rezasalehnejad.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

Starting math camp? 📉📈 @pmichaillat.bsky.social has you covered. pascalmichaillat.org/c3/

submitted 194 days ago • 0 comments

there is something so quintessentially australian (in fact particularly with the smaller feeling of adelaide) about terry tao having a jam sandwich for breakfast youtu.be/MXJ-zpJeY3E

submitted 2 days ago • 2 comments

For everyone thinking that we simply tell doctors they should not use LLMs for decision making and that they will listen to that, just look at these numbers: "76% of respondents reported using general-purpose LLMs in clinical decision-making" www.fiercehealthcare.com/special-repo...

submitted 4 days ago • 3 comments

In partnership with @schmidtsciences.bsky.social, @simonhrjohnson.bsky.social, David Autor and I are seeking funding proposals from researchers conducting innovative field experiments on the labor economics of frontier AI. Expressions of interest due March 31. shapingwork.mit.edu/call-for-pro...

submitted 4 days ago • 0 comments

This was a very enjoyable collaboration with @3blue1brown.bsky.social . Tanya Klowden and I are continuing to work on developing this material into a popular science book; we will share our progress on our Instagram at www.instagram.com/cosmic_dista...

submitted 6 days ago • 3 comments

This is a great paper by Mitsuru Igami framing breakthrough AI algorithms in the familiar logic of structural economic model estimation. doi.org/10.1093/ectj... But by implication, if structural models behave like board games, such as Chess and Go, how high can their degree of realism really be?

submitted 8 days ago • 4 comments

@lionelhenry.bsky.social and I are so excited to finally announce Air - an extremely fast R code formatter! 🎉 With Air, you'll never need to worry about styling your #rstats code ever again. All you need to do is save, and Air takes care of the rest. www.tidyverse.org/blog/2025/02...

submitted 8 days ago • 19 comments

As AI models grow in size, their values seem to increasingly converge on the same preferences, whether the model is made by OpenAI or Musk’s X or China’s DeepSeek Not everyone will like resulting AI preferences, so value engineering is likely to become a topic of discussion arxiv.org/abs/2502.08640

submitted 7 days ago • 4 comments

@prashantgarg.bsky.social & @trfetzer.com introduce a methodology that uses #LLMs to process and synthesise tens of thousands of #economics papers into a 'causal knowledge graph', which identifies which economic concepts are linked. cepr.org/voxeu/column... #EconSky

submitted 10 days ago • 3 comments

Print this out and put it on your pinboard! Most compact and at the same time clear and intuitive figure on possible associations between predictors and outcome (or variables). (source: journals.sagepub.com/doi/10.1177/..., by @dingdingpeng.the100.ci and @achatton.bsky.social)

submitted 11 days ago • 7 comments

was tinkering with some OCR and decided to update an old econ history OCR example. What a difference a couple of years makes: old: github.com/apoorvalal/h... new: github.com/apoorvalal/h...

submitted 12 days ago • 5 comments

Courtesy of @johnjhorton.bsky.social 🎓🍎

submitted 15 days ago • 0 comments

Any experience with text-to-voice AI tools for research papers such as Listening? Are they useful? www.listening.com

submitted 15 days ago • 5 comments

My piece for @productivity.bsky.social about the importance of technicians. We won't deliver growth and opporotunity without enough of them in our labs and across universities and schools. It's an urgent conversation to have www.productivity.ac.uk/news/the-rol...

submitted 15 days ago • 1 comment

1/ 📢 New paper alert! 📄 Together with Vitalii Tubdenov, we propose a measure of the "Average Period of Production" (APP) to capture the temporal dimension of production processes. It is inspired by Böhm-Bawerk’s capital theory.

submitted 16 days ago • 2 comments

Deadline for submission for the NBER Organizational Economics Working Group is tomorrow, send us your best work!

submitted 16 days ago • 0 comments

Half of “prompt engineering” was actually just prompting LLMs to act like Reasoners before the labs realized that was a thing. (Chain of thought/think step-by-step was the first powerful prompting technique that was discovered, now Reasoners do it automatically

submitted 16 days ago • 3 comments

www.youtube.com/watch?v=8tMP...

submitted 17 days ago • 0 comments

There is a lot of important stuff in this new paper by Anthropic that shows how people are actually using Claude. 1) The tasks that people are asking AI to do are some of the highest-value (& often intellectually challenging) 2) Adoption is uneven, but many highly compensated fields already are high

submitted 18 days ago • 1 comment

What would be the impact of Trump's announced 25% tariff on all steel and aluminium imports? Lydia Cox's (coxlydia.com) job-market paper, now an R&R at the American Economic Review, provides good guidance. Latest version is here (coxlydia.com/papers/cox_s...).

submitted 18 days ago • 1 comment

1/ 🚨 New working paper! 🚨 We investigate a novel determinant of labor market power: a firm's product scope. Do multiproduct firms (MPFs) pay lower wages than single-product firms (SPFs)? A 🧵 on our key findings.

submitted 18 days ago • 1 comment

This is a super interesting paper cognitive endurance.

submitted 61 days ago • 8 comments

🚨NEW PAPER ALERT🚨 Does job training reduce workers’ automation risk? And does its effectiveness vary by country, gender, or training type? We study these questions in a new WP (w/ Oliver Falck, Mo Guo @christinalanger.bsky.social & @lindlacher.bsky.social). #EconSky CESifo WP: bit.ly/3D4CDS5 🧵 👇

submitted 82 days ago • 2 comments

You prob follow @3blue1brown.bsky.social already. This latest video is great. I watched it with my son, who saw how all that middle school trig he learned solves important astronomy problems. I just wish he had learned it with these examples instead of "how tall is the tree". youtu.be/YdOXS_9_P4U

submitted 18 days ago • 2 comments

The negative externalities of GenAI never seem to end.

