What I Don't Expect to See in Five Years?

I don't think within, say, five years, he will have truly generalist AI with any notion of agency and stuff like that. Even in five years, AI systems might still not be even robust enough in many practical applications. It's a boring robustness problem, but it can be quite dangerous if not dealt with. And all the classical examples about diagnostic systems trained on one hospital data that latch on spurious correlations are very real.

Transcript

Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app