
1 - Adversarial Policies with Adam Gleave
AXRP - the AI X-risk Research Podcast
00:00
Observation Rate Increases With Self Placement Techniques
We are looking now in some follow up work on a rock paper scissors. That's a very, very simple gamey probably played atin no kindergarten. If you're playing against an r and n tat sees all the sequence of your actions, then is is actually quite a high dimensional space. And its still early work, but it looks like it is possible with some kinds of training set ups.
Transcript
Play full episode