I find this perspective quite difficult to reconcile with my own experience. I've saved countless hours using AI, solved problems I wouldn't previously have known how to approach, and used it to venture into areas well outside my existing skill set, often learning a great deal in the process. So when you insist it's useless, destroys critical thinking and erodes skills, I can't help but feel you're demonstrating a lack of imagination about how it can actually be used, along with a bit of doomerism.

If you're blindly accepting everything it tells you and outsourcing all your thinking to it, I can see how that might be detrimental. But that's one way of using it. You can just as easily use it to explore unfamiliar subjects, interact with complicated concepts, challenge your assumptions and develop skills you didn't previously have. These are largely questions of education and critical thinking, and teaching people how to evaluate information and use tools responsibly is invaluable whether LLMs exist or not. We could honestly use a bit of motivation to push this harder don't you think?

Again, the environmental costs and concerns about cognitive dependency are valid, but neither establishes that the technology has no utility. Likewise, the enormous financial losses of some AI companies raise serious questions about their business models, but you're making the common mistake of confusing that with whether the technology itself is economically viable. The cost of running models at comparable levels of capability has fallen enormously in a handful of years, and there's no reason to expect that to suddenly cease.

Your argument here conflates three different questions: whether AI is useful, whether the current industry is financially sustainable, and whether its widespread adoption is benefiting society. These are related, but fundamentally different questions with potentially very different answers, and you seem more interested in dismissing all three outright than examining them individually.

And once again, several of your criticisms concern corporate behaviour and economic incentives, which is precisely what my original comment was about. Others are legitimate questions about the technology's limitations and how people use it. But you're lumping all of these together as though they demonstrate that AI is fundamentally useless, when they don't. A technology can have serious limitations, be deployed irresponsibly and still offer enormous practical value.