I have two issues with AI coding. The first is how giant trillion dollar companies are profiting off the accumulated labour of mostly open source developers who mostly offered their work for free. Open source had this pattern too but there was a kind of give and take that is not present here. The other is the atrophying of my own skills. That’s one of the reasons I don’t use it for my side projects since it’s become a requirement in my day job and I don’t want to forget how to code.
This creates a grey area for things like a personal blog. If you’re not trying to maintain skills, yeah I wish we could avoid supporting these companies, but I can’t be a hypocrite seeing as I have to use them for work, and I use my provided account as a personal tutor to learn cool stuff.
The implicit arrangement in corporate use and contribution to open source is that everyone benefits. People contribute work and testing that improves libraries and tools that everyone else can use, making all our projects better. This includes corporations, many of which also engage in this mutual benefit by hiring actual developers to back open source projects. I think it’s a brilliant system, huge fan.
Enter LLMs training off that data. It looks similar on the surface, but if you think about the balance of cost and benefit it becomes clear it is no longer balanced. The training algorithms and models are closed source, and come at a hefty usage cost. AI companies saved hundreds of billions of dollars of paying humans to help train their algorithms by taking advantage of an open system that was open on with another kind of balance in mind.
So in short the work was offered for free within a specific kind of relationship. Now a powerful outsider is steamrolling and taking advantage of that existing open dynamic.
Most FOSS is offered free of charge, but not free of conditions. You still have legal requirements spelled out in the licenses, from attribution to actually requiring you to license your work accordingly.
I feel like the standard "Free as in Gratis vs. Free as in Freedom" argument doesn't quite explain the issue people have here.
Open-source code still had terms, and AI training was an unfortunate blind spot in those terms. If a human being learned how to code 100% from open-source code, nobody saw a problem if that person went on to write closed-source code.
But it feels fundamentally different when the open-source code is just being fed into a machine that uses absurd amounts of processing until closed-source code comes out the other side. It feels like theft.
And that means nobody can make money off of their derivative work? If I open source my work, that means it is open source (by definition). It is a gift[1] - to everyone. If someone uses it to create value, good for them!
I have done a few open-source projects (none of those took off, but that’s OK). I did them because I wanted to give to the people, not greedy corporations. It was an implicit contract with the world, which has been broken by AI. You can say “but it was never written!” Maybe. It still hurts. (And it is also not really true: for example MIT requires attribution, which I’m still waiting to see in all of the AI-generated/AI-helped projects…)
It was also a joy to write things and know certain like-minded people would like it.
Now I don’t really want to do that anymore.
“I've learned a lot about using a tool like this but the biggest result is that development has lost its joy. There were no challenges, there were no puzzles, no hurdles to overcome. The biggest lesson is that software development is now about the destination not the journey, the journey is for the agent. If that's the case then the joy now has to be thinking about what is possible. I have built features into this site that I never would have on my own.”
I really appreciate these words. I am not a professional developer, and my code to this point has been limited to VBA and writing excel formulas (if that’s even coding).
I’ve recently started work on a full-blown Mac App. I’ve had the idea for years, but never the expertise. It’s not something I’d ever pay someone for either, mainly because it would be expensive and I don’t know if there’s a market for it. But I got a Pro subscription, fired up Xcode and got to work. Within minutes I had a proof of concept. Less than a month later, and I’ve got what I consider to be a nearly finished app with more features than I’d ever considered.
In the interim, I’ve also vibe-coded up two quick single-page web tools to my (and my wife’s) specifications. Things that I could describe, but would have taken me so much time to learn and build that I just wouldn’t have bothered.
> I've learned a lot about using a tool like this but the biggest result is that development has lost its joy. There were no challenges, there were no puzzles, no hurdles to overcome. The biggest lesson is that software development is now about the destination not the journey, the journey is for the agent.
Which is why I minimise my use of these tools. I'm not going to delegate the fun/meaningful parts to a machine.
I always feel bad and guilty when coding with ai support or with an agent. This is the new way and things can be done that seemed impossible but I always have a bad feeling
I always feel bad and guilty when coding with emacs, giving me such an unfair advantage over people using vi. This is the old way and things can be done that seemed impossible but I always have a bad feeling for vi users.
