there aren't many complex tasks were fable delivers a significant value increase over cheaper models.
Strong disagree on this. Any decently complicated task like a refactor is going to be more likely to be solved by Fable than by Gemma 3B or whatever.
I have personally tried to use Sonnet over Opus for tasks and Sonnet gets things right sometimes and at other times I wish I had just paid higher.
This is the standard pattern I keep seeing and I can have a bet with you that it would stay like this.
It's not cheaper if the price of admission is $20 for the first taste. And it's definitely not cheaper to pay per-token versus using my GLM-5.2 quota.
Again this is a resolution problem. Your tasks are small enough that fit into a nice $3 quota. If you are an enterprise or a power user, the right-sizing argument doesn't work.
I'm talking about API prices - subscription is a different game.
> And gemma downloads also can be from auto CI pipelines etc. Nothing concrete
I have always found NPM download numbers truly suspect. Is no one caching? Are they estimating true number of downloads base on some estimate of cache hits?
Absolute download numbers are a pure vanity metric. Relative download numbers compared to other models tell you a little bit.
> And gemma downloads also can be from auto CI pipelines etc. Nothing concrete
Thank you for adding some clarity to this. When I calculated 900 million downloads divided by 8.3 billion people in the world, I came with a number that made it look like about one person in 10 were downloading this model.
This might also be related:
Why I left Google Deepmind: https://news.ycombinator.com/item?id=49067285
Google is now working together with Palantir: https://cloud.google.com/solutions/palantir
The Gemini 3.5 Pro delay has been catastrophic for Google. I'm really curious what has gone on behind the scenes there.
But missing out right as AI coding agents become genuinely deeply capable and useful is just an immense failure.
Surely the two leading candidates must be (a) the model is just not that good, or (b) it is misaligned in a pretty obvious way that can't be swept under the rug.
Sure, but the question is why either of those things happened, given Google's immense resources and talent.
The fact that OpenAI, Anthropic, SpaceXAI and 3 different Chinese companies were all able to train big models without these issues, yet Google could not, seems shocking.
They're definitely "down", but by no means "out". They're probably back in another "code red" and will need to deliver something leading edge in some area next. My bet is that Google will be the first to crack continuous knowledge cutoff updates to their model.
From https://www.ft.com/content/61d41764-f2f7-4906-a112-ff3073972...
> The moves suggest that DeepMind will be absorbed into Google’s broader business. [...] > They added that while Google DeepMind would not become purely commercial, it was integrating further into Google’s business.
I guess might be quite a change (but writing was on the wall, the Deepmind -> Google Deepmind part was the first step)
Why form a "Public Benefit Corporation"? There must be some kind of access available that a regular for-profit corporation is excluded from. Politics perhaps? AI tells me PBCs can be shielded from shareholder lawsuits. Also "Founders can maintain vision control even as outside venture capital enters the cap table".
Seems like a good way to spend investor dollars without consequences while maintaining control. Maybe a bit cynical but i can't figure out why they'd form a PBC over anything else.
I formed a PBC and worked at a well-known PBC. Personally, I opted for a PBC because I liked that I could balance a specific cause with shareholder benefit.
In most cases, it doesn't really matter. The board + management is still in charge, and they have significant legal leeway regardless of the structure. But there's little additional cost to opt for a PBC, and it does give you more legal defensibility to be truly mission driven. Standard C Corps weren't really intended for mission driven companies (see the shareholder primacy norm).
I think it's popular for AI startups, because many great researchers understand the risks involved, and they don't want what they build to be controlled solely for shareholder benefit.
While non-profits are also an option for a mission driven org, it's harder to raise the large amounts of cash that some AI startups need, and laws around deferred compensation and private inurement (e.g. options-like structures) make employee compensation harder.
The important part is PBCs protect you from a shareholder primacy directive. Eric Ries describes it in his new book, but the example he gives is if the most evil company you know tried to buy out your company you have to do it in a normal "best practices" C corp because it is your fiduciary duty. PBC helps prevent that based on your declared mission statement. If the sale doesn't facilitate your mission then you aren't obligated to sell.
Google's stock dropped 5% right away, 200billion of value
It's weird for an announcement like this to drop in the middle of a trading day. I wonder if it got leaked, forcing the timing.
And? I bought more. This is excellent news.
This over-reaction is why people can't see over a long time horizon.
