Gord Burtch

Just updated **panelView** in R: it now visualizes multi-way fixed-effect structures as a network graph, building on earlier work by Sergio Correia. You can easily identify singletons and duplicate observations before estimation. yiqingxu.org/packages/panel…

@causalinf Learned about this recently too w.r.t. LLMs' non-deterministic output, even when temp is 0! GPU parallel processing doesn't guarantee a fixed ordering of FP ops and the small rounding errors accumulate to produce different output.

I am embarrassed to say I didn’t know jack about “big numbers” being handled differently by different programming environments on the same computer. I have avoided learning all this float stuff for so long and now I have to. “Claude make a beautiful deck abt floating numbers”

Announcing a new Claude Code feature: Remote Control. It's rolling out now to Max users in research preview. Try it with /remote-control Start local sessions from the terminal, then continue them from your phone. Take a walk, see the sun, walk your dog without losing your flow.

A good step on Prolific's part

eBay has just banned AI agents that buy products on people's behalf. As @alexolegimas and I describe in our article (link below), this behavior is a barrier to consumer friendly agentic commerce. To counteract it, we may need a right to bring your own agent.

@alexolegimas Might find this interesting / relevant. nature.com/articles/s4159…

This is the worry. It is not just sentiment around the fall of Stackoverflow: the q&a forum was a continual source of new knowledge and was likely an important source of training data for LLMs. Until AI starts reliably producing new knowledge, this is very bad news.

@jt_kerwin I would worry that the assumptions may be overlooked by many (that unequal variances cause the math to fall apart), though I realize the blog post mentions this. And, yeah, the 'oddness' of this seems like it would create more trouble for interpretation than it resolves?

If you're going to plot means and CIs to show your results, they should be 83% CIs (not 95% CIs). Why? Because then when people look at whether they overlap, the implicit statistical test has the correct size.

We've a new paper posted, measuring audience imbalance (divergent delivery) across large samples of advertiser A/B and Lift tests at Meta (only A/B tests show any). We evaluate how much imbalance can be reduced by configuring the experiment the 'right way'. Check it out!

New piece out today in the @AMJPublicHealth "Abortion Restrictions and Infant Mortality in the United States, 2018–2023" ajph.aphapublications.org/doi/abs/10.210…

What's the optimal number of steps of physical activity that is associated with multiple (9) favorable outcomes? The largest, systematic review of 57 studies, 35 cohorts. 7,000 steps is clinically meaningful; some benefit also seen for 4,000 steps cf 2,000 steps @TheLancet thelancet.com/journals/lanpu…

📜 New working paper @apo_filippas @johnjhorton & diego AI is *rapidly* changing the job application process. 📈Employers getting 1000s of applications for a single job, all seemingly well-qualified. 👥Qualified workers are finding it difficult to stand out from the sludge of AI applications. 💡Our paper offers a market-based solution for this problem: allow workers to pay to send a costly signal of interest when applying for a job. (1/n)

New piece out today in Proceedings of the National Academies of Science... take caution using LLMs as human surrogates pnas.org/doi/10.1073/pn…

How much code now comes from AI? In new work with @simone_daniotti, @xiangnan_feng & @FrankNeffke we estimate that by end-2024 about 30% of Python functions pushed by US devs on GitHub are AI- generated. Adoption is rapid but diffusion lags globally. How did we do it?

@johnjhorton My first thought was... TSA: "everyone take off your shoes please."

The "there should be a rule/law/training/office" impetus after every bad event is both a) understandable and b) how you get the kinds of organizations most people hate to be in

Forthcoming in AER: Insights: "The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment" by Erik Brynjolfsson, Avinash Collis, Daniel Deisenroth, Haritz Garro, Daley Kutzman, Asad Liaqat, and Nils Wernerfelt. aeaweb.org/articles?id=10…

tl;dr: I agree with you that the relevant treatment effect is one that incorporates the delivery optimization implemented by the platforms (in this case Meta), and so I don't think this criticism is relevant for at least 99.99% of advertisers. [2 / 2]

How do #robots impact #warehouse safety? @gburtch @BUQuestrom and coauthors @GeorgeMasonU @IEuniversity show that warehouse robotics are associated with a 40% decrease in severe #injuries but a 77% increase in non-severe injuries. doi.org/10.1177/001979…

Articles in advance 03/25 (5 of 12) Frontiers: "Recommending What to Search: Sales Volume and Consumption Diversity Effects of a Query ...." by Shuang Zheng; Siliang (Jack) Tong; Hyeokkoo Eric Kwon; Gordon Burtch; Xianneng Li pubsonline.informs.org/doi/abs/10.128…

@MorganBarrettX @ATabarrok Really? reddit.com/r/funny/s/et0B…

Had a heart attack at breakfast this morning when the founder (young guy, dropout, just raised $5m) asked the waitress for a quickie I’m staring at the menu - then it clicks “Sorry, it’s pronounced, Quiche”

new: Microsoft and OpenAI are investigating whether a group linked to China's DeepSeek obtained OpenAI's data exclusively reported by @dinabass @shiringhaffary bloomberg.com/news/articles/…