If you invent/introduce made up concepts and rules on the spot ("let shmerple be ..."), which can't possibly be found in its training data, then, from my experience, when queried, LLMs are able to correctly reason most of the time... So it's not just Markov chains regurgitating sentences verbatim from their dataset.
Even if this kind of intelligence is just statistics, does it really matter? If it quacks like a duck, it's a duck. Maybe our own brains are nothing more than overrated statistical machines?
The only issue I see with the hype is when people attempt to anthromorphize it.
Even if this kind of intelligence is just statistics, does it really matter? If it quacks like a duck, it's a duck. Maybe our own brains are nothing more than overrated statistical machines?
The only issue I see with the hype is when people attempt to anthromorphize it.