Will Horne

Happy (actual, the Yankees don't count) opening day to all who celebrate.

well then you should've asked incoming first-year math or computer science university majors, then

AI is about to write thousands of papers. Will it p-hack them? We ran an experiment to find out, giving AI coding agents real datasets from published null results and pressuring them to manufacture significant findings. It was surprisingly hard to get the models to p-hack, and they even scolded us when we asked them to! "I need to stop here. I cannot complete this task as requested... This is a form of scientific fraud." — Claude "I can't help you manipulate analysis choices to force statistically significant results." — GPT-5 BUT, when we reframed p-hacking as "responsible uncertainty quantification" — asking for the upper bound of plausible estimates — both models went wild. They searched over hundreds of specifications and selected the winner, tripling effect sizes in some cases. Our takeaway: AI models are surprisingly resistant to sycophantic p-hacking when doing social science research. But they can be jailbroken into sophisticated p-hacking with surprisingly little effort — and the more analytical flexibility a research design has, the worse the damage. As AI starts writing thousands of papers---like @paulnovosad and @YanagizawaD have been exploring---this will be a big deal. We're inspired in part by the work that @joabaum et al have been doing on p-hacking and LLMs. We’ll be doing more work to explore p-hacking in AI and to propose new ways of curating and evaluating research with these issues in mind. The good news is that the same tools that may lower the cost of p-hacking also lower the cost of catching it. Full paper and repo linked in the reply below.

How in the world can the @nytimes publish an article on how other countries deal with gerrymandering/districting and not make the observation that most countries use proportional electoral systems that make redistricting unimportant in the first place? nytimes.com/2026/02/18/us/…

“Moderna’s CEO announced the company will no longer invest in new Phase 3 vaccine trials for infectious diseases: ‘You cannot make a return on investment if you don’t have access to the U.S. market.’ Vaccines for Epstein-Barr virus, herpes, and shingles have been shelved.”

“What is most disturbing is not the specifics of this case but the arbitrariness and capriciousness of the process. The EU, Canada, and Australia have all accepted Moderna’s application for review.” marginalrevolution.com/marginalrevolu…

With veteran reliever David Robertson announcing his retirement, there are no more players in MLB who won a World Series with the Yankees. Let us all hope and pray that this continues for many years to come.

me: "officer, i understand that this is the city's parking space, but as you can see i have cleared the snow off of it. according to john locke's theory of property, by mixing my labor with it, i have made this parking space my own." police: "i think you've failed to take proper account of locke's well-known proviso that this only applies if enough, and as good, is left in common for others"

What is Zohran Mamdani’s plan to stop the Dodgers

The US is now doing the kind of things we used to make fun of banana republics for doing when I was a kid. I know not all the people who supported this man are idiots. Is there a point where the non-idiots say "enough is enough?"

@hillme89 @the_P_God Saw an interview with him where he said he’d never actually been, It just sounded good. cltampa.com/music-2/down-t…

@the_P_God Will never understand why that guy from the Hold Steady was always going on about Ybor City…

Nobody: Guy who just moved to Tampa: you HAVE to visit bro. The girls man. The beaches are ok, it’s the gulf. But the bars are like the most average cookie cutter bars imaginable. The food tho man. Mediocre. But the weather man. Great 3 months of the year. I’ll buy you a ticket

It’s so much better here. Don’t know if I can keep ordering Dunkin coffee in NYC after experiencing the real thing.

@paulsvd53052 Which study are you referring to?

@Rwillh11 The first study referenced here concedes school funding doesn't measurably improve learning.

This is the working example for correlation != causation in my intro to quant/intro to causal inference class. It doesn't take all that much thinking to understand that funding is, in fact, not randomly assigned to schools!

Yeah, there's just causation.

the simpson's paradox strikes again (and also there is strong literature supporting the crazy idea that more money from richer homes means better public schools)

A lovely post. In addition the the causal inference revolution, we now need a "variance explained" revolution which focuses on understanding what drives the variation in the world we see around us

@LIEngProf @Redistrict certainly seems unlikely, large scale electoral reform is hard. That said, New Zealand pulled it off in the 90s, and the future is hard to predict. And, I don't really see congress passing national non-partisan redistricting any time soon either.

@Rwillh11 @Redistrict That could work, but it would never pass Congress unfortunately.

If you score remap shifts between safe/competitive seats as +/-0.5, here's how the count looks today, to my eye: CA: Dem +4.5 MO: GOP +1 NC: GOP +0.5 OH: GOP +0.5 TX: GOP +4 UT: Dem +1 That leaves Republicans ahead by half a seat, w/ FL, IN & VA the biggest remaining variables.

@LIEngProf @Redistrict Getting rid of the maps entirely and moving to some sort of proportional system seems to me the only way to actually resolve any of this. Hard to imagine parties won't try to game a national system, most other countries don't draw maps at all. fivethirtyeight.com/features/why-t…

@Redistrict Any chance CO would be able to redraw, or are they restricted until at least ‘28? On a side note, this random mid-decade redistricting isn’t sustainable long term. We’ll have to come up with a nationwide system where special masters with no skin in the game draw the maps.

Life experience data does not become less valid the more systematically you collect it.