useful piece. again and again it seems LLMs can be very useful *if you know what you are doing and are familiar with their strengths and limitations* lucumr.pocoo.org/2025/9/29/90...

That's quite the chart annotation. www.nature.com/articles/d41...

Re-reading Skowronek’e “Building a New American State” in 2025 as the administrative state continues to be whittled away is quite the experience.

Trying to phrase as delicately as possible to abide by TOS and cover my ass: Nullifying the plain text of the Constitution as explicitly meant by those that penned it for the purpose of voiding a right to citizenship crosses a threshold that cannot be uncrossed and I hope even this SCOTUS sees that.

"Student arrivals to the US dropped 19% in August year-on-year to just over 313,000 – a fifth consecutive month of declines" www.bloomberg.com/news/article...

i don’t think people on bluesky understand the severity of this situation. i woke up to literally hundreds of texts from friends and family thinking about either moving back home to india or applying to other countries for jobs. this includes me.

Ah, the well known “efficiency of operations” exception to the guarantee of speech free from punishment by the government. “It adds the university's interest in the efficiency of operations may outweigh an employee's rights of free expression.” krcgtv.com/news/local/m...

New work about how perceptions of social consensus shape belief. My sense is that the design of social media, w/ signals of engagement on top of massive attention flows, shapes social consensus… a key factor (IMO ) of information disorder.

One of Bari Weiss’s reporters over at the Free Press was shocked to discover what Columbia students are reading in class: lol wat

Cannot recommend enough the @bretdevereaux.bsky.social series on what Sparta was actually like, deep and accessible dive into a monstrous society that built an outsized reputation on a fragile system of pure cruelty and ho-hum but ultimately fleeting military success acoup.blog/2019/08/16/c...

This “explicit” is a weird editorial choice. The constitution gives the president no power over this at all. It’s not ambiguous.

On Prolific, "we estimate that about 34% of online study participants use LLMs to answer open-ended questions atleast some of the time..." Seems like a very timely paper for behavioural scientists using online samples: osf.io/preprints/so... ; We really need more papers on this issue

Nothing but respect for my Burning Man

NEWS: CVS, the largest pharmacy chain in the country, is currently not offering Covid vaccines in 16 states and DC — even to people who meet the new FDA criteria — because the CDC hasn't signed off and isn't expected to do so (if it does so) for at least three weeks.

A century to build the expert civil service, a few months to burn it down for generations

I hate to say I told you so but...centralizing government data infrastructures and turning them over to inexperiences actors generates big risks to society. We are starting to see the first consequences for social security. Others will follow. www.nytimes.com/2025/08/26/u...

Looks like we'll get a test of what, exactly, SCOTUS meant in Trump v. Wilcox sooner than we thought.

another orientation, another banner year of "strategic finance."

Notable that AI tools could not even generate accurate meeting summaries beyond “a couple hundred words.” That’s useless for minutes. And for lit review they *completely* shit the bed. But these aren’t people’s favorite tasks & AI is way quicker, so many will keep using them, truth be damned.

A stunning forensic account of how ignorance of vaccine efficacy was systematically manufactured. Steier exposes the anti-vaccine playbook: how flawed studies by discredited researchers created false doubt about vaccine safety, while 40+ rigorous studies of 5.6M+ people found no autism link.

Wow. The University of Chicago, a world-class institution whose humanities faculty in the past has included Homi Bhabha & Lauren Berlant & Ralph Ellison & Hannah freaking Arendt, is getting rid of basically every graduate department involving the acknowledgment that other cultures & languages exist.

Yet another helpful explainer to counter the ignorant disinformation spouted by the unqualified extremist who is currently destroying the HHS

Pleased to share the latest version of my paper with Arthur Spirling and @lexipalmer.bsky.social on replication using LMs We show: 1. current applications of LMs in political science research *don't* meet basic standards of reproducibility...

The entire business model of R1 public universities rests on 4 revenue sources: 1. Federal grants 2. Private gifts/endowments 3. Tuition (esp. from foreign students) 4. State $ For decades, 📈 in 1-3 offset a secular 📉 in 4. Now, 1 & 3 are being decimated & 4 ain't coming back. The math is clear.

the Wizard of Oz AI sphere thing is interesting (and gross) aesthetically to me because of the way the generated material draws on a certain type of hi-def photorealism stitched onto a movie whose charms, in my mind, are largely attributable to everything clearly transpiring on a soundstage in 1939

Dear God, look at the jobs report revisions for May and June! This is not at all the rosy picture it looked like to economists over the past few months.

"Fast forward a little further, and there is the looming and very real possibility of a national or global financial crisis precipitated by a rapid devaluation of hundreds of billions or trillions of dollars worth of real estate values."

The problem is not that we don't have enough laws on the books, it is that the administration is breaking the law and so far SCOTUS has let it. In 2022 Congress passed a law for force Presidents to provide information about spending. Vought is ignoring it. donmoynihan.substack.com/p/when-youre...

Hey is it bad if google AI search results stop people from leaving google dot com 99% of the time? All our news and information creators can survive just fine with one percent of their previous search traffic right? I’m sure it’s fine. www.404media.co/googles-ai-i...

Matt Walsh's imaginary professor: "good morning student-comrades. today we learn how to bayonet a white Christian capitalist while taking trans hormones" actual average professor: "folks, I am begging you, please do the reading for tomorrow this time. it is five pages"

My god these guys are such spectacular morons gizmodo.com/billionaires...

The endgame of AI slop is not the mass virality that we have seen to date. It's hyperspecific, ridiculously niche content about things that didn't happen but which tick the boxes of your "interests" as determined by a surveillance algorithm, and everyone living in their own little worlds of bullshit

The draft syllabus for my fall course on the political economy of AI. One missing reading that I need: something good and comprehensive (academic, policy report or serious longform journalism) on how the AI labs raise money. www.programmablemutter.com/p/the-politi...

This lawsuit is brutal. One of the exhibits is the letter Universal sent to Midjourney, including a quote from a 2022 Forbes interview that I bet midjourney regrets

yep

Cutting NIH or NSF funding sounds abstract. Seeing what research is actually being cut makes it real. Kudos to @nytimes.com for this piece: www.nytimes.com/interactive/...

Incredible. Reads like a conversation with a complete sociopath who is also really, really lazy. amandaguinzburg.substack.com/p/diabolus-e...

Happy to share this paper with Jeevan Baniya, @smeserve.bsky.social‬, and Brigitte Seim. We used multiple survey experiments to see how Nepali voters think about the trade-off between vote-buying and candidate provision of local public goods (schools, plumbing) in office. doi.org/10.1111/ajps... 1/

The geographic scale of these attacks are extraordinary. Irkutsk is 3,000 miles (4,900 km) from Kyiv, at the same longitude as central Mongolia. Put it another way, that's like Anchorage successfully launching an attack on Indianapolis.