This is great news for Google as they realize that Sundar is the problem and he will soon leave Google for Demis to be the new CEO of Alphabet (Google) which I am predicting. [0]
[0] https://news.ycombinator.com/item?id=39868160
AI is critically important to Google, but there's a lot more to Google than just having a frontier AI model. Do Demis skills line up with what the whole company needs? It's going to be tough to beat Sundar's 1200% increase in stock price.
I sold out of my position. I can imagine a story where it works out in the long term, but I don't see how this doesn't cause terrible retention problems in the short to medium term. I felt a pull to launch a startup when I heard Jeff Dean was leaving, and I'm a long time big corp employee who hasn't been at Google in over a decade.
Why would they do this extra step if that were the case?
In theory I think the stock should be going up because his tenure has found Google getting left behind.
I would have thought Jeff Dean would never ever leave Google. What on earth is going on?
Who knew the real revenue unlock wouldn’t be based on how much paranoid red-teaming the model underwent to resist users jailbreaking its ‘alignment’, and instead more on whether the model is post-trained to use ‘sed’ and ‘git’? Poor Gemini
Scientist-heavy orgs that want to solve everything in token space may overtook tool use; meanwhile Anthropic has been super focused on MCP, Claude Code etc for over a year
Out of the three main US AI companies' models, Gemini is obviously the less aligned (read: censored). So I really don't know what you're talking about.
Shhhhhhhhhh!!!!!
The more people find out that gemini 3.1 pro on the google AI playground or via API is de-facto uncensored, the more likely google will put up actually working guardrails and end the fun for everyone!
Gemini isn't as heavily aligned as OAI / Ant models ..?
If you just talk to it over API (no web search) the Gemini models are extremely resistant to thinking the user may be living in a universe outside their training data. Try to discuss any news etc and they assume it’s fake or fiction
Yes, Gemini let's me do SARS-CoV-2 evolution research (perfectly safe, should never be blocked but is impossible with OAI/Ant)
What do you mean by scientist-heavy orgs solving everything in token space? I feel like they use them to make or run tools almost exclusively.
I'm saying AI researchers have a bias towards thinking what needs to happen is prompt -> [crunching tokens] -> response rather than prompt -> [orchestrates 5 tools] -> response
In other words 'just add a calculator tool' is not as sexy research-wise as making the model accurately eyeball arithmetic in its chain of thought. Maybe I'm wrong but that seems to be the case
If all the other AI companies are promising AGI by Q4 of next year, what else can you do to satisfy shareholders than also jump on that same bandwagon?
Between Shazeer and these guys
Looks like Gemini's sub-par performance is claiming heads
Bets on Jeff and Sanjay joining anthropic?
My guess is 18-24 months.
Chinese labs are the proof that there is no need of big names, but of the right mindset and agility. It's those last things that Google truly misses, but now they are missing for a long time, and outside the AI divisions too, in almost every department of the company.
Jeff Dean leaving Google should be the headline but big news all around for Google AI nonetheless
It's chuckle-worthy that Axios webiste reads:
> ... independent publicity [sic] benefit corporation in which Google ...
A Freudian slip? https://archive.is/SxFjr.
If you lost your only 2 Senior Fellows, you deserve to be replaced yesterday
Google loves to kick senior execs upstairs.
Whatever happened to Prabhakar Raghavan? Got kicked upstairs and we barely hear from him nowadays.
It's only a matter of time before Demis leaves and joins Anthropic or OpenAI.
As what? CEO? Demis will not take up the MTS role.
The labs are all very interested in bio right now. Demis is working on Isomorphic rn (Google's version of that) but I could imagine a lab tempting him away to work on their equivalent if it had stronger momentum
He is probably itching to get to Anthropic. After all he is one of the early investors too.
> Google loves to kick senior execs upstairs.
The Japanese corporate culture has a name for this: Madogiwa-Zoku (窓際族) or “The tribe by the window”.
https://de.wikipedia.org/wiki/Fenstergucker_(Japan)
He was a protected entity (tambram)
This seems bad for AI safety/risk. Does DeepMind have any checks on model alignment now? What's stopping them from using AI for military/surveillance purposes?
They already quietly agreed to "all lawful use" with the Pentagon, to no real fanfare. Gemini Slaughterbot Edition, coming soon to a DHS facility near you?