I've written at least hundreds of thousands of lines of Java code in my career, possibly over a million. I used to be extremely fast at building new features when I worked on a MLOC Java app I'd architected from the ground up. A huge percentage of the characters in that code were not typed by me, but were generated by my IDE.
Did I have an unfair advantage compared to the guys who were trying to get by using Notepad and the commandline?
Your mission is not to build software, it's to share your collection of cool typewriters with the world. And your AI built website helps you do that! It shares your YT channel, and now I know a bit more about typewriters. Thank you for building your site with AI and for building your collection with passion. Do you have a typewriter room somewhere?
I do actually have a typewriter room and I love to go in there and appreciate the craftsmanship and history. Much as the other replier commented, there are two separate things in play here. I AM glad to have a website for the collection, I'm very glad to share the collection and I do enjoy thinking of new and interesting ways to make that history more present on the site.
The other side of it is my job. My personal mission was to display my typewriters, but my professional mission IS to build software. This post was me wrestling with the latter more than the former.
I think you missed the message of the post -- sure, the typewriters are ostensibly the objective of the website, but there was a separate, interesting challenge that he enjoyed for, you know, human reasons: it's fun to learn new things and apply what you learned. In this case, it's software development, but it could easily be learning how to fix a car, how to cook a certain cuisine, how to play an instrument.
Sure, the mission is the end goal: having a fixed car, a completed meal, a sonata you can rehearse. These are all good for you, the viewer, to admire. But that's not where the enjoyment comes from for the creator. It's completely empty satisfaction.
>or they can't figure out how to become profitable after all this.
The AI white pills will argue they already are vastly profitable - although they always, always leave out the small detail of externalities (AI R&D costs and data center-related capex). Of course if you remove the two largest costs, the thing will be profitable.
My intuition is that ultimately the US federal government sustains the whole thing financially, and profitability questions will be dismissed for as long as it continues to do so
This is a silly post. No one is forcing this guy to do anything. He did it because he didn't want to take the time to learn to do this all himself or simplify his expectations so that he'd have to learn less. Or he did it because he wanted to and liked the way the website looks and it was easy and he may not actually care too much about the ethics (at least not as much as who ever he's asking forgiveness from).
the real risk with an ai built personal blog isn't getting hacked, it's what happens when something breaks and you have no mental model of the code to even know where to look
Lol fixing a bug in a simple personal blog is WAY above the paygrade of latest Claude/OAI models.
It will literally one shot fix any bug that is ever found in such simple application.
This is not much different for hiring a contractor to build your personal blog. What happens when something breaks and you have no mental model of the code to know where to look? You call up the contractor to fix it. You have simply outsourced the problem.
The mistake in logic here the assumption that having AI build your personal blog means that you build the personal blog. You did not. There should be no more expectation of having a mental model than if you got another human to do it.
The agent is your mental model. As someone who held pride in his ability to dive into unknown codebases and quickly figure out how they tick, Agents just do it more efficiently. Even if I have to follow up on what the agent finds, it's so much faster and efficient to have an agent in the mix.
By construction, a personal blog is very likely to break unexpectedly compared to other types of vibecoded projects because a) the scope is very narrow and b) there is only one user.
No problem, just have the AI fix it, or build it again from scratch. If it's done it once it can do it again. Unless you lose access to the AI, which is what I worry about.
It's just a matter of time. The time you get freed up by these new tools are going to be used by new things, possibly learning something even more complex, and that also includes, potentially, creating more complex software and experiences.
No one is saying nowadays that they feel bad when they code in a high level language that gets converted to assembly with a compiler, wondering about the loss of not knowing how to code in assembly anymore. Because... most people didn't want to actually code assembly at the end of the day. They did it because it was the means to an end (produce software).
Of course there are people who love code in assembly and coding in whatever language with their own hands. For them, maybe the code itself IS the destination. Nothing wrong with it. It's just discovering who you are and how do you want to enjoy your own time. The same can be said about anything in life, not just code.
