
4 - Risks from Learned Optimization with Evan Hubinger
AXRP - the AI X-risk Research Podcast
00:00
Openin Five Agents - Is There a Coherent Objective?
Openin five was a project byk opening to solve dona o two, where they built agents that are able to sort of compete on a relatively professional level. They probably havelike a lot of sercuristics. Potentially there's some sort of search going on, whough i guess would be not very much. But nevertheless, we can look at it be like, there's clearly a sort of coherent behaviour objective in terms of trying to win the dotagan.
Transcript
Play full episode