> I’ve always believed the No.1 application of AI should be to improve human health. It’s time for AI to prove its unequivocal value to the world, and what better way to demonstrate that than to help finally cure diseases like cancer.
Here is someone spelling out explicitly what we should be discussing. More power to you, Demis Hassabis. Thank you!
It may be why he's moved role. He can focus on that, someone else can work on getting Gemini above Claude on the LLM Arena leaderboards.
Google is a major investor in Discovery Loop, btw.
Makes sense to me... leaving to start their company with Google being an investor.
Clayton Christensen [1] says (paraphrasing) it's a good idea to spin-off (or invest) in a startup that you've a say on, before the ecosystem sprouts a seemingly-non-entity and gently disrupts your market.
Interestingly (diff strategies i guess) Apple tends to acquire (Q.ai) rather than invest/venture (as google in OP).
[1] The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail.
Publicity benefit corporation HAHAHA typo, think they mean public benefit corporation. Anthropic however is DEFINITELY a publicity benefit corporation.
can anyone with corporate background decipher if that is good for demis or not?
It is good for Demis and Google overall.
This is a promotion for Demis and this could be a path for Demis to be CEO of Alphabet in the future in the AI era.
Google already invested in Discovery Loop (Jeff Dean, Orol Vinyals, Quoc Le and Sanjay Ghemawat's company), so what is happening is still a win in investment terms.
The question is about Sundar's future at Google, he is a mobile era CEO at Alphabet and I would hazard a guess he will probably step down in less than 3 years.
I don't see Demis becoming CEO. He's a scientist and researcher. He wouldn't want to be bogged down by minutiae of corporate politics, org structure, government relations, mobile hardware, etc. Chair lets him have authority to explore any path of interest without overhead of operations.
When someone is moved to Chair at their startup it’s usually a gentle way to change leadership.
I’d guess Google’s execs are unhappy with the way they’ve been left behind in LLMs.
The position at Anthropic seemed to be that LLMs were not the way forward.
They may prove to be correct over the long term, but commercially that view is “wrong”.
It’s a divide between a research oriented and commercial point of view.
For DeepMind, this is what happens when you take the money. Very rational to do so at the time, but it does have inevitable long term consequences.
For Google, it’s repeating the same mistake they made before: they should be taking the long view and looking beyond the tech we have today to leapfrog their competitors. But tbh I’m not sure such a behemoth can do that (I have no idea how Steve Jobs managed to do that). They’d have been better to reallocate some of the capex to letting Demis do his thing.
Another way of putting it: this is the opposite of “founder mode.”
> this could be a path for Demis to be CEO in the future in the AI era
His current title is CEO of Google DeepMind. Becoming Chief Scientist of Alphabet seems to be a step away from the path to replacing Sundar, no?
P.S. I think his real interest is Isomorphic, and his new role will offer fewer distractions.
I agree sundar will step down but I’m not sure if this means they’re grooming Demis to takeover. I suspect not, and someone else might take over.
Similar to how Sundar was promoted, isn't it more likely to be a promotion from existing leadership (one of the product leads/SVP).
My impression, inside and out of G, was that Sundar (and Ruth) were about scaling down R&D expenses (as a fraction of revenue), and focusing on exploiting the monopolies.
Perhaps now there could be a shift back to investing in R&D to get fresh monopolies.
> mobile era CEO
He was the guy branding Google an AI-first company back when they invented the transformer.
ty!
You're crazy of you think Demis is CEO of alphabet material. He's an accomplished AI researcher, Google does about 10000 other things than AI.
Demis is CEO of Alphabet material. He would be exceptional in that role, and Google would be wise to put him there
My read: managing bureaucracy is a boring job esp. if you’re financially in a good spot.
Some think Gemini is falling behind in benchmarks so there was a shakeup at the top. I don’t agree with it.
That said, winning this LLM race is difficult given the number of talented people working on it across the world.
"Dean and Google senior fellow Sanjay Ghemawat are starting Discovery Loop, an independent publicity benefit corporation in which Google will be an investor and cloud provider.:
Kinsley Gaffe: A mistake whereby a politician inadvertently says something truthful which they had not meant to reveal.
https://en.wiktionary.org/wiki/Kinsley_gaffe
I don't get what you mean?
To me the biggest news is Jeff and Sanjay leaving Google... but then not really, since Google will be "an investor". I guess that's sorta their retirement plan? And is Discovery Loop actually part of Alphabet? So complicated.
