Four: Brian Christian on artificial intelligence
Effective Altruism: Ten Global Problems – 80000 Hours
00:00
Is There an Escapeye?
i've been very interested in some of the work that's happening at opena by people like daniel zeegler and others on kind of fine tuning language models or doing these sort of interactive feedback procedures to get task specific performance. I think we're maybe ready to retire some of the thought experiments the field of ai safety needed, you know, at the beginning of the 20 tenths. We don't need the hypotheticals when we have just this, like, bevy of real catastrophes to point to. And so, for better or worse, it's easier to make that the case now. What does musero do? Is at the stockcraft one.
Transcript
Play full episode
