The Sentience Institute Podcast cover image

Raphaël Millière on large language models

The Sentience Institute Podcast

00:00

The Limitations of Current Models

There are a number of ways in which current models are still very limited. They undergo this training phase so they could train them a lot of data. And then once they're trained, they are frozen, as people say. So there's no more learning. There is no more evolution of the system. It's totally frozen in that sense. That's extremely different from the way in which biological systems leave evolve and learn and acquire intelligent capacities. We need something like a meta-cognition to think about their own representations and reflect upon that.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app