I think he was referring to an obvious typo "publicity benefit corporation"
Google is just an investor, it's not a subsidiary.
the typo was the Axios article, not from Google's own announcement
He's failing up and out of the blast zone. He won't be responsible if Gemini 4 fails.
> It’s time for AI to prove its unequivocal value to the world, and what better way to demonstrate that than to help finally cure diseases like cancer.
What a worthy goal!
Wow. The list of names of people who've left recently - a who's who, but also quite damning.
Think it's time for Google to reboot itself.
It's difficult to imagine a more damning departure than Jeff Dean.
Welp, the market didn't like that. GOOG dropped about 5% on that news.
The biggest win for AI dev efficiency is cutting down what gets loaded into context. Semantically matching tasks to the top tools helps a lot.
Dunno what the longer term effects of this and the other departures will be, but in the world of vibe-finance Alphabet dumped 4.15% of share value as of this minute.
This move was expected when Google shifted focus to Gemini-only.
It's funny seeing the doom and gloom cycles over Google's AI future on HN. Google was basically considered dead business not even a year ago on here. If there is one thing true it's that HN has always had a habit of conflating engineering sentiment with business reality.
I think what you’re missing is the long term trend.
Google’s cloud business looks great. I’ve always thought they should have completely focused on cloud - a company full of excellent engineers would always win there.
But its main income stream is advertising, and AI is an existential threat. Google is far behind Anthropic and OpenAI, so right now their advertising business is “default dead”, so to speak.
The company won’t die, but without a radical change in velocity it will be a shadow of its former self.
Absolute earthquake at Deepmind the last few months...
What do you mean?
If you search "Deepmind departures" you'll see a string of high profile ones. This is also coupled with the numerous 3.5 pro delays (and strongly suspected underperformance when released).
Something was definitely going down internally.
they are not going to get rewarded as much as they will make going to openai or anthropic. Google is not insane to pay tens and hundreds of millions to individuals like Zuck is. Crazy as it seems, Zuck might have been right, they seem to have stepped back into the arena with Muse.
There is no Gemini 3.5 pro despite multiple announcements?
miss me with the AGI nonsense
Retention is all they need
That's..huge. Hassabis has been the most important figure of the last 15 years in AI.
Even in indirect ways. OpenAI itself was founded because Musk got fixated on "stopping" Hassabis.
Same day Google is rolling out layoffs.
Alphabet has almost 200k employees, it would be surprising if they weren't constantly laying off a ton of people (and hiring them).
Woah, Jeff and Sanjay are big losses. The hits keep coming for google ai
So how many people from DeepMind have left now? Not looking too good for Google.
AI hurts ads money. It's a classic Innovators' dillema.
My read is that AGI is a long shoot and nobody expects it in the next 10 years. They will push for it as always. But the cost of AI is exploding and requires a strong hand and a pragmatic approach if you don’t want to go down with it once the bubble bursts.
Thus, efficiency and pragmatic enterprise application is the next focus. And for that, the old guards are not necessarily the best. Google needs new blood for this venture, preferably from China with their practical engineering view. Every big AI lab is on the verge. Steering the right way will help avoid a catastrophe and come out stronger than ever. A weak captain could sand the ship and drag down who knows what with them.
i mean as a european solo developer entrepeneur I have horror stories of google taking down services due to COMPLETELY broken chain of command / services. Basically a complete failure to help any small business at all. More scarily whenever I raised issues I had weird employees coming onto HN to accuse me of being a fibbing.
This was now years ago, but the thing is when I raised this, it should have been triaged immediately. Clearly there were now cracks that needed fixing, people that needed disciplining / firing / and processes corrected.
Yes I know this woudl have been for "small business" but in my mind , this is indicative of a ROOT issues, like a hairline crack in a windscreen, that if not addressed end up with the utter fragmentation of trust both outside ( in smaller clients), INSIDE ( not doing their job, ganging up), etc. Years later the bigger clients start to get affected. I guess what I mean is institutional rot.
Now take into account there are people through the chain of command who are smart enough to realise but effecting change will just reduce their returns or job security ( like with the old tesla hardware setups) , so unless they are completely selfless for the sake of the company due to greatitude or long held stock, its just not worth the grind to try and correct.
