Jason Crawford

Or is there one process that generates innovations that have a power law distribution?

And we think there's two types because we see a lot of really big ones and we see a lot of tiny ones.

You know, we don't see like a full like a whole spectrum of things like we sort of expect because we're just not that attuned to parallel distributions.

And he has sort of a good argument that that that the latter is the case.

Um, so yeah, there are huge changes.

There are small changes.

Um, we shouldn't over-focus on the huge changes and, and neglect the role of, uh, of a lot of incremental innovation.

Um, a lot of progress just comes from compounding, you know, 1% improvements like every year, two or 3%.

Um, so, uh, yeah, absolutely.

I rolled my magic eight ball and it says, future is cloudy, try again later.

I don't know.

I don't have a great... And I'm not sure anybody can really know.

And if there are people who can really know, it's probably people who've spent a lot more time thinking about AI than I have.

I don't know.

I think it could be big.

It's one of these things we're on the verge of, and it's hard even to know how far away some of the big breakthroughs are, let alone really how it's going to transform.

Let me say one thing on this.

And this is sort of maybe... I think this is a contrarian position, especially in rationalist circles.

It is not at all... I think it is often assumed that...

The type of intelligence that machines have or that we can create in them is essentially of the same kind as the type that we possess through our biological brains and consciousness.