submitted 19 days ago • 10 comments

AI is unlikely to speed up science by multiples acting solely as a co-pilot to humans. There are some early successes using current LLMs to autonomously research, plan & conduct experiments, including learning how to use equipment by reading documentation, with big potential rewards (& risks).

submitted 19 days ago • 5 comments

One of the biggest areas of AI hype right now is the notion that it will hyperaccelerate scientific progress. I understand why people think this — AI is already accelerating scientific _production_. 🧵

submitted 19 days ago • 4 comments

A good review of China's industrial structure: open.substack.com/pub/highcapa...

submitted 20 days ago • 1 comment

Success 🏆 Arthur Guillouzouic (CNRS/AMSE) and his co-author Antonin Bergeaud (HEC) received the 2024 ADRES Best Young Paper Award ! More info ➡️ https://buff.ly/4aRmOLC @arthurgzouic.bsky.social @cnrs.bsky.social @cnrs-dr12.bsky.social @abergeaud.bsky.social

submitted 23 days ago • 0 comments

strike fear into the hearts of data scientists with one sentence. [from the newest 3B1B featuring the one and only youtu.be/YdOXS_9_P4U'

submitted 21 days ago • 0 comments

A few implications of tricks like this: 1) We are still VERY early in the development of Reasoners 2) There is high value in understanding how humans solve problems & applying that to AI 3) Higher possibility of further exponential growth in AI capabilities as techniques for thinking traces compound

submitted 22 days ago • 1 comment

What does the world need the most? It needs more focus on how sustainable policies can become politically feasible. Partha Dasgupta is thus kicking off our conference on this topic today: www.gsb.stanford.edu/events/polit...

submitted 22 days ago • 1 comment

Final final o3 thought (you won't like it): it will be good for science! All math will need formal proofs in e.g. Lean. All empirics will need replicable one-click files from original data with docs. All papers open-source. "Trust me, I did what I said" won't be acceptable.

submitted 26 days ago • 3 comments

This paper is wild - a Stanford team shows the simplest way to make an open LLM into a reasoning model They used just 1,000 carefully curated reasoning examples & a trick where if the model tries to stop thinking, they append "Wait" to force it to continue. Near o1 at math. arxiv.org/pdf/2501.19393

submitted 22 days ago • 8 comments

Forthcoming article by @janeeckhout.bsky.social "The Value and Profits of Firms" @eeanews.bsky.social doi.org/10.1093/jeea...

submitted 22 days ago • 2 comments

Like Vincent says, this is a big release with a shedload of new features. Same small footprint, though ;-) Here’s quick rundown of some of my favourites... #rstats #dataviz

submitted 22 days ago • 2 comments

A concise explainer of “reasoning” by large language models, such as those of DeepSeek:

submitted 22 days ago • 2 comments

Not an easy chart to read, but worth spending a minute on. It tells you a few key things: 1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper 2) The cost of GPT-4 capability dropped 1,000 fold in 18 months 3) Pace of improvement is swift

submitted 23 days ago • 8 comments

10 short videos about LLM infrastructure to help you appreciate Pages 12-18 of the DeepSeek-v3 paper (arxiv.org/abs/2412.19437) www.youtube.com/watch?v=76gu...

submitted 53 days ago • 4 comments

What to know about DeepSeek youtu.be/0eMzc-WnBfQ?... In which we attempt to figure out MoE, o1, scaling, tech reporting, modern semiconductors, microeconomics, and international geopolitics.

submitted 25 days ago • 1 comment

Highly recommend watching

submitted 23 days ago • 0 comments

🚨New Paper 🎻 On Bob Dylan: A Computational Perspective In this work, I use LLM (o3) to explore how Dylan’s ever-evolving lyrical themes break expectations and challenge convention (i.e., dishabituation). This method can be used an artist’s cultural and creative evolution in general.

submitted 24 days ago • 3 comments

Dick Nelson was a giant in the economics of technological change. Incredible work including the very influential Simple Economics of Basic Scientific Research. Potentially the very first of appearance of the "simple economics" genre. www.journals.uchicago.edu/doi/abs/10.1...

submitted 28 days ago • 1 comment

Wow, that's exciting and a great public good! "The Global Macro Database: A New International Macroeconomic Dataset" by Karsten Müller, Chenzi Xu, Mohamed Lehbib, and Ziliang Chen. www.globalmacrodata.com/index.html

submitted 25 days ago • 3 comments

The new OpenAI model announced today is quite wild. It is essentially Google's Deep Research idea with multistep reasoning, web search, and the o3 model underneath (as far as I know). It sometimes takes a half hour to answer. Let me show you an example. 1/x

submitted 26 days ago • 8 comments

In a big representative sample of humans against GPT-4: "the creative ideas produced by AI chatbots are rated more creative [by humans ]than those created by humans... Augmenting humans with AI improves human creativity, albeit not as much as ideas created by ChatGPT alone” docs.iza.org/dp17302.pdf

submitted 27 days ago • 5 comments

Lovely write up of Ohid's Research Policy paper on serendipity in science in Nature. www.nature.com/articles/d41...

submitted 28 days ago • 1 comment