Sendhil Mullainathan

@sendhil.bsky.social

Really excited about this paper. Part of our continued efforts to make sense of what ML models are and aren’t doing.

Keyon Vafa @keyonv.bsky.social · Jul 14

Can an AI model predict perfectly and still have a terrible world model?

What would that even mean?

Our new ICML paper (poster tomorrow!) formalizes these questions.

One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵

6

Reposted by Sendhil Mullainathan

Keyon Vafa

@keyonv.bsky.social

Can an AI model predict perfectly and still have a terrible world model?

What would that even mean?

Our new ICML paper (poster tomorrow!) formalizes these questions.

One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵

1 14 38

by Sendhil Mullainathan

Reposted by Sendhil Mullainathan

by Ajay Agrawal — Reposted by Sendhil Mullainathan

Reposted by Sendhil Mullainathan

Reposted by Sendhil Mullainathan, Norbert Schwarz, Olivier Corneille , and 2 more Sendhil Mullainathan, Norbert Schwarz, Olivier Corneille, Jana Lasser, Kris Hardies

Stephan Lewandowsky

@lewan.bsky.social

Honest people don’t lie. Or do they? Liars aren’t honest. Or are they?
One puzzling conundrum in contemporary politics is that politicians who seem to be estranged from facts and evidence are nonetheless considered honest by their followers.
1/n

14 92 250

Reposted by Sendhil Mullainathan, Miles Corak

Jann Spiess

@jannspiess.bsky.social

Last chance to apply to this year's Machine Learning in Economics Summer Institute, our AI/ML summer school for graduate students, co-organized with @sendhil.bsky.social @asheshrambachan.bsky.social @aleximas.bsky.social @lindseyraymond.bsky.social: www.chicagobooth.edu/research/cen... #econsky

Machine Learning in Economics Summer Institute 2025 (MLESI25)

<span id="docs-internal-guid-53f97f20-7fff-b55c-c89c-4f85d663a48c" style="color: #39393a;"><span style="color: #000000;">MLESI<span id="docs-internal-guid-394e21c4-7fff-1437-16ff-fdbb1ce1d34c" style="...

www.chicagobooth.edu

9 12

Reposted by Sendhil Mullainathan

Ben Waber

@bwaber.bsky.social

First was #SSAC day 2 w/sessions on basketball analytics (@deepakmalhotra.bsky.social, S Bird, E Wasch, M Zarren, D Morey), NFL front office (L Horton, @scottpioli51.bsky.social, K Demoff, @espnbillc.bsky.social), & AI (@sendhil.bsky.social, @pablo.show, Morey) www.youtube.com/watch?v=jG1A... (2/5)

SSAC25: Main Stage: Day 2

YouTube video by 42 Analytics

www.youtube.com

1 1 2

Reposted by Sendhil Mullainathan

by Cass R. Sunstein — Reposted by Sendhil Mullainathan

Reposted by Sendhil Mullainathan

Reposted by Sendhil Mullainathan, David Lusseau, Dimitris Xygalatas , and 2 more Sendhil Mullainathan, David Lusseau, Dimitris Xygalatas, Bill Thompson, Gregor Wolbring

Melanie Mitchell

@melaniemitchell.bsky.social

An incredible opportunity for the right person:

The Santa Fe Institute is seeking applications for full-time resident faculty positions at all academic levels

For me, this has been the best possible job ever.

More info here: santafe.edu/about/jobs/r...

sfiscience

SFI seeks applications for full-time, 12-month resident faculty positions at all academic levels.

santafe.edu

9 210 380

by Andreas Fuster — Reposted by Sendhil Mullainathan, Ralf Martin

Andreas Fuster

@fuster.bsky.social

For those who didn't make it to #ASSA2025: strongly recommend @sendhil.bsky.social's AEA distinguished lecture, available at www.aeaweb.org/webcasts/202... (starting at minute 16)!

American Economic Association: AEA Excellence Awards and Distinguished Lecture

www.aeaweb.org

4 21

Reposted by Sendhil Mullainathan, Arvind Narayanan, Arun Sundararajan , and 1 more Sendhil Mullainathan, Arvind Narayanan, Arun Sundararajan, Taku Ito

Melanie Mitchell

@melaniemitchell.bsky.social

Some of my thoughts on OpenAI's o3 and the ARC-AGI benchmark

aiguide.substack.com/p/did-openai...

Did OpenAI Just Solve Abstract Reasoning?

OpenAI’s o3 model aces the "Abstraction and Reasoning Corpus" — but what does it mean?

aiguide.substack.com

17 99 340

References

Fields & subjects