If LLMs can capture human values,but first-order logic struggles,the only explanation is now that first-order logic lacks majority-vote quantifiers.Metaethically, this makes sense:Scanlon, Parfit,…

Posters: davidad. davidad

2 likes recorded by the source corpus.

Topics: AI & LLMs, AI safety

View Post: https://x.com/davidad/status/1662433209332768768