Lightnews — Scholar-powered news

Reposted by Andrey Fradkin

Brett Hollenbeck @bretthollenbeck.com · Jul 11

🚨Free data alert!! 🚨 Please share.

Large new dataset of Amazon product reviews, including full text and photos and product characteristics, with individual *reviews labeled as fake reviews*.

I believe this is the first publicly available data of this kind.

github.com/bretthollenb...

1 43 130

Andrey Fradkin @andreyfradkin.bsky.social · May 13

Thanks Erik!!!

1

Andrey Fradkin @andreyfradkin.bsky.social · Apr 22

3. There is substantial multi-homing, with users of the same app employing a mix of models.

This is still very preliminary work, so comments are welcome!

1

Andrey Fradkin @andreyfradkin.bsky.social · Apr 22

2. New model releases differ in the degree to which they cause substitution from existing models or expand the market.

1 1

Andrey Fradkin @andreyfradkin.bsky.social · Apr 22

Through the use of three case studies of model releases (Sonnet 3.7, Gemini 2.0 Flash, Gemini 2.5 Pro), I document three stylized facts:
1. New models are adopted quickly, with increased demand stabilizing within a few weeks.

1 2

Andrey Fradkin @andreyfradkin.bsky.social · Apr 22

🚨 New working paper 🚨
Demand for LLMs: Descriptive Evidence on
Substitution, Market Expansion, and Multi-Homing

A key question for the business of AI is the extent to which LLMs are differentiated from each other. I use data from OpenRouter to take a first look.

andreyfradkin.com/assets/deman...

1 8 18

Reposted by Andrey Fradkin

Erzo F.P. Luttmer @erzoluttmer.bsky.social · Apr 19

Congratulations to Frank Verboven (@frankverboven.bsky.social) for being appointed as AER coeditor, handling papers in empirical IO and related fields. His term starts on July 1. I’m delighted to have him on our team. @aeajournals.bsky.social #EconSky

1 11 53

Andrey Fradkin @andreyfradkin.bsky.social · Apr 7

A new episode of Justified Posteriors just dropped. We discuss the recent papers of Chad Jones about existential risk and AI safety investment. Check it out!

open.substack.com/pub/empiricr...

How much should we invest in AI safety?

We discuss Chad Jones' papers on existential risk and investment in society.

open.substack.com

4

Reposted by Andrey Fradkin

Alex Imas @aleximas.bsky.social · Apr 6

Great point from @andreyfradkin.bsky.social: as a movement’s policies become more incoherent and harmful, the quality of “experts” who are willing to publicly defend it decreases precipitously.

1 6 25

Andrey Fradkin @andreyfradkin.bsky.social · Apr 5

It was fun engaging with the EconCS community!

Kira Goldner @kiragoldner.bsky.social · Apr 4

David Laibson on present bias and @andreyfradkin.bsky.social on experiments regarding dark patterns in cookie settings.

5

Andrey Fradkin @andreyfradkin.bsky.social · Mar 11

In the latest episode of our podcast, Justified Posteriors, we discuss whether interest rates should rise in anticipation of AGI (as predicted by @basilhalperin.com). Our priors are quite different! Do check it out.

empiricrafting.substack.com/p/if-the-rob...

If the Robots Are Coming, Why Aren't Interest Rates Higher?

Transformative AI, Existential Risk, and Real Interest Rates

empiricrafting.substack.com

1 3

Andrey Fradkin @andreyfradkin.bsky.social · Jan 24

I'm just using their GUI. The papers submitted to MS are mostly available online already, so am not worried about leakage.

1

Andrey Fradkin @andreyfradkin.bsky.social · Jan 24

Write a referee report for Management Science based on this paper. Take the perspective of an economist who cares about causal methods.

1 1

Andrey Fradkin @andreyfradkin.bsky.social · Jan 24

But what if we're interested in studying AIs?

1 1

Andrey Fradkin @andreyfradkin.bsky.social · Jan 24

Can report that Deepseek writes perfectly serviceable referee reports of empirical papers submitted to Management Science. Note, I will still write my own, but you should update accordingly and ask it to provide feedback on your paper prior to submission.

1 2 12

Andrey Fradkin @andreyfradkin.bsky.social · Jan 18

I think your example of the rent curve is great, but doesn't represent much of social science research (based on things I referee). A lot of (most?) empirical papers are driven by hammer and nail thinking (something changed, run a DID), and not by careful descriptive work followed by modeling.

2 7

Andrey Fradkin @andreyfradkin.bsky.social · Jan 17

Perhaps the best David Lynch appearance was on Louie.
"Get that belly moving son or you're out."

www.youtube.com/watch?v=HlEJ...

Louis C.K - not that kind of funny

YouTube video by mapusti

www.youtube.com

3

Andrey Fradkin @andreyfradkin.bsky.social · Jan 11

Thanks, right back at ya!

1

Andrey Fradkin @andreyfradkin.bsky.social · Jan 11

It's a good time to mention that Seth Benzell and I have a podcast called Justified Posteriors discussing the economics of AI and other technologies. Our latest episode considers @timobres.bsky.social's paper on AI and Aggregate Growth.
empiricrafting.substack.com/p/beyond-tas...

Beyond Task Replacement

Artificial Intelligence Technologies and Aggregate Growth Prospects by Tim Bresnahan

empiricrafting.substack.com

5 13

Reposted by Andrey Fradkin

Aaron Roth @aaroth.bsky.social · Dec 30

EC 2025 will be held at Stanford from July 7-12. Itai Ashlagi and I are the chairs. The abstract deadline is February 3, and the paper deadline is February 10. The scope is inclusive of many topics across CS, economics, and operations research. Submit your best work!

1 20 39

Andrey Fradkin @andreyfradkin.bsky.social · Dec 21

Write down a vague idea of a model and it will write the math. Tell it to translate the math into code. Tell it to make it run quickly and to try other variations. I think the key speedups are actually in steps 2 and 3, which prevented most from using them.

1 7

Andrey Fradkin @andreyfradkin.bsky.social · Dec 20

Will be a huge complement to most of it, but the efficiency improvements for tasks in IO for me have been immense. So maybe IO / anything with structural modeling. But also, it might now be cheap to include a structural model in any type of paper, making econ papers even longer.

1 10

Reposted by Andrey Fradkin

Avi Goldfarb @avigoldfarb.bsky.social · Dec 6

NBER Digital Economics and AI meeting at Stanford. Deadline December 16. www.nber.org/conferences/...

Digital Economics and AI Meeting, Spring 2025

www.nber.org

9 11

Andrey Fradkin @andreyfradkin.bsky.social · Dec 4

It's the best time ever to experiment with new coding tools. Coding with o1-preview is very effective for hard research tasks (structural modeling with pyTorch, functions for custom latex tables, computing Hessians). 4o/Claude do well on simpler data cleaning with just a little guidance.

Paul Goldsmith-Pinkham @paulgp.com · Dec 4

Just a reminder that @gmcd.bsky.social and @kylefbutts.bsky.social have made an incredible public good describing how to use data.table and fixest to encourage moving from Stata to R!

stata2r.github.io

Translating Stata to R

Learning R coming from Stata

stata2r.github.io

1 5 29

Andrey Fradkin @andreyfradkin.bsky.social · Nov 30

I should mention that I'm a fan of one of the authors, Byrne Hobart, and his newsletter The Diff, which provides interesting analysis at the intersection of tech and finance.

3