I too was a bit surprised at first at this sharp turn on #AI podcasts.. then it occurred to me: We might learn a whole lot about how LLMs hallucinate reasoning.. 💡

Even LLMs can't generate that text apparently.. 😅

"Today, OpenAI has more than 1,700 employees, and 80 percent of them started after the release of ChatGPT in November 2022." nytimes.com/2024/09/03/tec… via The New York Times Cade Metz

Gearing up to speak in ~15min at the 3rd Neuro-Symbolic summer school. It is being streamed at youtube.com/live/aQ7dq8Kjg… Here are the slides I am planning to use: bit.ly/3ZdNFh6

Some fond (not necessarily high-brow) memories of Laveen from my grad school days at UMD Department of Computer Science (the best saved for the last..) --> The many times he would stand outside Dana Nau's room and ask "So Dana, how does this email thing work again?". At least on one occasion, Laveen

I have agreed to speak at UPenn NLP's CLunch next Monday on planning/reasoning abilities of LLMs.. First ACL and now this--not quite sure whether *I* am having an identity crisis or the #nlproc community is.. 🤷‍♂️😋 nlp.cis.upenn.edu/clunch.html

I just can't believe ScarJo is on Time's #AI 100 and Alicia Vikander isn't! The brain trust behind this august list really needs to look even more carefully (than they already are obviously doing; bless their souls!) into the AI contributions of people they include/exclude..