5 ms·
Current LLMs are optimized to produce output most resembling what a human would generate. Not surpass it.
by jimbokun 2y ago
Current LLMs are optimized to produce output most resembling what a human would generate. Not surpass it.
- ben_w 2y agoThe output most pleasing to a human, which is both better and worse. Better, when we spot mistakes even if we couldn't create the work with the error. Think art: most of us can't draw hands, but we can spot when Stable Diffusion gets them wrong. Worse also, because there are many things which are "common sense" and wrong, e.g. https://en.wikipedia.org/wiki/Category:Paradoxes_in_economics https://en.wikipedia.org/wiki/Category:Paradoxes_in_economic..., and we would collectively down-vote a perfectly accurate model of reality for violating our beliefs.