Kevin Hartnett

It was not just kind of a clever cobbling together of like obvious techniques.

And the result itself was just like so good.

Pretty unanimously, people agreed this could be published in the Annals of Math, the top journal in math.

In a way, over the last year, there'd been this kind of shifting of goalposts.

AI did this, but it can't do that.

Oh, it did that.

Nope, it still can't do this.

But this result, this proof of the unit distance conjecture really said AI can do absolutely top tier research.

Yeah.

So Terry is an extremely interesting figure in this because he is so important, as you said.

I have a whole chapter in my book about some of the early work that he did with AI, this thing he called equational theories.

And he's always been interested in different ways of doing math.

He's very intensely collaborative.

He's interested in kind of new ways of working.

Yeah.

Terry is, I think, kind of representative of one of three attitudes towards AI and math right now.

A couple of weeks ago, I was at the Institute for Advanced Study in Princeton, New Jersey, which is like the citadel of modern math, like the biggest, the most dense collection of kind of great math minds in the world live and work there.

And in a single afternoon, I was walking the campus there and I had two strikingly different experiences.

I ran into two 40-year-old mathematicians, top of the field.

One of them told me he just tried to do some math with Gemini and he's like, that stupid thing told me like XYZ thing, which we know is wrong, is true.