Tracy Allaway

And like people were, they were trying to club together to like invest in these companies.

So clearly there are people out there who are using these charts as investment tools.

Actually, on this note, this reminds me of something I wanted to ask.

So when you look at the domain specific time horizon charts, so the ones that show like I think you call them task suites or something like that, like I guess productivity by specific job and you see these different lines.

So sometimes you see like almost horizontal lines and sometimes you see squiggly or steeper lines.

What is actually happening there?

Like, how are we supposed to interpret that?

Like, is this a measurement problem or is it saying something very fundamental about, like, what AI can and can't do under current conditions?

But for other tasks, there's... Beautiful code, elegant code, people always talk about.

For other tasks, there's going to be... If Anna Wintour was coding, this is what it would look like.

Can I just ask very quickly, since you brought up China, and I don't want to forget to ask this question, but QAN doesn't show up on your main charts.

I think you did a preliminary assessment of it a while ago.

But what's the difference between assessing one of the closed models in America versus one of the open source models over in China?

So they're so irrelevant, they just don't make it onto the chart?

You mean they're gaming the benchmark?

Is that what that means?

Yeah.

Like who are you interacting the most with at the moment?

People listening to this podcast, presumably.

But no one seems to be really thinking about it in a lot of detail.