Anders Sandberg on ChatGPT and the Future of AI
Future of Life Institute Podcast
00:00
How Do These Models Actually Work?
I had a discussion on moral philosophy with this model, and it felt real to me. It felt as if the model understood what I was saying. And then I asked it to correct the point that had made because it was wrong. But then it repeated the point over and over again, which breaks the illusion. So in one sense, as you say, this line of models showed understanding but in another sense it's it's brittle understanding. Also not sensitive to inconsistencies, which I think is very interesting.
Transcript
Play full episode
