News for hackers. Stuff that matters.

A comment on Peter Norvig critically reviews AlphaCode's code quality. Parent comment

Re:Peter Norvig critically reviews AlphaCode's code quality by olalonde on Friday December 16, 2022 @05:35PM (#34021560)
> This is a great review but it still misses what seems like the point to me: these models don't do any actual reasoning.

Hmmm... I have seen multiple examples of ChatGPT doing actual reasoning.

[ Parent | on HN ]
Re:Peter Norvig critically reviews AlphaCode's code quality by teaearlgraycold on Friday December 16, 2022 @08:01PM (#34022964)
I've also gotten it to make very obvious logical errors. It reasons by chance. Better than a dice roll but there is no internal knowledge graph.
[ Parent | on HN ]