Raw LLM Responses

Linus Torvalds spot it all. He said "It is [just] pattern matching on steroids". Another way to see it is that it is "Search engines repackaged". But Hinton continues to instill the idea that this bubble is something superior, that it has intentions (beyond what trainers decide to submit to it for the sake of 'seducing expression' - muck like pot sellers), that is has some kind of consciousness. The problem lies in not what [by the way misnomered] AI is, but the problem lies in these kinds of foggy talks about it. They submitted to it a lot of human input, together with [supposed] quality of that input (written by humans), and it does a nice job for saving me time going on wikipedia or sometimes technical info and/or programming patterns. Nothing new except the speed of obtaining information. But people forget that human brains were thinking while they were googling for results. So it can augment the productivity when it comes to some low-level, but not where thinking is involved. The usefulness from the statistical point of view is there, and it was shown by Deep Seek, well before the Hinton's hyped AI, by cross-pattern matching when it comes to diagnoses in medicine and similar. If there was some of whatever small amount of intelligence they wouldn't publish it, and wouldn't create of this hype around. It is not "Artificial Intelligence". It does not have whatever kind of cognitive model. Just stochastic probability generator.