Further:

> They were trained on stolen information, and where they didn't break the law they corrupted social contracts.

I completely reject this framing. Learning on something has always been part of the social contract (public libraries are a great example).

But I do care about, for example, a recently published author, or someone who posts their paintings on social media. Regardless of where the law stands in each particular case, I think it's repugnant that statistical models were trained on these inputs with the intent of offering the ability to produce similar outputs without any payment or even attribution to the original artist.

> Information wants to be free!

If someone chooses to make their information free, then so it may be, but the issue at hand is very obviously not about the use of, e.g., copyleft materials. The AI companies knowingly chose to take material that they knew they should not have and used it to make their copycat machines, building those machines in such a way that those responsible can cower behind the letter of the law to defile its spirit.

I think about how disappointing it is when a new project makes it to the front page of HN and I click through and it's just... vibe-coded slop. Looks pretty at first, but it's clear that there isn't really much human underneath. Lots of obvious interface problems or just a distinct lack of a personal touch. It's empty, and it makes me sad.

Or even not in software, I think of all the people who have been so quick to offload their thought processes to a statistical model. What do you mean you don't know what you want to eat at this restaurant so you asked Claude and got recommended something that isn't even on the menu?

And that's just individual stuff. I am genuinely fearful of the integration of AI into, say, the legal and judicial processes, or governance, or whatever else. People give the models wayyyy too much credibility considering they (the models) cannot give an accurate confidence interval for their results, and I think it's going to make things much worse for disadvantaged communities to a degree that the communities will not be able to overcome by using the technologies themselves.

For the time being, I'm content to simply not use it myself. I don't think it's reasonable for me to foist my morals onto those around me, in the same way that my very close vegan friend still talks to me when I order a whole milk latte. I don't think there's a way to exist in modern society without incurring some degree of (likely indirect) immorality, so it's all about setting boundaries for myself where I can and accepting where I end up.

But, of course, this is also part of the immorality of the AI industry.

If they come to me with a cartoon-like contraception that comes straight from Acme, you can understand how reluctant I am to not be blown up like Coyote.

Money is an abstraction, not an approximation. There is not of necessity any real basis for any price that has ever been attached to any product or service, ever. Money is an artificial mechanism by which we exchange debt owed by the issuing state that itself also has no real basis.

> If you don't care about creating value, you don't care about people.

On this we agree, but, of course, my point was that when I say "value" I don't mean monetary value. A person's value is intangible, and the value of their works (whatever those may be) is derived entirely from the relationship formed between the artist and their context and the connection from this relationship to whoever wishes to partake of the art.

> As, ultimately, it costs money to care for people, and if you don't have it - all you have are prayers and best wishes, and you functionally can't help anyone.

What a woefully capitalistic perspective. I wish the best of luck to you and your shareholders.

> where the anti-ai crowd ends up in a few years.

I think if LLMs turn out to be unviable in a few years, then anti-AI crowd (hello!) can stand there proud as they did not deskill themselves and kept themselves more independent than most of their peers.

I admit that the above sounds more like a delusion at the moment, but hey one must dream.

So, I'd love to imagine what the future looks like once we're stuck with a model (no matter how capable) that can never be updated ever again. What future versions of "do it for my grandma" can never be patched out?

I guess the main features I'm thinking of when talking of a bubble is 1) demand (especially at the real price) and ultimately return of investment is not there and 2) circular financing is keeping the frenzy ongoing despite that. This by definition is not profitable and sustainable and if that becomes the mainstream belief then nobody would lend the industry money, rendering it bankrupt.

> The AI models arent going away. The tech is still there.

I'm under the impression that AI data centers (which are very different from other data centers) are only good for a handful of years at most, because of the GPUs limited lifetimes, and also are expensive to run (electricity) which makes AI fundamentally different than DotCom as the infrastructure built back then which is fiber cables was very much not like that. And you need AI data centers for both inference and to train new models. Which is why I asked you if you think we could keep using the same models we have right now, in which case we'd need much less compute and get lower maintenance costs to keep this going even after a crash.

Was Pets.com burning tens of billions of dollars per year? This comparison is not favorable to Anthropic.

Do generally agree that the models aren't going away. Neither OpenAI nor Anthropic are financially viable without government intervention, but it's not hard to be profitable in this sector and it will be even easier to be profitable after the unprofitable guys stop spending hundreds of billions in fake money inflating hardware prices.

If they're never going to be updated again (because, according to the labs themselves, training and updating them is the part that loses money), then how soon will the next "do it for my grandma" be discovered? And how useful will the models really be if every public document, and I imagine a decent number of open source libraries, will inevitably have that magic phrase embedded somewhere. php2shell was bad and that was patchable.