Fun with ChatGPT ...part2
So we know ChatGPT3.5 is not very good at math: couldn't even factor Georgia Techs 222-0 score. Being trained on language alone might be the problem, and I speculated that if humans find math hard then an LLM might also have difficulties.It would be ...
carlek.hashnode.dev5 min read
carl ek
(learn + do) * repeat
testing testing 123