Greg Veramendi

I read @itsafronomics.bsky.social new book “The Double Tax” this afternoon. It’s engaging and one of the best popular ways to bring social science to the public and policy that I’ve read in a long while. Black women are uniquely disadvantaged in our economy. 😍 she gave Janelle James her flowers!

Saying the quiet part out loud, again.

Please join me in signing and sharing:

For Day 2 of our celebration of work using the Census Tree we highlight @lukestein.com & co's work on the gendered impacts of perceived skin tone. They find that among African Amer. sisters, women perceived to be darker-skinned were disadvantaged.

*They also cite a great QJE pub by Lisa Cook et al*

SI 2022 Methods Lectures - Empirical Bayes Methods, Theory and Application

Another good place to start: www.nber.org/conferences/...

Worth a read if you don't already follow Phil. TLDR:

Conservative Man Proudly Frightened Of Everything FLOWER MOUND, TX—Condemning the “woke left” for what he called the “modern evisceration of masculinity,” local conservative man Hank Daniels confirmed Monday that he was never going to stop being prou...

theonion.com/conservative...

what gif pops up when you type your name?

Hey Luna, what do you mean when you say "i got too excited"?

Stable genius!

"Young political vigilantes" have historically been very useful to dictatorships--from Berlin to Beijing--because they gleefully engage in mass violent destruction of social institutions whose value and purpose they do not understand.

Ex: Mao's Red Guard:
www.theguardian.com/world/2023/j...

i understand the american people on average not reading history books what with the failing education system etc but what the fuck is the university leadership's excuse

📝New @cepeo-ucl.bsky.social working paper

Why are students from elite high schools much more likely to go to high ranked university courses than equally qualified students from the state sector? 🤔

w @opmc1.bsky.social @lindseymacmillan.bsky.social & Jo Blanden

econpapers.repec.org/RePEc:ucl:ce...

When an English person shares how they feel about good news: "I feel, um, unnecessarily emotional about it."

from Thames Water documentary. Worth a watch.

This 👇

Recognising the growing difference between predicted grades and achieved grades over time, UCAS are piloting personalised reports to schools

These will show schools how their UCAS predicted grades compare to achieved results

www.ucas.com/corporate/ne...

Author of "One Long Night: A Global History of Concentration Camps"
👇

I Randomly Decided To Pay Off A School’s Lunch Debt. Then Something Incredible Happened. “I immediately wondered if this was real.”

Great article on school lunch debt relief

www.huffpost.com/entry/utah-s...

YouTube video by UNSW eLearning Modern Day Oracles or Bullshit Machines? Seminar with Prof. Carl Bergstrom, University of Washington

Yesterday, @carlbergstrom.com presented his course "Modern-day Oracles or BS Machines? How to thrive in a ChatGPT world". Neat way to make students aware of the capacities and limitations of LLMs👇🎰

Recording: www.youtube.com/watch?v=TZC0...

Course: thebullshitmachines.com

@unswbabs.bsky.social

Is it better to have standardised exams or teacher assessments?

Myself and my colleague @opmc1.bsky.social have read the literature so you don't have to. Our @iza.org world of labour summarises below #econsky

This study by MIT authors shows students using LLMs exhibit decreased cognitive activity. I admit that it is a small sample and confirms my priors. Lots of crazy shit happening atm, but this still worries me a lot as an educator and when thinking about the future. www.brainonllm.com

Chatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.

They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling.

This is the third story I've read in a month about how AI chatbots are leading people into psychological crises.

Gift link

I wonder what this guy is thinking as he takes a potshot at a journalists. Where do they find people happy to violate their oath?

The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity Parshin Shojaee∗† Iman Mirzadeh∗ Keivan Alizadeh Maxwell Horton Samy Bengio Mehrdad Farajtabar Apple Abstract Recent generations of frontier language models have introduced Large Reasoning Models (LRMs) that generate detailed thinking processes before providing answers. While these models demonstrate improved performance on reasoning benchmarks, their fundamental capabilities, scal- ing properties, and limitations remain insufficiently understood. Current evaluations primarily fo- cus on established mathematical and coding benchmarks, emphasizing final answer accuracy. How- ever, this evaluation paradigm often suffers from data contamination and does not provide insights into the reasoning traces’ structure and quality. In this work, we systematically investigate these gaps with the help of controllable puzzle environments that allow precise manipulation of composi- tional complexity while maintaining consistent logical structures. This setup enables the analysis of not only final answers but also the internal reasoning traces, offering insights into how LRMs “think”. Through extensive experimentation across diverse puzzles, we show that frontier LRMs face a complete accuracy collapse beyond certain complexities. Moreover, they exhibit a counter- intuitive scaling limit: their reasoning effort increases with problem complexity up to a point, then declines despite having an adequate token budget. By comparing LRMs with their standard LLM counterparts under equivalent inference compute, we identify three performance regimes: (1) low- complexity tasks where standard models surprisingly outperform LRMs, (2) medium-complexity tasks where additional thinking in LRMs demonstrates advantage, and (3) high-complexity tasks where both models experience complete collapse. We found that LRMs have limitations in exact computation: they fail to use explicit …

If I have time I'll put together a more detailed thread tomorrow, but for now, I think this new paper about limitations of Chain-of-Thought models could be quite important. Worth a look if you're interested in these sorts of things.

ml-site.cdn-apple.com/papers/the-i...

@luhrmann.bsky.social @absidd.bsky.social @gafap.bsky.social

@causalinf.bsky.social @pfallesen.bsky.social @paulhufe.net @pietrobiroli.bsky.social

Postdoctoral Research Associate in Youth Mental Health at Royal Holloway, University of London Discover an exciting academic career path as a Postdoctoral Research Associate in Youth Mental Health at jobs.ac.uk. Don't miss out on this job opportunity - apply today!

I am happy to share a 3-year postdoc opening. My fantastic colleagues @kailirimfeld.bsky.social and Dan Anderberg recently won a grant to work on youth mental health at the intersection of psychology, genetics, and education. It is open to economics PhDs, among other fields. #EconSky #econ_ra

This is such a great demonstration of how LLMs function.

Gen Z Doesn't Want to Start a Bar Tab

Above the fold at

The Guardian: Ukraine launches major drone attack against Russian bombers

Wall Street Journal: Breaking
FBI Investigating Possible Terror Attack in Boulder, Colorado

New York Times: