Have a question that is challenging for humans and AI? We (Center for AI Safety + Scale AI) are launching Humanity's Last Exam, a massive collaboration to create the world's toughest AI benchmark. Submit a hard question and become a co-author. Best questions get part of $500,000 in

One of the things I really notice at the Cambridge existential risk conference #CCCR24 hosted by Centre for the Study of Existential Risk is how many of the participants think they got the best job in the world.

Panel at #CCCR24 generally think opaque funder goals are bad - causes too much overthinking about how to adapt research to it, instead of just selecting reasonable one. I guess opaqueness is due to trying avoid gaming, and thinking that it represents a wide and open-ended goal.

Can you get time off from a prison sentence because you did not attempt omnicide? #CCCR24 (Generally, I agree with panel that real omnicidal terrorists are rare, but accidental existential terrorism might be a risk.)

list of unusual presentations of classic graphics or familiar information [please contribute]

Matt Boyd presents the project leading to the NZCat report at #CCCR24 adaptresearchwriting.com/2023/11/16/mai… - island nations are potentially valuable natural refuges in big disasters, but need adaptive governance. islandfutures.earth

A new tome has arrived. #CCCR24. openbookpublishers.com/books/10.11647…

(peak behind the scenes - Anders and I are sat next to each other and we were nerding out *hard* during this talk. Really scratches the mad scientist meets prepper itch)

Final panel of #CCCR24 today, about GCR risk reduction, response and resilience. Very cool!