Raw LLM Responses

I'm kinda lazy to watch the vid, but I'll just say, as an "AI guy" myself (getting masters in AI and work with it, but wouldn't consider myself a very good specialist) I've never "understood" the people who hype the shit out of LLM or some other new models. Yes, models are getting better, incremental improvements here and there, but it's not even remotely close to completely replacing human in anything. The fundamental problem is that modern AI still isn't a real intelligence in any way, it's just literally some linear algebra with complicated memory storing, that's all. We need a bunch of breakthroughs to even come close to human cognitive power, and I imagine it will take some time. Anyone who says otherwise is just ignorant or flat out lying (especially for investor's money). For now, you still need a human, that is able to review and apply some changes to AI output. Even if we would be able to have real AI brain, I would argue it's not worth it to try and replace homan completely. But for now, another problem is, why to even use AI, if you need a human, and if you can to BECOME the human that is capable for particular task? Another thing is, more you use AI to do things - more you become dependent on it, losing your abilities even (and now we even have a research about it from MIT). If a task is not very trivial, but you want results fast - AI won't do any good (had that experience), especially if you can't fix bad output... what then? Tbh I wish nothing good on people, who thinks replacing people in creative fields, engineering, politics, etc. is good, or AI art is real art, or some other bs, from my point of view, if you say this -- you're not really passing as a human... AI is very interesting in terms of research, nothing bad about it, but when we get to applying it -- machines should be applied to the robotic tasks... It's sounds like a hot take but it shouldn't be like that..