The question remains, if google works, it works. And it has, for decades. Its magnificent. Sure I was burned but I can forgive them for that. My question is if there are powerful people who want google to succeed it would only take one good hire to keep an eye on things before they get out of control.
My suspicion is that perhaps that will organically come about with the advent of their own LLM tooling organically starting to do this for them .
So I guess I have no answers, perhaps they will find a way. I was certianly impressed by Gemini.
Sorry for the long possibly misdirected spiel.
Jeff Dean leaving... as is Sir Demis...
What's happening?!
Have they had enough of Google?
Demis isn't leaving, he's taking on a larger role.
I've had enough of Google and I don't even work there.
I guess Demis was about world models and the old talent saw the writing on the wall. Noam was a big one and this is just a continuation of that.
Don’t understand how Sundar is still running things, though.
Real next chapters:
1) Repair search that had been broken by AI initiatives.
2) Include less invasive AI with ads for those that need to be spoon fed.
3) Pretend to work on AGI and data centers in space.
4) Sell shovels and TPUs to the gold diggers.
1) I think about this a lot. The AI overview has already destroyed the need to even look further down the page for a lot of people. And if you look further down the page, there is often a whole page of AI generated blog spam, which is in turn being regurgitated by the AI overview up-top. The AI overview often regurgitates a completely false reddit comment from two days ago as well. That AI generated blog spam is probably reading the AI overview.
Where do real results live?
Hassabis is being promoted to Chair of Deepmind
Who is going to be sitting on him?
Demis should have been promoted to full CEO. This is the reason to sell $GOOG :-(
well, at least they are not leaving to be MTS at Anthropic.
My uninformed hypothesis: Demis was gently pushed out. Gemini, while having made major strides over the last year, continues to lag behind Anthropic and OpenAI.
Lots of talk of us being at the cusp of AGI, but how will we even know when we get there? When AGI develops a religion for itself?
We'll know that we've reached AGI when a key conservative political belief in the US is that AIs are not people and do not deserve rights.
We'll know that we've reached the singularity when there's a mysterious and superintelligent AI entity with technology and motivations that we as humanity do not understand and have no control over. Not sure why we want that, exactly, but that'll be the big sign.
Humans are really terrible at most things, being moral being exhibit A.
I want a singularity yesterday, and I would even go as far as to claim that all people have an ethical obligation to build super intelligent AI systems as fast as possible to reduce the maximum amount of suffering through the universe possible in living things.
The moment that we can get viable non-human leadership I will embrace it with open arms. The yoke of human existence is extraordinarily oppressive. Give me Ghost in the Shell style cyberborgs. We are the demiurge.
Of course, the real red pill is to realize that there's many singularities. Can't believe I'm going to link Orions Arm on HN but its relevant here: https://www.orionsarm.com/page/298
I can't agree until I have some reason to believe that the non-human leadership will be moral in a way that I can accept. Giving up control forever to some unfathomable, alien intelligence is basically the last meaningful choice our species will ever make, so don't be in a hurry to commit.
That's the fun part. We don't. Could have happened 10 years ago without us noticing. I don't expect my skin cells to be able to recognize "me" anymore than I expect we will be able to recognize a superintelligent AGI. If it arrived 10 years ago, then the last 10 years could simply be its PR campaign. We wouldn't even be able to tell if it was successfully achieving its "goals" or not, assuming an AGI even has "goals"
Science fiction.
Here's my bet: Demis will be gone within 6 months. A year at most.
Hugely disappointing and shocking news. Demis seemed like the inevitable successor to Sundar. A move of this magnitude couldn't have been made without consent of Larry and Sergey, so it makes me wonder why from their perspective.
Perhaps an unpopular opinion: Google would greatly benefit from this AI bubble to pop and take down OpenAI and Anthropic.
“Demis seemed like the inevitable successor to Sundar.”
Demis is a ai researcher. No indications of being a polyglot beyond that? Running alphabet means a focus on revenue growth and across more boring products.
I think most people consider AI a threat to Google's business
All of these would happily come back if Sundar left.
Genuinely curious why. I'm not a Googler and don't follow closely.
Can we please stop saying anything about "AGI"? I remember when people would be like "AGI in 3 months" / "AGI in 2024" / "AGI is confirmed in 2025" like stop, you don't know if it's even a thing, let alone if it's coming/imminent.
Google is indeed collapsing.
"Onwards!"
These guys...
