Max Chupilkin

Principal economist at @EBRD & pursuing DPhil at @Politics_Oxford. Interested in all things international.

[Max Chupilkin @maxchupilkin]

https://t.co/eUu81k6IOe

[Max Chupilkin @maxchupilkin]

When frontier AI models evaluate economic policy trade-offs, which outcomes actually move their judgments? (1/6)

[Max Chupilkin @maxchupilkin]

Link to full essay in the comments. (6/6)

[Max Chupilkin @maxchupilkin]

Over the past year, I have been experimenting with applying methods from the social sciences to the study of AI.

I created a small website/research initiative to share working papers and research blogs based on this work:
https://t.co/AxcOfAEB1E

[Max Chupilkin @maxchupilkin]

The Eurasian roundabout: Trade flows into Russia through the Caucasus and Central Asia” by @BJavorcik , @alplekhanov , and myself is now out in the European Economic Review!

The temporary free-access link to the paper is here:

https://t.co/96GU2yUMEb

[Max Chupilkin @maxchupilkin]

The link to the paper: https://t.co/5v3sgxniFw (7/7)

[Max Chupilkin @maxchupilkin]

How do LLMs evaluate economic policy? They care about unemployment, inequality, environment, and financial stability. Less so about growth, inflation, and government debt. A small exercise. (1/7)

[Max Chupilkin @maxchupilkin]

What causes LLMs to provide such consistent scoring? Similar training data? Higher level instructions? To be discovered. (6/7)

[CEPR @cepr_org]

New CEPR Discussion Paper - DP20471
Exorbitant Privilege and Economic Sanctions
@maxchupilkin, @BJavorcik, Aleksandra Peeva @HumboldtUni, @alplekhanov
https://t.co/EDY9mo9IPt
#CEPR_IMF #CEPR_ITRE #EconTwitter

[Max Chupilkin @maxchupilkin]

The paper is called “The Prompt War: How AI Decides on a Military Intervention” and can be found here: https://t.co/fc728ZVCNY (10/10)

[Max Chupilkin @maxchupilkin]

How does AI decide whether to go to war? A short experimental paper done as part of my DPhil and in strictly personal capacity. (1/10)

[Max Chupilkin @maxchupilkin]

This is an early working paper and I would be very grateful for any feedback/comments! (9/10)