Just an enormous difference in Russian support for the war when comparing direct question estimates with experiments from list experiments. Also, peep those CIs! sensitivity bias has to be HUGE (which it apparently is in this case) to be detected with Direct - List

Georgy Tarasenko@tarasenkogeorgy.bsky.social · last wk.

Every wave after that splits. Direct support holds at ~75%. Indirect support drops to 44–58%. The gap runs from 16 to 30 percentage points.

Remote work is a feminist issue and this study (and NYT op-ed) panning it are just plain silly reason.com/2026/07/27/n...

No, remote work is not making us miserable

The rise of remote work is a tech success story, despite what one new and very convoluted study claims.

reason.com

Very cool that Shopify’s CEO thinks the poor should have far less of a vote than the rich. Still worth that ad read money, folks?

🤩 bingkaiwang.github.io/RCT_Bench-ma...

Imagine that in pre-publication peer review, the reliability of a 3-reviewer composite is .70. The journal accepts the top 10% of papers according to average review score. What is the reliability of the original reviewers in the subset of published papers? (Answer in alt text)

The US Justice Department is prosecuting Samuel Tunick for providing border agents with a “duress” password that wiped his GrapheneOS phone during a border search.

US accuses American of allegedly wiping his phone using a 'duress' password during border search | TechCrunch

A U.S. citizen has asked a court to throw out the government's claim that he gave over a passcode to border authorities that wiped his phone's data, opening up fresh questions about a person's constit...

techcrunch.com

I'm grateful to the authors of this paper for sharing their data and code, but I don't think they did the right analysis! pubpeer.com/publications...

We're excited to be hosting CODE@MIT again this fall! We invite a mix of work involving experimentation — methods, cool experiments, industry reports, etc. Look for more info later on this year's invited speakers.

CODE@MIT@codemit.bsky.social · 2w ago

📢 Call for abstracts for the Conference on Digital Experimentation (CODE@MIT), which takes place Nov 13-14. 📨 Submit your best work by *Sept 13th*! ide.mit.edu/events/2026-...

*walks into room of mathematicians* Have you considered DOING A BREAKTHROUGH.

mr. TIM@timkellogg.me · 2w ago

the Dinitz-Garg-Goemans conjecture, a graph theory problem that’s gone unsolved for 30 years, was solved with this prompt

This episode is interesting as it's about a research group explicitly trying dozens to hundreds of quasi experiments, progressing only with the ones that "work", backing them up with a bunch of robustness tests, and then extrapolating a lot from the results. podcastaddict.com/freakonomics...

On Nolan's previous adaptation of a classic epic: Inception and Scrooge McDuck

"Cool Zone New Year" was meant to be a joke@discchoco.bsky.social · 3w ago

I am not joking by the way www.cracked.com/article_1902...

It seems like one barrier to partially automated review of scientific literature is commercial AI agents refusing to do CAPTCHAs. I can imagine this being something where various open models are better tools?

Dean Eckles@eckles.bsky.social · 5mo ago

Seems like one task Claude Code / Cowork can't reliablty handle is finding and downloading PDFs for dozens of academic papers. (Or what am I missing?) Good reminder how difficult access to these are — even when you're on a university network giving you subscription access.

This gorgeous essay from the social-media-less Lily Hu is everything: on what it means to "do" philosophy amidst pressure to automate it (with lessons that travel much more broadly). Rhymes with brilliant insights from @biblioracle.bsky.social @janerosenzweig.bsky.social Ted Chiang & others

Lily Hu · Un-Doing Philosophy

Analytic philosophers in the United States are increasingly embracing AI, both by taking advisory positions at tech companies and by using large language models for their own research and writing. Som...

blnreview.de

The correlation coefficient is not a proper scoring rule. An agent who ranks models by correlation can be beaten by an infinite class of alternative rules. Here's a 2007 article on proper scoring rules: sites.stat.washington.edu/raftery/Rese...

This skeet will self destruct@pardoguerra.bsky.social · 4w ago

This is quite the finding for a device thats just a plagiarism machine, or whatever we are calling it this week. www.nature.com/articles/s41...

Reuters confirms. 11 more people captured by ICE have been sent to a maximum security prison in Eswatini (formerly Swaziland), a country where they have no ties and committed no crimes. reut.rs/4wpbIGU

Eleven more Trump deportees arrive in Eswatini from the United States, lawyer says

Eleven more migrants deported by the Trump administration under a ​policy of sending them to third ‌countries arrived in Eswatini from the United States on Wednesday, a U.S. lawyer following the case ...

reut.rs

Gillian Brockell@gillianbrockell.com · 4w ago

A few minutes ago, two ICE flights operated by Journey Aviation landed in Eswatini, likely another third-country removal of immigrants to a maximum security prison there. Nineteen people have already been sent there on previous ICE flights in the last nine months. Only two have been released.

