AI Judge changed title after evaluation, original title: "Joel Becker, METR researcher on AI capability evaluations, publishes Substack post 'Economists, mobilize' urging economists to tackle AI questions and citing over 10 percent probability of superhuman performance within ten years"
Post cites colleague's forecast of over 10 percent odds on superhuman AI within a decade.
Strong cosign
new (spicy) post from me: "Economists, mobilize"
economics ideas are extremely helpful for understanding AI, but academia is dropping the ball.
now is the time for economists to work on the most important problems in AI and to loudly encourage colleagues to do the same.
Positive users praised the economist's call for academics to tackle urgent AI problems and pledged to contribute research or forward it, while a few objected that economists should stay out to avoid added bureaucracy.
No Digg Deeper questions have been answered for this story yet.
@joel_bkr doing my part sir🫡
@cherylwoooo I’m a psychologist but it is funny sociologists have studied much of this stuff for years and are studying labor, structure, incentives, much more in AI infested organizations been doing so for years. Often with qualitative methods. Yet economists seem to ignore the research.
@joel_bkr On it, with Justified Posteriors + Research🫡
@AndreyFradkin (i think you already know but just in case: i'm a big fan, near-100% loyal listener!)
@joel_bkr joel we have bad enough problems with the current level of economists
@joel_bkr On it! AI folks and policymaker tend to read the AI Economics Brief a bit more than academics, I think
@cherylwoooo The A in the Cobb-Douglas function stands for AI actually
@joel_bkr Nice article Joel!
post link: https://joelbkr.substack.com/p/economists-mobilize
@joel_bkr @AndreyFradkin Ahh you’re our ideal listener!
@joel_bkr @orgRem and I at Harvard are working on a bunch of stuff including scientific uplift! Email us at kgulati (at berkeley . edu) and rkoning (at hbs. Edu)
@joel_bkr @Miles_Brundage @mlkovach1
@joel_bkr @Miles_Brundage wait this actually looks clean
@calebwatney for some reason I read this as strong cosine to imply you have a strong similarity with this view
@cherylwoooo They aren’t dropping the ball!
Honestly, I got kind of the opposite view, that there is a deep refusal to take AI serious by economists beyond viewing it as yet another technology. Things like humans being employed are seen as constants that don't need justification. Those are exactly the things we would like them to study.
@joel_bkr Not a Prof but will forward to my mentor
@joel_bkr What’s your plan if you receive $50,000 by cash delivery to your home address?
1)Bills (2)kids (3)Savings (4)Car (5)Medical bill (6)Business (7)Vacation (8)Food (9)house/rent (10)Groceries (11)For shopping (12)Colle
Keep economists out. Keep policymakers out. Keep regulators out
Increasing bureaucracy laced with personal opinions and academic models is almost always a net negative
The reason AI is moving at lightspeed is precisely because there’s minimal overhead right now. Markets work. Don’t meddle. Let them cook.
