Every time I probe LLMs on questions on fields in which I have expertise, it goes the same way.
They give me some sort of American news media "both sides" type of summary. The answer is always a variation of "it's not one thing, it's many" and the many things are all the opinions that exist on the topic as if they were all equivalent.
Hinton talks are hilarious. He goes from explaining perceptrons to huge claims like "this is what understanding is", "that's how our brain works", "linguists are cranks if they don't agree", and "this proves there is no god" but the last one as some weird little anecdote involving some religious migrant worker, and implies heavily that disagreeing with him is equivalent with having religious beliefs.
I am also tired of this whole "hallucination" nonsense
These LLMs are buggy as hell. They say they can do certain things - reasoning, coding, summarizing, research, etc - but they can't. They mangle those jobs. They are full of bugs and the teams behind them have proved they can't debug them. They thought they could "scale laws" out of it but that proved as unfruitful as it was illogical.
What class of software can work this bad and still have people convinced the only solution is to double the amount of compute and data they need, again?
I hope you are doing this for "virality". This is nonsense and if you believe any of this, seek help.