Jacob Steinhardt, UC Berkeley: Machine learning safety, alignment and measurement
Generally Intelligent
00:00
Why Is It So Hard to Be Calebrated?
"It's not clear that you should think it's trivial to do this," says Evgeny Kuznetsov. "You can have a bunch of verbal debates, but often just adding more data points is going to be a lot more valuable in terms of mausation", he adds. The next step for the researchers will be finding out if their models are as robust as they first thought and how best to improve them.
Play episode from 28:02
Transcript
