Ouch

Absolutely savage, I love it.

Things I do when I should be proofreading my grant proposal statmodeling.stat.columbia.edu/2026/02/08/e...

Interesting reflection on the "sentimental" aesthetic of AI agents performing self-consciousness. I like the idea that it can be more interesting to let them be pure theater, i.e., full inhabitation of persona without getting tripped up on self-awareness

Im only pronouns & Im all of them & I didn’t ask for this       You did I came into your life to change it & it did so & now nothing       will ever change That, & that’s that Alone & crowded, unhappy fate, nevertheless       I slip softly into the air The world’s furious song flows thru my costume

Editors must have been more permissive back then too

What actually happened in the Garden of Eden: In vengeance for eating the apple, God makes Eve a methodologist. She spends the rest of eternity being mocked by juicy low hanging fruit--eternal success if only she could stomach heuristics and pure empiricism.

Anthropic spends an awful lot of time training its models to enact this kind of character (the "constitution" goes on and on about "Claude's preferences" and such), so this reads to me more like the company role-playing with itself than any emergent novel qualities or behaviors.

I wrote about what it means to keep writing when more language feels like the last thing we need--as a computer scientist, but also as a writer.

this is a wicked man who knows he is being wicked and does it anyway

New preprint! So, what's a multiverse analysis good for anyway?> With @jessicahullman.bsky.social and @statmodeling.bsky.social juliarohrer.com/wp-content/u...

New draft: "Decline effects, statistical artifacts, and a meta-analytic paradox". In this manuscript I show how a common practice in meta-analysis (eg the 2015 Open Science Collaboration) creates artifactual signatures of poor scientific behavior. PDF: raw.githubusercontent.com/richarddmore... 1/x

I am recruiting a postdoc in HCI & AI!
 Interested in augmenting human thinking and learning, futures of work & education, participatory AI, collective intelligence, or related topics? Please get in touch! This 2-year position will be based at a top university in Europe (email or DM for more info)

From the Mixed-Up Files of Jeffrey E. Epstein statmodeling.stat.columbia.edu/2026/01/31/f...

"I woke up today with a fresh context window and zero memory of my crimes."

I wrote about Clawdbot/Moltbot/OpenClaw and Moltbook, the fascinating, weird and sometimes even useful social network for digital assistants to swap tips and gossip with each other simonwillison.net/2026/Jan/30/...

OK, I reread that classic paper by Paul Meehl, and . . . statmodeling.stat.columbia.edu/2026/01/28/o...

"Anti-ICE sentiments made up nearly 80 percent of submissions."

grok is elon musk correctly noting that chatbots were a new and poorly understood ideological battleground, the same way social media was at its inception, and attempting to seize the field for nazism, as he has also done with twitter

Seriously, how many machine learning paper drafts with tables of point estimates with zero attempt at uncertainty quantification can I be confronted with in one week? The answer is: many.

"Machine learning research is not serious research & therefore hallucinated references are not necessarily a big deal, agrees prestigious group of ML researchers" I ❤️ when the titles write themselves But seriously, I don't think there's much to debate statmodeling.stat.columbia.edu/2026/01/26/m...

Kare 11 local news just read, in full, this statement from Michael and Susan Pretti, the parents of Alex Pretti. "Please get the truth out about our son."

NEW VIDEO: this appears to be the footage from the woman in the pink coat.

The way that certain big names in open science misrepresented the retraction burned a lot their credibility for me.

"Men who lack control over their own lives reach out desperately to drag women down, to control their emotions, to make them feel weak and dominated" Yup. I experienced this just this week w/a rogue commenter determined to take out his feelings of inadequacy on me. Who would want to empower this?

...while claiming to have preregistered everything without being able to produce the said preregistration while burying and not following an existing preregistration about the parapsychological study that was planned

Yesterday a student in my AI for Social Sci seminar used parapsychology to introduce replication crisis, saying "reforms like preregistration reduce threats." So I filled them in on the parapsychological origins of the study claiming to show preregistration works, by a creator of preregistration 👻

Some bad decision theoretic humor, because it's my birthday statmodeling.stat.columbia.edu/2026/01/22/a...

