🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Why fund highly novel research? It can improve the evolution of knowledge by guiding future researchers. We propose a model in which researchers decide which questions to address and at what intensity to search for the answer based on existing knowledge. econometricsociety.org/publications/e…

New test: Operator vs Claude taking a B2B CEO survey. Both give sensible answers, but Operator dumps text all at once and Claude streams in chunks. Typing like a human is hard!

🚨 New data alert! Curious about how people really understand inflation—its causes, impacts & what governments should do about it? 📊 You can dive into the data from our project "People's understanding of inflation" here: socialeconomicslab.org/research/publi… Excited to see new analyses!

Today's "DeepSeek selloff" in the stock market -- attributed to DeepSeek V3/R1 disrupting the tech ecosystem -- is another sign that the application layer is a great place to be. The foundation model layer being hyper-competitive is great for people building applications.

Summer Institute in Bounded Rationality, Berlin, 2025 decisionsciencenews.com/?p=7210

Our paper was accepted to ICLR 2025. Check out the latest version below. openreview.net/forum?id=Tn8EQ…

Why permute when you can cheaply permute?

It gets harder and harder to identify experts through tests because of cheating with AI. Forecasting performance will become the primary metric on which experts compete. Those who consistently outforecast AI are either psychic or possess valuable knowledge and judgment.

While there are big crowds on the "AI hurts learning" train, let's realize that's like saying "computers hurt learning" Two papers where unaided learners do better after AI interaction: arxiv.org/abs/2502.02880 ssrn.com/abstract=46416…

Attention NYC undergrads: Applications are open for our 12th annual Data Science Summer school at Microsoft Research NYC! Apply here by April 15th: bit.ly/3pCQENh

🚨Out today in PNAS PNASNews🚨 pnas.org/doi/10.1073/pn… Why do people overestimate the size of politically relevant groups (immigrant, LGBTQ, Jewish) and quantities (% of budget spent on foreign aid, % of refugees that are criminals)? We analyze 100k estimates to find out🧵👇

The Society for Judgment and Decision Making is pleased to announce that the latest newsletter is ready for download: sjdm.org/newsletters/

What happens when a static benchmark comes to life? ✨Introducing ChatBench, a large-scale user study where we *converted* MMLU questions into thousands of user-AI conversations. Then, we trained a user simulator on ChatBench to generate user-AI outcomes on unseen questions. 1/

🚨 Predoc at Microsoft Research 🚨 Do you have strong front-end programming and statistics skills but need research experience before heading off to graduate school for a computational social science PhD? Apply now for this summer! microsoft.com/en-us/research…

Best job ever :)

A nice discussion of why it is so hard to outpredict models. References one of my favorite papers: "Improving out-of-population prediction. The complementary effects of model assistance and judgmental bootstrapping" dangoldstein.com/papers/Hardy_e…

New from Microsoft Research and Harvard Business School colleagues

Check out the Stanford AI Index Report: AI Index It's the best one-stop-shop for credible, current data on AI trends.

The Society for Judgment and Decision Making is pleased to announce that the latest newsletter is ready for download: sjdm.org/newsletters/ This issue contains announcements, conferences, jobs, and a new section entitled HotFresh Research News! decisionsciencenews.com/?p=7220