So is Gemini doomed or do they have a roster strong enough to carry the torch?
The momentum is not being conserved here.
This doesn’t sound like stepping down this sounds like stepping up within Google.
It usually isn't, Chairman doesn't involve day to day supervision and decision making.
When Eric Schmidt became chairman or Ruth Porat became president, this was more a transition towards less involvement than an increase of impact.
He means the chief scientist part. Before Demis had no influence on cloud, tpu development, day to day dev tools, etc.
Demis took Jeff former title. (Jeff became Chief Scientist when Brain was dissolved into deepmind)
I suspect the SVP/CEO leading Deepmind has a lot more sway over the rest of the company than someone with a title and no reports. :)
(Also the 20y+ of contributions and having built much of the base Google infra probably gave more influence to Jeff than the remaining title)
Related:
The next chapter of our AI momentum
https://news.ycombinator.com/item?id=49184755
Jeff and Sanjay both leaving!
Source: https://x.com/demishassabis/status/2085034334914769203
(https://xcancel.com/demishassabis/status/2085034334914769203)
Evil reshuffles.
Well, damn. Jeff, Sanjay and Demis leaving all at once, it's hard not to read that as someone fucked up pretty bad.
Hassabis isn't leaving
Ah, you're right. Thanks for the correction. I read that other post too fast.
Let's see... they ruined their core search product. What's next after that bold move?
Two thoughts on this.
First, it is incredibly difficult to pivot a large organization because there are too many competing interests, too many fiefdoms people have built and too much organizational inertia. A new company has the advantage of a singularity of purpose. Google has to compete with internal interests about AI disrupting search, ad revenue and so on. This can slow you down and limit the resources you get. It's why companies get disrupted, particularly when they reach monopoly status. Steve jobs said it best [1].
Second, Gogole's path here (IMHO) is to make their own hardware. They've already done this with their TPUs but need to be able to compete with NVidia offerings. NVidia controlling the price, features and, most importantly, who gets to buy them is bad for Google. This is what Google should be pouring billions into.
The beauty of this is that success is easily measurable against metrics like price-per-petaflops, performance-per-Watt and so on. Throw money at some key Nvidia engineers and have them design silicon for you for TSMC to fab.
I still believe that Google is positioned to survive the (IMHO) inevitable AI bubble popping.
[1]: https://www.youtube.com/watch?v=NlBjNmXvqIM&t=3s
The tech world, in particular the Silicon Valley type, is the most dramatic thing.
anybody watching the pace of output from Gemini team has been nothing but disappointing. meanwhile anthropic and openai has surpassed them.
I do wonder if anybody working at Google can give us an insight why the chaos and lack of competitiveness?
Not GDM, but tech island was strangling when I was there. Everyone will probably say different things.
OpenAI has commited 5 felonies.
Anthropic 2. Fines for $1.5B.
There is no word for copyright in Chinese.
These are the companies that are moving forward.
No AGI on the horizon and that ROI horizon approaching fast is going to make Ed Zitron a superstar.
If Demis Hassabis got a Nobel prize for being a Project Manager, is Zitron up for the Nobel on Economy for excellence in economic forecast?
I don't understand any of the praise for Zitron. He is saying very obvious things and gets timelines wrong all the time. What's special about that?
> If Demis Hassabis got a Nobel prize for being a Project Manager
His accomplishments are far beyond project manager. https://en.wikipedia.org/wiki/Demis_Hassabis
Nobel was awarded due to project management work. There are many thousands of scientists with work much much better than what Demis has done.
Ed Zitron has predicted 16 of the last 0 AI bubble bursts...
But more seriously, this has to do with DeepMind falling way behind the frontier in intelligence and cost per task. There's basically no reason to use a Gemini model today.
Tbf, he only has to get it right once.
Being right at the wrong time is indistinguishable from being wrong.
Maybe. Was it wrong to call Madoff a fraudster in the early 2000s?
Now, I don't think this is necessarily a Madoff situation, or even a dotcom one. Ed Zitron is certainly not Markopolos. Especially evidence wise.
But, getting a date wrong doesn't necessarily mean that he's wrong in general.
A stopped clock is also right twice a day.
Public warning: please don't trade on Ed's idiotic analyses. Or if you do, look at his track record so far, and apply the Kelley Criterion to your bet sizing.
Superstar? This Ed Zitron?
