Four: Brian Christian on artificial intelligence
Effective Altruism: Ten Global Problems – 80,000 Hours (October 2021)
00:00
Is There a Future for Transparency?
Satslik: I think we're still in the regime of good old fashioned misalignment. It doesn't know that it's tricking someone, or hasn't gone through some ell asshume. We've already moved on from convolutional nets to these huge transformer attention models. But this is a place where transparency is maybe fivish years behind the cutting edge of capability. And so we want to make sure that that would really like it to catch upyes.
Transcript
Play full episode