Speaking of means to an end, previous to AI, I've been using already existent CMS for my web projects. Be it Payload, Strapi, Directus. It was always painful, I've never enjoyed it, but something I had to do in order to get the frontend rolling for me and clients. Now I can literally architecture backends (with their own frontend!) from scratch with sqlite, bun, and minimal libraries. Control is tighter, much easier to achieve personalized experiences for each website, less buggy, creating tests suites for even the weirdest edge cases is a painless process, I'm not dependent anymore of the direction the CMS maintainers decide upon. Being able to mold shape a backend this way creates, for me and my clients, a far better experience. All while leaving more time for what I enjoy much more, which is fiddling with the UI/UX and fine-tuning down to the pixel the frontend itself with my bare hands :)
> I wouldn't be using this if I could avoid it really. I don't use the "gig economy" at all. I don't use crypto either. But in this case, I couldn't beat them, so I had to join them.
> But until then, just to keep up, I have now learned to make a website using AI.
Come on, man. You could have taken the extra time to learn it and do it yourself. I use LLMs for development at work (and I do kind of hate what it's done to the process of discovery and joy in puzzle solving; this is why I still do my hobby programming without LLMs), but your typewriter website didn't have to be done on an enterprise schedule. This is like taking photos and using a printer to do your "painting" because you have to "keep up". It's a hobby, you definitely don't have to "keep up".
edit: I'm not commenting on the moral quandary here, but if you don't want to or have to do something, why do it anyway and feel bad about it?
> To be clear, none of the content, but the site engine itself is all AI built.
Feels a bit odd to apologise for this. I mean the only downside is you might get hacked if you don't understand your code or what it might be missing, but it's a personal blog, so I don't really see what real damage this would do or why you'd have to apologise for trying something out, whatever tools you use.
If you'd used AI to generate the content that would be a different matter, and is what people complain about as it is killing the internet. Instagram for example is now completely infested with AI generated slop, posting huge dollops of slop under videos that nobody ever reads.
> I don't really see what damage this would do or why you'd have to apologise.
FTFA:
> I didn't learn these libraries. Hell I barely know which libraries I've built on top of! Open source development used to run on attribution, how can I attribute what I barely know I've consumed?! Why would those developers continue to work on these projects when they now go even more unappreciated than they used to be?! I assume many have stopped.
> And those are just the libraries. The code it is generating is synthesized from being trained on code they have sucked out of the public (and perhaps not so public) internet. Code like my own from my own Github page and many many others, all without attribution or compensation. I could have paid someone to build it, freelance software development used to be a thing, but I didn't. I bet that market is pretty dry these days.
You know how we used to have little badges on sites like "built with react" or occasionally Wysiwyg editors?
Why not just have that at the bottom for whatever agent you used. That would give a good picture of the state of the art of agents and tooling and devs could show off their set up in a standard way.
Then people could create their own custom models with silly names like Linux distributions or cannabis strains, and then it would be all about fashion and branding.
I recently redesigned my own blog (https://minimaxir.com) during some experiments into how well modern LLMs can be conditioned into not outputting the stereotypical telltale designs of web vibecoded slop; it worked well enough (although more difficult than expected) and "how to design a website using agents that doesn't look like an agent made it" might be a fun topic for a blog post as there were a few gotchas.
The main reason I did the redesign is because my blog is running Hugo + PaperMod theme, which is becoming too popular and it makes blogs no longer stand out. There were also a few features/shortcodes I wanted added (better video support, better sidenote support, etc) that give me more tools for blogging.
How is your site not an example of your stereotypical web design? No unique design choices, no unique color palettes, just one single column for content with massive amounts of white space; it looks like any other generic dev blog.
Stereotypical modifies "vibecoded slop", not web design in general. As someone who has worked with blog coding for over a decade and attempted to actually do unique design choices, it is a generally bad idea to do so unless you have both a) a good design sense and b) analytics to test such that your unique designs don't actually make things worse. Blogs have converged to a one-column layout for a reason.
Your question seems disengenuous, and it’s not clear you read the post. But I would reckon many of us are having very similar experiences with how our daily activities are shaking out. The author’s story likely resonates with many devs here, who lament the loss of what many of them found interesting or compelling about the work they did.
I have two issues with AI coding. The first is how giant trillion dollar companies are profiting off the accumulated labour of mostly open source developers who mostly offered their work for free. Open source had this pattern too but there was a kind of give and take that is not present here. The other is the atrophying of my own skills. That’s one of the reasons I don’t use it for my side projects since it’s become a requirement in my day job and I don’t want to forget how to code.