"Based on estimates of their burn rate and historic analyses, I hypothesize that OpenAI will collapse in the next 12-24 months unless it raises more funding than in the history of the valley and creates an entirely new form of AI."
-July 29, 2024
https://x.com/edzitron/status/1817955630784917548
And they did raise new funding as he predicted...Since July 2024, OpenAI has raised or secured roughly $170 billion in new capital. Numbers never see before. So yes its the one, and remarkably right so far.
Notice how he didn't say "I predict they will raise new funding", he predicted collapse, with the implicit retort that there's no way they will raise more funding than has ever been done before and invent a whole new AI. He was mocking the supposed things that would need to happen. He strongly assumed collapse.
2 years these charlatans have had their "imminent collapse" predictions repeatedly wrong, meanwhile all valuations, AGI progress in unsolved problems, and frontier leadership have been repeatedly vindicated that we are beginning ASI / the singularity.
Oh, come on. He's not perfect, but compared to CEOs and researchers predicting AGI and complete economic upheaval every 3 months, he's looking very good. He's no more a charlatan than Altman is.
His predictions (while often wrong) are at least somewhat justified and backed up by genuine scoops and original research. Altman goes on podcasts and talks about building Dyson spheres for energy.
You can't imagine OpenAI being a multi-trillion dollar company?
Open weights will drop costs, but distribution matters. OpenAI has that.
Costs will come down. Deep entrenchment will not.
How will costs come down?
Each generation of model is more expensive to train and run than the last.
More efficient hardware? That means throwing out billions of dollars worth of existing hardware and buying billions more in hardware. The cheapest hardware is the stuff they already have because spending $70B in hardware to halve a $2-4B electricity bill just doesn’t make financial sense (and that’s if doubling hardware efficiency can happen before they go under).
If they cut costs by using smaller models, they are then racing to the bottom vs China which seems hard to do (especially with China’s cheap solar energy).
As it stands, the leaked financials prove him correct. Given their losses (not counting restructuring) they are going to need enough funding to buy any of the bottom 300-400 of the F500 companies outright just to keep the lights on for the next year.
To say that’s an outrageous amount isn’t overstating in the slightest.
Given they are already being forced to drop inference prices in an attempt to compete with Chinese providers, I don’t know if billionaire investors will be willing to hand out that much money this time around if they can’t generate enough value to break even despite the hype.
Sundar Pichai: @DemisHassabis is stepping up to become Chair of @GoogleDeepMind & Chief Scientist of Alphabet, in addition to leading @IsomorphicLabs.
https://x.com/sundarpichai/status/2085033425736745093
So it's not stepping down, right?
If it's anything like what happened with others (like Eric Schmidt), Chairman is basically "you're out from day to day stuff but we'll give you a pot of money to say you're still senior and still here, to save face and the stock price, while you give a bunch of talks for the next few months"
So yes fancy title, but basically, someone who can move the stock price is quitting and we're trying to ease the perception of it.
I feel likely Demis wants to focus on research and healthcare with AI. Gemini is under huge pressure to monetize. This seems like a good step.
But he's not just Chairman, he's taking the title of Chief Scientist. It sounds more like moving into an IC role.
The idea that “Chief Scientist” means anything more specific about what he does than “he’s a big shot who was in technology” is misguided.
A chief scientist can be super influential, or a guy who’s on the slow path to retirement but is keeping a paycheck to keep up appearances.
Jeff Dean became Chief Scientist when Demis took over his org.
It's still the typical title as a stepping stone towards something else (often outside).
Chairman is a stepping stone for a retirement/resignation. Apple's Tim Cook. Google's Eric Schmidt. The list goes on.
But Alphabet is just a holding company, what does Chief Scientist of Alphabet even mean?
It means "please, he's still here, don't let the stock plummet"
It still did.
Could've been a lot worse
4% drop is a nothingburger. Stocks drop more than that for overreactions to random news all the time.
Everyone commenting here seems to take this to mean he's leaving Google, when that's not what the article says.
Clearly his influence is going to be less though.
And most likely this is a calculated step to avoid freaking people out, even though he is effectively leaving.
Not at all clear.
To me it looks like he’ll be in a position to actually be unhobble DeepMind - reminder DM was sitting on a ChatGPT product for a whole year before ChatGPT got released (LMChat) and Google didn’t let them release it.
