This category of complaint is useless until it is unfolded to contain better economic reasoning.
In the case of a VC-subsidized service like Uber, the subsidization is utilized by the company until a a monopolistic or network effect takes hold and allows for price increases.
LLM economics are very different. If the tokens are being subsidized now, they must stay subsidized until some form of monopoly, network effect, or pure R&D advantage is achieved.
In the case of LLMs, the open weight models are nipping at the heels of the proprietary models, and this may be a fundamental condition. Perhaps subsidizing tokens enhances increases engagement, and thus, equity value of the company training the proprietary models, and they reinvest this value back into the energy costs needed to train them. Perhaps this dynamic gives proprietary models the performance distance they need in order to increase their margins.
Hasn't happened yet. It's not clear that it will happen.
Next time I read this take, I want to see eight to ten more paragraphs of analysis before it feels like a contribution to the discourse.
I've seen people compare the situation we are in now with AI to early days of Uber. Basically "You're excitement is artificially inflated by the fact that a VC just paid half your bill."
That definitely happened with Uber, but I would argue that one key difference between the Uber situation and the AI situation is COST. How much can COGs be reduced via optimization and technology.
In Uber scenario, the cost is labor, there's a hard lower limit where people will find something else to do for work.
In AI scenario, we've already seen the labs make major reductions in cost-per-token. I think it's fairly uncontroversial to say they have more possible cost reduction levers than Uber.
So I don't agree that at some point VC money will run dry and the unit economics for tokens will dramatically change.
The part where this will be fun is where the VCs lobby the US government, via peace board donation or golden iphones, to ban Chinese Open Source AI because otherwise they won't ever make their money back.
Everything you need to know about the Tech industry (or any industry for that matter) can be learned from watching the Simpsons - Season 4 Marge vs. the Monorail
Hasn't the state of the art in genAI tools been making its way to open source/local inside of a year? Between that and pressure from Chinese firms, it would be hard for the current leaders to monopolize and enshittify their products. There isn't really a network effect to lock people in like with social media.
Another tech writer I used to respect who is now posting about how the future is now fantastically wonderful and wide-open because anyone can create an app to do whatever they want.
As I said before, the ones boosting this everywhere to get as many tech folks on the train have either the following:
* Fully invested in those portfolio companies (no disclosures until IPO / acquired)
* Employees / Social Media Influencers at those companies with stock options which they are effectively paid boosters until they reach the vesting period.
* Frequently screaming the loudest and appearing on TV interviews, opinion articles spreading the scam further.
Just like the "Full Self Driving" scam that still requires a human behind the wheel, now we have "AGI" which still requires humans to supervise the AI agent to not make mistakes.
The tech folks don't question the euphoria and fall for it easily because they want to be on the next Google. But they could also be on the next Enron.
This category of complaint is useless until it is unfolded to contain better economic reasoning.
In the case of a VC-subsidized service like Uber, the subsidization is utilized by the company until a a monopolistic or network effect takes hold and allows for price increases.
LLM economics are very different. If the tokens are being subsidized now, they must stay subsidized until some form of monopoly, network effect, or pure R&D advantage is achieved.
In the case of LLMs, the open weight models are nipping at the heels of the proprietary models, and this may be a fundamental condition. Perhaps subsidizing tokens enhances increases engagement, and thus, equity value of the company training the proprietary models, and they reinvest this value back into the energy costs needed to train them. Perhaps this dynamic gives proprietary models the performance distance they need in order to increase their margins.
Hasn't happened yet. It's not clear that it will happen.
Next time I read this take, I want to see eight to ten more paragraphs of analysis before it feels like a contribution to the discourse.
Highly trained but poorly educated.
And "education" is not limited to the classroom. How the world really works isn't always obvious from in front of a computer.
Hence the current initiatives to incorporate world models into LLMs.
Always open to an opinion if its backed up by deep analysis and some well presented facts but this whole blog is just short angry comments.
I've seen people compare the situation we are in now with AI to early days of Uber. Basically "You're excitement is artificially inflated by the fact that a VC just paid half your bill."
That definitely happened with Uber, but I would argue that one key difference between the Uber situation and the AI situation is COST. How much can COGs be reduced via optimization and technology.
In Uber scenario, the cost is labor, there's a hard lower limit where people will find something else to do for work.
In AI scenario, we've already seen the labs make major reductions in cost-per-token. I think it's fairly uncontroversial to say they have more possible cost reduction levers than Uber.
So I don't agree that at some point VC money will run dry and the unit economics for tokens will dramatically change.
The part where this will be fun is where the VCs lobby the US government, via peace board donation or golden iphones, to ban Chinese Open Source AI because otherwise they won't ever make their money back.
https://this.os.isfine.org/blog/posts/us-ai-labs-love-the-ai...
My bet is on model signing to run on US (Nvidia hardware stack), akin to what Nvidia already has built for various game console customers.
no doubt something like this is coming
Everything you need to know about the Tech industry (or any industry for that matter) can be learned from watching the Simpsons - Season 4 Marge vs. the Monorail
I am still waiting for a donut to save us.
Tech is pretty split here actually, just depends who you’re talking to and their incentives.
One potential path is to support and use open source AI models.
Hasn't the state of the art in genAI tools been making its way to open source/local inside of a year? Between that and pressure from Chinese firms, it would be hard for the current leaders to monopolize and enshittify their products. There isn't really a network effect to lock people in like with social media.
OP is right. Surely we will all go back to on premise any day now.
The article prophet is saying that AI companies are going to jack up the prices for their products - hence the evil, evil scam.
But don't we already have open source AI models which you can run on your own machine?
Another tech writer I used to respect who is now posting about how the future is now fantastically wonderful and wide-open because anyone can create an app to do whatever they want.
Alternative option, the writer did not fall for the scam, but is now in on the grift: https://www.cnbc.com/2026/02/06/google-microsoft-pay-creator...
As I said before, the ones boosting this everywhere to get as many tech folks on the train have either the following:
* Fully invested in those portfolio companies (no disclosures until IPO / acquired)
* Employees / Social Media Influencers at those companies with stock options which they are effectively paid boosters until they reach the vesting period.
* Frequently screaming the loudest and appearing on TV interviews, opinion articles spreading the scam further.
Just like the "Full Self Driving" scam that still requires a human behind the wheel, now we have "AGI" which still requires humans to supervise the AI agent to not make mistakes.
The tech folks don't question the euphoria and fall for it easily because they want to be on the next Google. But they could also be on the next Enron.
Only time will tell.