This creates a grey area for things like a personal blog. If you’re not trying to maintain skills, yeah I wish we could avoid supporting these companies, but I can’t be a hypocrite seeing as I have to use them for work, and I use my provided account as a personal tutor to learn cool stuff.
What’s wrong with using work offered for free?
The implicit arrangement in corporate use and contribution to open source is that everyone benefits. People contribute work and testing that improves libraries and tools that everyone else can use, making all our projects better. This includes corporations, many of which also engage in this mutual benefit by hiring actual developers to back open source projects. I think it’s a brilliant system, huge fan.
Enter LLMs training off that data. It looks similar on the surface, but if you think about the balance of cost and benefit it becomes clear it is no longer balanced. The training algorithms and models are closed source, and come at a hefty usage cost. AI companies saved hundreds of billions of dollars of paying humans to help train their algorithms by taking advantage of an open system that was open on with another kind of balance in mind.
So in short the work was offered for free within a specific kind of relationship. Now a powerful outsider is steamrolling and taking advantage of that existing open dynamic.
Most FOSS is offered free of charge, but not free of conditions. You still have legal requirements spelled out in the licenses, from attribution to actually requiring you to license your work accordingly.
I feel like the standard "Free as in Gratis vs. Free as in Freedom" argument doesn't quite explain the issue people have here.
Open-source code still had terms, and AI training was an unfortunate blind spot in those terms. If a human being learned how to code 100% from open-source code, nobody saw a problem if that person went on to write closed-source code.
But it feels fundamentally different when the open-source code is just being fed into a machine that uses absurd amounts of processing until closed-source code comes out the other side. It feels like theft.
> mostly offered their work for free
And that means nobody can make money off of their derivative work? If I open source my work, that means it is open source (by definition). It is a gift[1] - to everyone. If someone uses it to create value, good for them!
[1] https://en.wikipedia.org/wiki/WTFPL
I have done a few open-source projects (none of those took off, but that’s OK). I did them because I wanted to give to the people, not greedy corporations. It was an implicit contract with the world, which has been broken by AI. You can say “but it was never written!” Maybe. It still hurts. (And it is also not really true: for example MIT requires attribution, which I’m still waiting to see in all of the AI-generated/AI-helped projects…)
It was also a joy to write things and know certain like-minded people would like it. Now I don’t really want to do that anymore.
[dead]
“I've learned a lot about using a tool like this but the biggest result is that development has lost its joy. There were no challenges, there were no puzzles, no hurdles to overcome. The biggest lesson is that software development is now about the destination not the journey, the journey is for the agent. If that's the case then the joy now has to be thinking about what is possible. I have built features into this site that I never would have on my own.”
I really appreciate these words. I am not a professional developer, and my code to this point has been limited to VBA and writing excel formulas (if that’s even coding).
I’ve recently started work on a full-blown Mac App. I’ve had the idea for years, but never the expertise. It’s not something I’d ever pay someone for either, mainly because it would be expensive and I don’t know if there’s a market for it. But I got a Pro subscription, fired up Xcode and got to work. Within minutes I had a proof of concept. Less than a month later, and I’ve got what I consider to be a nearly finished app with more features than I’d ever considered.
In the interim, I’ve also vibe-coded up two quick single-page web tools to my (and my wife’s) specifications. Things that I could describe, but would have taken me so much time to learn and build that I just wouldn’t have bothered.
> I've learned a lot about using a tool like this but the biggest result is that development has lost its joy. There were no challenges, there were no puzzles, no hurdles to overcome. The biggest lesson is that software development is now about the destination not the journey, the journey is for the agent.
Which is why I minimise my use of these tools. I'm not going to delegate the fun/meaningful parts to a machine.
I always feel bad and guilty when coding with ai support or with an agent. This is the new way and things can be done that seemed impossible but I always have a bad feeling
I always feel bad and guilty when coding with emacs, giving me such an unfair advantage over people using vi. This is the old way and things can be done that seemed impossible but I always have a bad feeling for vi users.
I've written at least hundreds of thousands of lines of Java code in my career, possibly over a million. I used to be extremely fast at building new features when I worked on a MLOC Java app I'd architected from the ground up. A huge percentage of the characters in that code were not typed by me, but were generated by my IDE.
Did I have an unfair advantage compared to the guys who were trying to get by using Notepad and the commandline?
