Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Pricing

Dwarkesh

👤 Person
1735 total appearances

Appearances Over Time

Podcast Appearances

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

becoming superhuman at coding competitions doesn't make you a more tasteful programmer more generally.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

Maybe the thing to do is not to keep stacking up the amount of environments and the diversity of environments to figure out an approach which lets you learn from one environment and improve your performance on something else.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

But then what is the analogy for what the second student is doing before they do the 100 hours of fine-tuning?

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

I think it's like they have it.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

I think it's interesting to distinguish it from whatever pre-training does.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

So one way to understand what you just said about we don't have to choose the data in pre-training is to say, actually, it's not dissimilar to the 10,000 hours of practice.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

It's just that you get that 10,000 hours of practice for free because it's already somewhere in the pre-training distribution.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

But it's like maybe you're suggesting actually there's actually not that much generalization from pre-training.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

There's just so much data in pre-training.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

But it's like it's not necessarily generalizing better than RL.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

Here's analogies that people have proposed for what the human analogy to pre-training is.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

And I'm curious to get your thoughts on why they're potentially wrong.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

One is to think about the first 18 or 15 or 13 years of a person's life when they aren't necessarily economically productive, but they are doing something that is making them understand the world better and so forth.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

And the other is to think about evolution and

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

as doing some kind of search for 3 billion years, which then results in a human lifetime instance.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

And then I'm curious if you think either of these are actually analogous to pre-training or how would you think about at least what lifetime human learning is like if not pre-training?

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

What is that?

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

Clearly not just directly emotion.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

It seems like some almost value function-like thing, which is telling you which decision to be made, like what the end reward for any decision should be.

Dwarkesh Podcast
Ilya Sutskever – We're moving from the age of scaling to the age of research

And you think that doesn't sort of implicitly come from... I think it could.