Chris Miller

And the answer is not really, at least from what we can tell.

Closed source models had roughly 70% of the market 12 months ago, and it looks like they've got 70% of the market today from some of the

latest data that I've seen, most of the evidence suggests that most AI startups in Silicon Valley are building their startups on closed models because of the performance gap.

They're just better in terms of their quality.

So we haven't seen the commodification of AI at all.

I think the big question is, will we see the continued revenue growth that we've seen over the past couple of years?

you know, when you look at Anthropic going from 100 million to a billion to almost 10 billion in revenue the last three years, that is staggering growth.

And it can't continue at that rate for a long time.

But even if the rate slows somewhat, that implies a ton of new revenue coming in.

And I guess I remain on the optimistic side that yes, we're investing a lot, but I'd rather be investing a lot than not investing in a technology that has the potential to be so transformative across so many different sectors.

You know, I think Jeff's right.

It's a key question.

It's also a very, very hard question to know how to begin to measure.

I guess my theory would be that the first era of AI will be primarily applying AI to the enterprise space, making companies more efficient.

finding ways to automate legal services and accounting services and coding already happening.

And there, I think the empirics are that the US has done really well.

Most enterprise software firms around the world are based in the US.

The US has always been a leader in applying enterprise software to other sectors, to financial services, for example.

And so if that holds, I think you should expect to see U.S.

firms be early adopters of AI and enterprise as well.