Your mission is not to build software, it's to share your collection of cool typewriters with the world. And your AI built website helps you do that! It shares your YT channel, and now I know a bit more about typewriters. Thank you for building your site with AI and for building your collection with passion. Do you have a typewriter room somewhere?
I do actually have a typewriter room and I love to go in there and appreciate the craftsmanship and history. Much as the other replier commented, there are two separate things in play here. I AM glad to have a website for the collection, I'm very glad to share the collection and I do enjoy thinking of new and interesting ways to make that history more present on the site.
The other side of it is my job. My personal mission was to display my typewriters, but my professional mission IS to build software. This post was me wrestling with the latter more than the former.
I think you missed the message of the post -- sure, the typewriters are ostensibly the objective of the website, but there was a separate, interesting challenge that he enjoyed for, you know, human reasons: it's fun to learn new things and apply what you learned. In this case, it's software development, but it could easily be learning how to fix a car, how to cook a certain cuisine, how to play an instrument.
Sure, the mission is the end goal: having a fixed car, a completed meal, a sonata you can rehearse. These are all good for you, the viewer, to admire. But that's not where the enjoyment comes from for the creator. It's completely empty satisfaction.
>or they can't figure out how to become profitable after all this.
The AI white pills will argue they already are vastly profitable - although they always, always leave out the small detail of externalities (AI R&D costs and data center-related capex). Of course if you remove the two largest costs, the thing will be profitable.
My intuition is that ultimately the US federal government sustains the whole thing financially, and profitability questions will be dismissed for as long as it continues to do so
This is a silly post. No one is forcing this guy to do anything. He did it because he didn't want to take the time to learn to do this all himself or simplify his expectations so that he'd have to learn less. Or he did it because he wanted to and liked the way the website looks and it was easy and he may not actually care too much about the ethics (at least not as much as who ever he's asking forgiveness from).
And, bit by bit, humans start to lose.
the real risk with an ai built personal blog isn't getting hacked, it's what happens when something breaks and you have no mental model of the code to even know where to look
Lol fixing a bug in a simple personal blog is WAY above the paygrade of latest Claude/OAI models. It will literally one shot fix any bug that is ever found in such simple application.
This is not much different for hiring a contractor to build your personal blog. What happens when something breaks and you have no mental model of the code to know where to look? You call up the contractor to fix it. You have simply outsourced the problem.
The mistake in logic here the assumption that having AI build your personal blog means that you build the personal blog. You did not. There should be no more expectation of having a mental model than if you got another human to do it.
The agent is your mental model. As someone who held pride in his ability to dive into unknown codebases and quickly figure out how they tick, Agents just do it more efficiently. Even if I have to follow up on what the agent finds, it's so much faster and efficient to have an agent in the mix.
By construction, a personal blog is very likely to break unexpectedly compared to other types of vibecoded projects because a) the scope is very narrow and b) there is only one user.
No problem, just have the AI fix it, or build it again from scratch. If it's done it once it can do it again. Unless you lose access to the AI, which is what I worry about.
idk when most ppl's cars break down, they dont have the mental model to fix their cars either
we've been here / done this before, but in other journeys and with other "vehicles"
this is just another kind of vehicle, in a field where vehicles never existed
aka the purpose of 'Ask' mode. These new tools require one to 'Think Different'
My problem with AI is never going to be "It replaced Squarespace for small websites"
It's just a matter of time. The time you get freed up by these new tools are going to be used by new things, possibly learning something even more complex, and that also includes, potentially, creating more complex software and experiences.
No one is saying nowadays that they feel bad when they code in a high level language that gets converted to assembly with a compiler, wondering about the loss of not knowing how to code in assembly anymore. Because... most people didn't want to actually code assembly at the end of the day. They did it because it was the means to an end (produce software).
Of course there are people who love code in assembly and coding in whatever language with their own hands. For them, maybe the code itself IS the destination. Nothing wrong with it. It's just discovering who you are and how do you want to enjoy your own time. The same can be said about anything in life, not just code.