I have just resigned from the board of "Statistics and Computing", along with 17 other Associate Editors. This was a difficult decision: this great journal has published many tremendous articles under the leadership of EiC Ajay Jasra, and before him David Hand, Gilles Celeux, and Mark Girolami.

Read access to private repositories + unstructured prompt filled in by user + post arbitrary LLM response as a comment = data leak No shit babes. It’s lethal trifecta, yet again Luckily you can fix this by understanding how access control works 🫡

amos@fasterthanli.me · 4w ago

omg THIS IS AN ACL PROBLEM NOT AN LLM PROBLEM STOP GIVING PUBLICLY ACCESSIBLE AGENTS UNFETTERED ACCESS TO RANDOM SHIT noma.security/blog/gitlost...

Whether or not radicalization is a good replacement, the political science concept of polarization is normatively dubious, descriptively thin and causally... eh, let's not even go there. I know you just got a big grant to study it but I'm sure they'll let you spend the money on something more useful

Petter Törnberg@pettertornberg.com · 4w ago

‘Polarization’ has become the master concept for diagnosing our current democratic crisis. But this is not polarization. It is the radicalization of the political right. By misdiagnosing it, we risk becoming complicit in the crisis. 🧵 Just out in Perspectives on Politics doi.org/10.1017/S153...

*sharing a screenshot from an article that cites another article quoting one (1) “New York financier” who says his firm prefers humanities grads to STEM* this is because the humanities teaches you critical thinking

What if I told you that your state rep can say they voted in favor of a bill, but you have no way of verifying that fact?

It's more than a little depressing how many academics I admire in other contexts keep citing an understanding of LLMs that hasn't been current since LLMs were wedded to code, webfetch, and search years ago.

conputer dipshit@davidcrespo.bsky.social · last mo.

starting to think this standard characterization is just false, not in a subtle way but in the bare and straightforward way “it‘s raining” is false when it’s not raining www.nytimes.com/2026/06/30/o...

This study only varies gender and not AI use. The headline results are compatible with a pure gender effect on competence judgments with no specific AI effect.

David Ho@davidho.bsky.social · last mo.

Women’s use of AI is perceived as a sign of incompetence, while men’s use of AI is seen as an indication of initiative. 🤦🏻‍♂️

Great moments in U.S. History.

I was in the Bay Area over the weekend, and Alyosha Efros twisted my arm to give a talk in his weekly research group meeting and to "keep it entertaining." I took this as an opportunity to think of five philosophy papers everyone working on AI/ML should read. realizable.substack.com/p/five-philo...

Five Philosophy Papers Everyone* Should Read

*in AI/ML

realizable.substack.com

I joined @ericseufert.bsky.social on the Mobile Dev Memo podcast — one of my favorite shows — with coauthor @garjoh.bsky.social to discuss our PNAS paper (with Zhengrong Gu) on removing cookies and privacy-enhancing technologies (PETs) in digital advertising. mobiledevmemo.com/podcast-quan...

Podcast: Quantifying the advertising value of cookies (with Garrett Johnson and Shunto Kobayashi) | Mobile Dev Memo by Eric Seufert

Podcast: Quantifying the advertising value of cookies (with Garrett Johnson and Shunto Kobayashi)

mobiledevmemo.com

I'm reminded when I see this much of our perception of the safety of hotels vs cruises is based on the fact that cruises must publicly report illness and hotels have no such requirement. Part of a larger point that most of our view of modern life is formed by what story is easier to assemble.

Forbes@forbes.com · last mo.

The norovirus outbreak comes just months after the much more deadly hantavirus spread through a Dutch-flagged cruise ship sailing through the South Atlantic.

People who spiral out overusing LLM chatbots are really just experiencing the kind of brain-damaging stream of unconditional validation that’s always been the norm for the hyper-wealthy, which explains a lot.

Jon Hendry@jonhendry.bsky.social · last mo.

Worse than Twitter, private discussions (text, email, whatever) among like-minded wealthy lunatics who are all surrounded by yes-men who talk like ChatGPT in maximum ass-kissing mode.

I'm in Europe talking about US criticisms of the Digital Services Act. I forgot to discuss one of its most basic problems: it shares an acronym with the Democratic Socialists of America.

Tamoor Hussain@tamoorh.com · last mo.

Incredible moment where Fox News puts up a graphic with all the reasons it opposes the DSA and why people should be afraid of them

First paper from my grad student, @d-spinozamartin.bsky.social, who is also new to Bluesky!

Daniel Spinoza-Martín@d-spinozamartin.bsky.social · last mo.

In a new paper out in @cognitionjournal.bsky.social, @gordpennycook.bsky.social and I argue the classic view that System 2 spontaneously corrects System 1 is flawed. We provide evidence that S2 is instead triggered by conflict between competing S1 intuitions www.sciencedirect.com/science/arti...