🤖🛑 Hitting the breaks on AI?

(1) Alignment worries: perhaps legit. But don't have much to say.
(2) Job automation concerns: LLMs ≠ robots. Economic arguments for slowing down automation are much weaker for LLMs.

Why? Let's dive in! #EconTwitter #FutureOfWork #AI #ChatGPT Economists have put forth two main arguments for slowing down technologies that automate jobs and displace workers. The first is based on redistributive considerations. The second is based on efficiency considerations.

Taxi to JFK. Driver asks: where are you from? Argentina, I say. Driver turns around. Opens his jacket. He is wearing an Argentinean football jersey with the number 10 in it. I’m from Bangladesh, he says. We are big fans! You don’t believe me? Let me show you. Hilarity ensues… First he shows me pictures of his house. Big Argentina flag is hanging from his balcony. Then pulls up Youtube. Shows me this video of hundreds of Bangladeshi on the street with the 🇦🇷 jersey.

The more credible approaches (in my view) match conditional moments, like responses to a particular shock, not unconditional ones, like this blog.

There is a nice exercise illustrating why in my paper on regional business cycles: economics.mit.edu/sites/default/…

1/n

https://twitter.com/paulgp/status/1587983815238189059

Think about trying to identify one parameter: how "sticky" wages are. This is a key parameter in many DSGE models. All other parameters are known.

Imagine you have a strong prior that wages are very sticky when the "truth" is that they are rather flexible.

2/n