Speaking of means to an end, previous to AI, I've been using already existent CMS for my web projects. Be it Payload, Strapi, Directus. It was always painful, I've never enjoyed it, but something I had to do in order to get the frontend rolling for me and clients. Now I can literally architecture backends (with their own frontend!) from scratch with sqlite, bun, and minimal libraries. Control is tighter, much easier to achieve personalized experiences for each website, less buggy, creating tests suites for even the weirdest edge cases is a painless process, I'm not dependent anymore of the direction the CMS maintainers decide upon. Being able to mold shape a backend this way creates, for me and my clients, a far better experience. All while leaving more time for what I enjoy much more, which is fiddling with the UI/UX and fine-tuning down to the pixel the frontend itself with my bare hands :)
> I wouldn't be using this if I could avoid it really. I don't use the "gig economy" at all. I don't use crypto either. But in this case, I couldn't beat them, so I had to join them.
> But until then, just to keep up, I have now learned to make a website using AI.
Come on, man. You could have taken the extra time to learn it and do it yourself. I use LLMs for development at work (and I do kind of hate what it's done to the process of discovery and joy in puzzle solving; this is why I still do my hobby programming without LLMs), but your typewriter website didn't have to be done on an enterprise schedule. This is like taking photos and using a printer to do your "painting" because you have to "keep up". It's a hobby, you definitely don't have to "keep up".
edit: I'm not commenting on the moral quandary here, but if you don't want to or have to do something, why do it anyway and feel bad about it?
> To be clear, none of the content, but the site engine itself is all AI built.
Feels a bit odd to apologise for this. I mean the only downside is you might get hacked if you don't understand your code or what it might be missing, but it's a personal blog, so I don't really see what real damage this would do or why you'd have to apologise for trying something out, whatever tools you use.
If you'd used AI to generate the content that would be a different matter, and is what people complain about as it is killing the internet. Instagram for example is now completely infested with AI generated slop, posting huge dollops of slop under videos that nobody ever reads.
> I don't really see what damage this would do or why you'd have to apologise.
FTFA:
> I didn't learn these libraries. Hell I barely know which libraries I've built on top of! Open source development used to run on attribution, how can I attribute what I barely know I've consumed?! Why would those developers continue to work on these projects when they now go even more unappreciated than they used to be?! I assume many have stopped.
> And those are just the libraries. The code it is generating is synthesized from being trained on code they have sucked out of the public (and perhaps not so public) internet. Code like my own from my own Github page and many many others, all without attribution or compensation. I could have paid someone to build it, freelance software development used to be a thing, but I didn't. I bet that market is pretty dry these days.
The Freelance web dev market for personal blogs dried up about 20 or more years ago, but I suppose the author doesn’t know that.
Using things inspired by open source for personal use is in the spirit of most licences IMO. Again not something to worry about.
I’d say security should be a bigger worry for them.
I forgive you.
Thanks for your perspective on things!
You know how we used to have little badges on sites like "built with react" or occasionally Wysiwyg editors?
Why not just have that at the bottom for whatever agent you used. That would give a good picture of the state of the art of agents and tooling and devs could show off their set up in a standard way.
Then people could create their own custom models with silly names like Linux distributions or cannabis strains, and then it would be all about fashion and branding.
I know: Web rings for agents!
I recently redesigned my own blog (https://minimaxir.com) during some experiments into how well modern LLMs can be conditioned into not outputting the stereotypical telltale designs of web vibecoded slop; it worked well enough (although more difficult than expected) and "how to design a website using agents that doesn't look like an agent made it" might be a fun topic for a blog post as there were a few gotchas.
The main reason I did the redesign is because my blog is running Hugo + PaperMod theme, which is becoming too popular and it makes blogs no longer stand out. There were also a few features/shortcodes I wanted added (better video support, better sidenote support, etc) that give me more tools for blogging.
How is your site not an example of your stereotypical web design? No unique design choices, no unique color palettes, just one single column for content with massive amounts of white space; it looks like any other generic dev blog.
Stereotypical modifies "vibecoded slop", not web design in general. As someone who has worked with blog coding for over a decade and attempted to actually do unique design choices, it is a generally bad idea to do so unless you have both a) a good design sense and b) analytics to test such that your unique designs don't actually make things worse. Blogs have converged to a one-column layout for a reason.
Reads like slop. Crappy text
[dead]
[dead]
Why is that on HN at all?
Your question seems disengenuous, and it’s not clear you read the post. But I would reckon many of us are having very similar experiences with how our daily activities are shaking out. The author’s story likely resonates with many devs here, who lament the loss of what many of them found interesting or compelling about the work they did.
Why are you?