> DM was sitting on a ChatGPT product for a whole year before ChatGPT got released (LMChat) and Google didn’t let them release it.
Heard this multiple time, to me this is pure history rewriting and post-rationalization. OpenAI also had a internal chat app before chatgpt, Microsoft had multiple, a bunch of other players also had internal chatgpt equivalents + many startups built some using OAI API.
The breakthrough of ChatGPT wasnt because OAI were the first to think of that (absolutely obvious and basic) product, it was because they were the first to get a model strong enough to be actually useful to talk to, vs a mere fun novelty.
There is no evidence whatsoever that DM ever had such a model that they decide not to talk about/release.
OAI was the first to throw safety to the wind.
Everyone else was scared to unleashed it on the general public because it was too powerful with too many unknowns (in ~2022, hindsight is different)
Altman didn't give a fuck, first mover was more important to him.
if him leaving unhobbles gdm, it'll be because he has less influence...
What? His influence is going to be larger.
Larger in what sense? He is stepping away from the day to day operation of the company.
He won’t have any authority in the company other than leading and voting in board meetings.
I’m willing to bet a lot of money that the “chairman” role is a contractual obligation and he be will be gone the day it expires.
omg, you are right, he is leaving google!
By the article, it looks like he's actually stepping up.
Your comment says he's stepping up, the HN title says he's stepping down, the article says he's stepping aside. Nobody says if it's a step forward or backward.
We know he's definitely stepping. That's a start.
It's just a jump to the left...
Yeah I actually read this more as them scrambling because Jeff Dean is leaving? Because that is the role he is taking up
There is no way a Chief Scientist (essentially IC) role is more important than leading the whole of DeepMind, 6000 people strong, working on the most important product for the future of Alphabet.
This means Demis wanted a change and to work on other things, it's not Google wanting this.
Dean has pretty big symbolic importance inside of Google regardless of what his role title is/was
Edit: sibling story in front page links to this which explains better in its lede text than what I just wrote: https://www.nytimes.com/2026/08/05/technology/google-researc...
Or Sundar wanted Demis to change to work on other things.
Grok TL;DR cut the corporate bullshit:
- Demis gets kicked upstairs: loses day-to-day power over DeepMind, gets fancy “Chair + Chief Scientist” title so he can do TED-talk AGI destiny stuff and Isomorphic Labs instead of running the actual product machine. - Koray takes the real job: owns the models, research, Gemini, and shipping. Reports straight to Sundar. Operators over visionaries. - Jeff Dean and Sanjay peace out after 27 years to start their own nonprofit-ish thing. Google seeds it so they don’t walk to a rival with all the institutional knowledge. - Official line: “Accelerate AI + shape AGI for humanity.” Actual move: tighten operational control, keep the science-fiction narrative alive for PR/regulators, push commercial models harder. - Full-stack bragging and “950M users / singularity foothills / cure cancer” language is pure corporate theater. The org chart is the only signal that matters. - Google is in a knife fight with OpenAI/Anthropic/xAI/Meta. This is them optimizing for speed and productization while still sounding noble.
I like most of that TLDR, but:
> Full-stack bragging (...) is pure corporate theater.
I think Google is just about as full-stack as possible when it comes to AI. They do design their own ML hardware, they do tons of research, and do everything in between including hardware drivers, ML libraries (JAX), and the physical datacenters.
when you attack the king make sure to kill the king.
Demis tried to be google ceo by pumping out self aggrandizing 'documentaries' and highfalutin interviews where said a whole bunch of nothing.
Oh wow demis and Jeff, that Ethics engineer who left GDM, and reached to both and posted on lesswrong a couple weeks [1] ago did make a difference?
1. https://www.lesswrong.com/posts/iKm2FhpWkuuBojm82/why-i-left...
...you think they've only been planning this for a couple weeks?
hmm not really, and it probably isnt the origin, but im glad to see him vindicated by having both C-level google leadership people who targeted to try and either stop google from doing the pentagon deal, both stepping down the same day.
At least it makes me feel like he reaching out might have made a difference on remembering this people killer robots are bad, and being part of it might not be on the best of their interests to go down on history for
Alex Turner discusses a bit here at this timestamp:
https://www.youtube.com/watch?v=pGlJQHVeKIA&t=14m27s
what was the post?
https://www.lesswrong.com/posts/iKm2FhpWkuuBojm82/why-i-left...