If you look at the last 3 years of AI progress, technologists have largely been right about the (massive) increase in capabilities, and economists have largely been right about the economic impact. I don’t expect that to change.
Ricardian comparative advantage ftw!
Post
Conversation
Post your reply
I'd be more confident about the negative version: technologists are v often naive about econ impacts, and academic economists have frequently ignored or assumed away capabilities progress.
So maybe economists continue to be more right on near-term effects, but wrong on
I don't think we know who will be right about the long term unconditional economic impacts yet...
Fwiw there's a bunch of resources recording predictions across the camps: tecunningham.github.io/posts/2025-10-
kevinbryanecon.com/ForecastingAI.
can you be more specific about what economists were right about? Many prominent economists predicted massive labor replacement happen by now since like 2016, but no one finds that (yet)?
Labor market impact and growth. I can think of only one economist who predicted massive impact at this point, everyone else was very firm on bottlenecks and frictions slowing things down. I’ve been pretty vocal about this.
Would you agree w 's raising the late 90s parallel (sub-1% productivity boost) in his recent interview with ?
To me, it both overestimates and underestimates the (implied) relevant coefficients (& assumes these iterative changes evidence in macro terms).
I disagree completely. Here are two examples that I think are representative of how economically important the two camps were expecting AI to be:
* Robin Hanson losing his revenue bet (x.com/robinhanson/st)
* OpenAI forecasting $1B ARR in 2024 as of 2022 (was $3.7B, not a bad
Quote
Robin Hanson
@robinhanson
I've just paid off on this bet to @ATabarrok, who won. x.com/robinhanson/st…
If you think US uses fewer resources than China in producing equivalent quality AI, you are living in a cave.
I don't understand why AI companies don't make more effort to make positive economic arguments, rather than their senior execs say AI will devastate employment.
There is a good positive case they should make. Even if there wasn't they should being trying to make one!
I think accurately predicting sub-AGI AI impact on the labor market requires ASI, and accurately predicting ASI impact on the labor market requires a Dyson sphere up and running.
The US unless it bans data centres will continue to grow at 2% every year--as fast as the industrial revolution
Meanwhile evey single person I know that's unemployed has been unemployed for > 10 months. Nothing to see here.
But technologists and economists haven't disagreed much on the economic impacts in the last 3 years, though they disagree a lot on the economics impacts in the next 3 years. We don't have much data on who to trust
Does this line of thinking account for the fact that reasoning models capable enough to make a dent on any type of job are at most, 9 months old?
Framing AI progress through Ricardian comparative advantage explains why massive capability gains haven't immediately displaced human labor en masse.
KEEP GRAPHIC TEES AFFORDABLE!
Premium tees at unbeatable prices. Just $6.99 each when you buy 10+! Thousands of designs to pick from. Printed & Shipped in the USA.
Slide 1 of 6 - Carousel
Third time's a charm
Quote
Vanta
@fangedlumina
The fascinating thing about all of this, and especially the @tylercowen retweet - when you'd imagine he knows better - is we are quite obviously in the *buildout* phase which is quite obviously job-creating.
The tech is not good enough to replace jobs yet but it is good enough x.com/sama/status/20…
Yeah, but techs are actually betting on economic outcomes now at macro scale. So one of them would probably be a lot more right then the other in 3 years 
This does not change gradually, but suddenly. An AI agent that is twice as good as a human at half the tasks and no good at all at the other half, is very different from an agent that is as good as a human at all tasks in a job.
What jobs are humans going to be paid to do when AI & robots can do them all better, faster & cheaper? As the technologists predict.
How does capitalism function as a social-economic-politcal system under something approaching 100% unemployment? Or even 30%?
Most men wait until it's awkward, you won't. It's time to be proactive about ED. 50% off sitewide.
