The Scale of Deep Learning Systems

We were big believers in the powers of evolution as a driver of improvement or other optimization to your environment. So effectively obviously the human species is an optimized outcome that fits well on the earthly environment. And so the process the same here, you're using the process of competition, culling, mutation and creating offsprings to progressively create a population of hyper-adapted, hyper-optimized beings. In some ways we're right that scale does matter. I think with what's happened today, there's still some work that's done about an architectural search at the bidding edge.

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