One contentious debate in AI research is about the ability of LLMs to reason like a human does. While there is still a lot of uncertainty, this paper from Google has a handy chart on the kinds of reasoning tasks AIs are tested on & how they do (GPT-4 wins) arxiv.org/pdf/2312.17661…pic.twitter.com/ZGC9NFVPS8
Thornton Neuroethics 