Jasmine Sun

They only have so many wants, right?

And so if you have a world that's very unequal...

which is something that we expect with AI because there's going to be more returns to capital, those rich people, they may be able to hire a few party planners and a few personal trainers, but they got 24 hours like the rest of us.

And so you're just not going to have as much demand in a very unequal economy compared to one where there's a really strong middle class and everybody is buying a lot of services and goods all the time.

So those are some arguments that I would consider making if I were trying to make the more extreme case.

But once again, I just want to say that my own beliefs are a little bit more moderate.

Yeah, I mean, so what do I think is going to happen?

I lay some of this out in the New York Times piece.

I do expect the near-term labor disruption.

I think that there are

certain categories of jobs that are way easier to automate than others.

And this is where a lot of my disagreement with people like Dario comes from is that software engineers, super easy to automate because code is verifiable.

It's all the context is in a code base.

You have this like open source data on the internet that you can go train on.

Most jobs are not like that.

Software engineering is a really weird type of job.

Maybe accountants are also like that.

There's like a few jobs like software engineering, maybe digital marketing, copywriting and freelance digital illustration, maybe like accountants or something, management consultants.

Let's call it like 10% or 5% of the US economy is jobs that are very, very easy to automate for like some slate of reasons like this.

Those I do think are going to get disrupted pretty quickly because financial incentives are just going to make bosses choose to use AI over hiring humans, especially like when a human gets laid off or they quit their job, you're just not going to replace them if an AI can do a good job.