𝗟𝗮𝘂𝗻𝗰𝗵𝗶𝗻𝗴 𝘁𝗵𝗲 𝗣𝗲𝗿𝗽𝗹𝗲𝘅𝗶𝘁𝘆 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗙𝗲𝗹𝗹𝗼𝘄𝘀𝗵𝗶𝗽 If you’re in STEM and exploring a transition into AI research, please consider joining us to push the frontier of intelligence systems! 𝗙𝗲𝗹𝗹𝗼𝘄𝘀𝗵𝗶𝗽: perplexity.ai/hub/research/f… 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻: jobs.ashbyhq.com/Perplexity/…

I’m once again on the program committee for the Workshop on Platform Analytics (WoPA), taking place April 2–3 at UC San Diego this upcoming year. It’s a great workshop, and I encourage anyone working on platforms to submit a paper! The submission deadline is December 6, 2026. platformanalytics.org/

PhD was my backup option :)

🚨🚨🚨 The call for papers for the 5th edition of the Workshop on Platform Analytics is open! Deadline: December 6: platformanalytics.org/

Worst post of the century

New substack from R.E.A.L.: Downtown Chicago office employment GREW 10.8% between 2019 and 2023 — from 359,000 to 398,000 jobs. However, downtown L ridership is still down 35%. These two facts are hard to reconcile, so we looked at the station level. Chicago has published daily turnstile counts for all 143 L stations since 2001, which we merged with Census LODES data on where jobs sit, block by block. The stations surrounded by the most office jobs are running at 58% of their 2019 traffic. The stations with the least office exposure are at 70%. A 12-point gap, on the same rail system, in the same downtown, in the same year. This isn't a story about jobs leaving the Loop. The jobs are still there. The people holding them are showing up two or three days a week instead of five, and that shortfall lands on everything downstream of the turnstile: lunch counters, transit fare revenue, ground-floor retail, and property assessments. Full analysis: realab.blog/p/what-urban-rai…

FYI, I've finished my year at Amazon and am visiting Stanford @DigEconLab for the year. If you're on campus and interested in the intersection of AI and markets let me know.

Economists: totally fine with RAs and predocs providing substantial assistance on a paper without being coauthors. Also economists: using AI to help with the writing is unacceptable.

Dear Econ community, I have a favour to ask. If you teach one of my papers in your courses, or include it in a reading list, could you please let me know? DM/email/comment. Thank you! 🙏

Excited to be co-organizing @CODEConference again this year Nov. 13-14 at MIT! Submissions due Sept. 13. We’re especially keen on work on new metrics for complex/AI products, methods for uncovering causal mechanisms, and experiment design under "treatment abundance." Links 👇

Pretty big break happening in academia between the “AI is banned for reviews” journals and the “AI is mandatory for reviews” journals.

This is happening so fast! Had some thoughts a couple of months ago about where we are headed and what it means for the review system currently in place: realab.blog/p/some-thoughts-…

Even if the criticism of Refine was unfair, publicly criticizing Rohan's paper in retaliation was completely unnecessary.

Waiting for LLMs to solve the P vs. NP problem in a few days.

🚨🚨 New REAL post 🚨🚨 A Metro-by-Metro Measure of the U.S. Housing Supply-Demand Gap The U.S. housing shortage is not one national problem. Using ACS data from 2019–2024, we measure the housing gap metro by metro. Across 105 large growing metros, new supply covered only ~85% of new household growth. realab.blog/p/a-metro-by-met…

Messi's vision of the game is impressive.

Really excited about this!

Economists checking whether other economists' writing is AI-generated is (another) low point for the profession.

🚨 New working paper 🚨 We looked at what happened to foreign sellers on Etsy after the U.S. suspended duty-free de minimis exemption in 2025.

New working paper with Pengxiang (Shawn) Zhou and @GoliAlii : "LLMs as Gatekeepers: Source Concentration, Factual Quality, and Political Slant in Information Search." (dadepro.github.io/papers/llm…) More and more, people ask ChatGPT, Perplexity, Gemini, Claude, or Grok instead of searching. That means these tools increasingly decide which news outlets and media sources people actually see. Focusing on the news and media ecosystem, we looked at how AI compares to regular search — and what happens as publishers start blocking AI crawlers. A few things we found: 1. AI platforms send traffic to a far more concentrated set of sources than organic search (3–8× higher HHI), but currently of higher average factual quality and comparable political slant. 2. The publishers blocking AI crawlers are disproportionately the largest, most factually reliable, and centrist-to-left outlets. Their exit shifts the available pool toward higher concentration, lower factual quality, and rightward slant. 3. Blocking Perplexity's crawler cuts a domain's referral traffic by 46–48%. That traffic flows back to the largest remaining outlets, reinforcing concentration rather than diluting it. 4. The result: redistribution reinforces concentration while failing to undo the shift toward less factual and more right-leaning publishers.