Connor Leahy on AGI and Cognitive Emulation
Future of Life Institute Podcast
00:00
Cohen's Safety Property
If I have a system that can do any level of human science without killing me, that would be like my mark of success that Cohen has succeeded. So if you tell a Cohen to shoot your leg off, it shoots its leg off. It doesn't stop you from shooting your leg off. That's the safety property I'm looking for. The safety property is not, will always do the right thing and it's completely safe no matter what the user does. And they were like, what I would do if Cohen succeed, then that's the next step, like to go towards things.
Transcript
Play full episode
