People in some places in the US have more long ties — ties between people without mutual contacts. This is a map from our newly published paper: https://t.co/GJIi5AVXW5 Why does this matter? How do people end up with these different network structures? 🧵

📢 Call for abstracts for the Conference on Digital Experimentation (CODE@MIT) 📨Submit your best work by *Sept 12th*! https://t.co/RX2ySBwKEf Here are three topics we're particularly interested in getting more submissions on this year:

We're recruiting a tenure-track faculty member to @MITSloan! https://t.co/aEFCcMhsLc We're looking for a researcher working in quantitative marketing, broadly considered. Our existing faculty include experts on privacy, mathematical psych, networks, and applications of ML & AI.

🚨 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 @OpenRouterAI to take a first look. cc

PSA: I am only hearing rumors, but they are circulating that NSF social and behavioral sciences is next on chopping block, maybe this week, maybe a bit longer. Here is what I wish I’d done and I recommend you do if you have a grant:

Associated Press reporters at the White House were barred from an Oval Office event today with the President - because the AP won't use Trump's new name 'Gulf of America' https://t.co/CKCqm2VRDh

AMFAR has published a report on how the foreign aid freeze affects PEPFAR, which actually still hasn't resumed. They estimate that hundreds of thousands of people lose access to HIV antivirals and 1500 babies are expected to contract HIV — *each day* 🧵

In a large US sample, approximately half of all women discontinue antidepressant usage during pregnancy (no change in usage for spouses). This is in a new paper that I just published with @claireeboone and @ccolina_pe that is primarily descriptive.

I'm hiring for two front-line engineering manager roles in PyTorch: someone to manage TorchTune and someone to manage TorchRec. Link in follow up post.

chain of thought is hilarious (distilled deepseek 14b running via ollama)

This should be a national scandal: Today, the LA County Board of Supervisors unanimously passed a resolution asking the state to suspend various legal protections for property owners in burned areas—their right to build a duplex or ADU, to reasonable permit timelines, and more.

You will get FOMO if you don't listen to our new pod about the social media trap. cc @SBenzell

We've revamped the **fect** package in R and integrated **gsynth** into it. The original **gsynth** is retired. The new package implements a series of imputation estimators, including FEct, IFEct, MC & Gsynth. New user manual: https://t.co/YEWG6LjW4R

It's a good time to mention that @SBenzell and I have a podcast called Justified Posteriors discussing the economics of AI and other technologies. Our latest episode considers @timobres's paper on AI and Aggregate. Link below and available at all fine podcast purveyors.

A few of us were at this event where Dwarkesh interviewed Tyler. I love that he got feisty here - right before the interview, some EA guys were trying to convince him of some craziness and he was giving them the business. But all good - important ideas discussed by smart people!

Super interesting post about a real-life implementation of a social impact bond. Kudos to @Arnold_Ventures for getting this done. A shame the CBT intervention was underwhelming, but extremely useful to know it!

Guys. We — as in Californian voters who voted in 1988 — voted for a system where every time an insurance company wants to raise rates for automobiles or homes, they have to have a public hearing and an elected insurance commissioner has to approve that rate increase.

i broadly agree that there is a common language among the quantitative social sciences, but I think this growth hasn’t just been “hail the conquering econ heroes.” 1) econ became more influential in QSS as it shed its more structural tendencies and embraced the reduced form…

Feeling a bit surreal—just heard that JP Morgan and Morgan Stanley issued equity research notes on our paper with @hristakevas and Leo Feler. This is probably the closest my research will ever come to influencing markets! 📉📈

Apropos of nothing, I recently wrote on my blog about how advertising is the model organism of studying how organizations make decisions. Link below!

This is pretty cool, but also, imagine if you could collect that same level of property tax revenue from that lot without waiting for someone to build something on it—and without discouraging them from doing so.