Jasmine Sun

And I hope that policymakers, political candidates start to think about what their role in this looks like.

Maybe they're sitting down with AI executives and saying, where are you seeing impacts on jobs?

What do you think we should do there?

Because

If my belief maybe naively is if you can come to a deal, if you can get to a bargain, we're going to be able to preserve the gains of technology, the growth that you get from technology without this kind of mass populist backlash.

Me neither.

Yes.

Oh, my gosh.

I mean, I feel like you guys are probably like masters at this.

So I don't know that I have any crazy tips.

You know, I mean, I use like I pay for like the best models I use every few months.

I will sort of I have my own personal eval.

So like mine is something like if I feed if I feed an AI like 10 interview transcripts and one paragraph about the kind of article I want to write, can it just spit out a reported article?

I never copy-paste these to be clear.

I do not actually use them, but that's the eval that I measure them on because I want to know at what point will they be able to do that kind of work.

And if they do start getting pretty good, I also want to know where's my comparative advantage going to be.

The way that I think about this, and I think this is what most economists would advise as well, is technology is going to get better, but so long as humans have a comparative advantage, then you're going to be okay, right?

As long as you're a complement to the technology.

And so I am actually almost more interested oftentimes in what it is the tech can't do yet.

The thing that the only way to find out what the tech can't do yet is to constantly be playing with AI so that you know, right?