Oz Veloshin

You mentioned that you started your career or your study of AI when it was a very unpopular field.

There was the kind of long AI winter where nobody really believed it would happen.

And I think you

reference, you didn't believe in your lifetime that you may never see AI be able to distinguish between a cat and a dog in your lifetime.

Safe to say it can now.

But with your doubts about where the field would actually take off, what made you want to pursue it before it was the hottest thing in the world?

So what changed?

I mean, it's interesting you mentioned like 20 years ago, it would have been like, say, I'm going to study teleportation.

I'm curious if you agree with my kind of hypothesis in my introductory remarks about this idea that if the royal institution founders could see what's happening today, their minds would be completely blown.

And yet there's this cultural feeling of despair or darkness.

So you sit in many ways at the coalface of the future.

What are you seeing when you're working in DeepMind, when you're interacting with AI, when you're building it?

What are you seeing the rest of us should know?

Except you know how to interrogate in a way that we don't, right?

And I think that's an important point because when we talked before this, you mentioned that for the last couple of years, you viewed yourself as a builder.

And now you've described the feeling of surfing.

Is this the open-claw revolution?

I'm experiencing right now yeah I mean you're you built something which now has emergent capabilities that you're trying to understand so you don't understand all the capabilities of the thing that you built which is mind-bending

You've described your own work as creative, I think, and I know you're also an artist.

What can we learn from your creative process in terms of how you