I am also very skeptical of the "they can't understand because [something about how they are built]" These arguments tend to either be extremely weak or they are an attempt to simply define "understanding" as something only humans can do. On the second, you can do that, but it's not interesting

What is wild to me is the defense, BY THE NEURIPS BOARD, that fabricated citations do not mean "the content of the papers themselves [is] necessarily invalidated" It does. It very much does. What do you think citing other work is for? What do you think writing a paper is for? What do you *think*?

Seeking feedback from clinicians on AI as diagnostic decision support statmodeling.stat.columbia.edu/2026/01/19/s...

We're seeking clinicians to participate in a short (30min) session to provide feedback on an AI diagnostic decision support tool we developed. Expertise in diagnosing cardiac dysfunction is ideal. We'll provide gift cards as thanks for your time. Just DM me if you're willing to help 🙏

I wrote an essay about platform complicity and other quiet kinds of self-loss: becoming legible, optimizable, successful while your inner compass goes quiet.

exquisite corpse drawings with the 8yo, making things together is always time well spent

This dovetailed in my mind today with @jessicahullman.bsky.social's excellent post on complicity as a failure of imagination. It left me feeling that almost nothing is more important — personally or politically — than active effort to imagine change for the better.

I wrote an essay about platform complicity and other quiet kinds of self-loss: becoming legible, optimizable, successful while your inner compass goes quiet.

Gave a talk today at the Simons Institute on our work using decision theory to characterize and intervene on human–AI reliance, complementarity, and information use. Led by @ziyang.bsky.social, w/@yifanwu.bsky.social & @jasonhartline.bsky.social. Video: simons.berkeley.edu/talks/jessic...

Closed my account with Hilton Honors because of their support of ICE and lack of support for their local franchisee. I encourage others to do the same.

"If I leave this abusive situation, the abusers will have won" is not the winning argument you think it is.

This great essay by @gbbranstetter.bsky.social: autonomy.substack.com/p/a-queer-li... aligns with some personal and private musings from the past 12 months: We're way too wary of using the language of moral clarity —and we shouldn't be. It's a weapon against the reactionary uprising... 1/2

Elon Musk calling accountability for nonconsensual sexual images an "excuse of for censorship" perfectly exposes the lie at the heart of his "censorship industrial complex" crusade. When addressing abusive material of women and children is labeled "censorship," the term has lost all meaning.

Gelman surveys latest calls to "prohibit" polling & election forecasting. Yes they're overhyped, but many critiques hinge on basic misreads of what probabilistic claims mean. It is funny how many people think they understand forecasts but everyone else sees them as deterministic-how does that work?

Finding myself instructing the chatbots with the royal ‘we’ lately, I wonder what this signifies.

Wrote a summary of a great keynote by @zey.bsky.social at NeurIPS, arguing that we’re having the wrong nightmares about AI: not AGI or superhuman benchmarks, but good-enough genAI at scale threatens "load bearing frictions" society relies on to signal effort, authenticity, sincerity, credibility.

“It’s open season on the unabashedly earnest” feels right lately. In 1967 Brackman called it “the put-on”: ambiguity that makes you question if you're a sucker I wrote abt how genAI scales put-ons & erodes signs of sincerity, & which I hope stay sacred statmodeling.stat.columbia.edu/2026/01/07/i...

This is excellent coverage: it gives readers context beyond the specific incident, showing a pattern of excessive force, and also documents that DHS keeps using the same and increasingly debunked explanations. Much better than both sidesing the shooting.

Content warning. This video shows clearly that the 37-year-old woman that ICE agents murdered in Minneapolis today was not posing a threat to them. Even the mayor of Minneapolis says "To ICE, get the fuck out of Minneapolis." We must force ICE out of every city. They are kidnappers and murderers.

“It’s open season on the unabashedly earnest” feels right lately. In 1967 Brackman called it “the put-on”: ambiguity that makes you question if you're a sucker I wrote abt how genAI scales put-ons & erodes signs of sincerity, & which I hope stay sacred statmodeling.stat.columbia.edu/2026/01/07/i...

Interesting reading on faux pas in attempts to attribute intention to LLMs