I'm a cognitive scientist with an interest in epistemic vigilance, and this essay that's been going around gave me pause. I don't think it's straightforward to apply the concept of epistemic vigilance to interactions with LLMs, as this essay does. 🧵/ sbgeoaiphd.github.io/rotating_the...

A preprint for a paper I've been contributing to on-and-off for over five years is now out! We look at model collapse, the phenomenon where AI models trained on their own output stop working, using the theory of Markov chains, and the music of Alvin Lucier arxiv.org/pdf/2602.19033

I had the opportunity this week to present a fun little poster at SIGCSE. We took data structures students on a field trip to the library to help them better understand how the library functions as a data structure, and the algorithms that librarians use to maintain it. The activity worked great!

Today I was proud to share my 2026 ACM SIGCSE Technical Symposium opening keynote, titled Love, Learning, and Computing Education. Here's a rehearsal, in case you weren't there. Thank you to the organizers for inviting me to speak and to our community for listening and learning. youtu.be/iaT_zvNN_zc

I'm very surprised at how AI-divided this conference is. I've heard everything from "my students started a luddite club with an no AI pact" to the view that skill using generative AI will be the main way software engineers are hired and promoted in the future.

Oh yeah, I'm at SIGCSE! If you follow me here and want to meet up, let me know.

It's worth adding to this conversation that the technologies behind genAI are fundamentally conservative -- they start from the assumption that the future will be similar to the past, that the kinds of sentences, actions, problem solving strategies, etc. that have existed will continue to exist.

I have this sort of crackpot theory that part of our problem is that popular imagination of robots continues to adhere to an expert systems model when in fact the paradigm shifted to ML some time ago. Think of Data from Star Trek. Data could reason beautifully but he struggled +

Amid the “higher-education hellscape,” there may be one place where college life persists in some semblance of its ideal form, @ibogost.com argues. Are small liberal-arts colleges the last, best hope for undergraduates?

Imagine a fugue that focuses on chord movement rather than voice leading. Imagine it has almost no non-chord tones. Now imagine it uses chromatic mediants at both large scale and small. Then realize it's actually by Shostakovich.

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*?

Have any of you ever taught game design (broadly construed) as part of a first course in human-computer interaction? I'm exploring an idea for a project assignment and looking for examples.

Happy holidays, Bluesky! I got you a megathrust earthquake, soil liquefaction, spine-tingling papers about the way our networks confound knowledge, and a PDF in a pear tree. It's my wrap on a year of trying to make sense of how we make sense of what's happening to us. www.wrecka.ge/landslide-a-...

Something to keep your eye on. Popular discourse is starting to turn against AI, so new frameworks are being built to continue selling it. We are about to see a flood of media that qualifies AI: “Gen AI” is bad and overhyped but “predictive” or “analytic AI” is responsible and good. +

YES! Everyone wants us to teach “ai skills” but no one has ever been able to tell me what an “ai skills” even is

Every year in my class I show my students the gender shades project and they are like but it has been fixed now… right?

10000 Bowls of Oatmeal But One Of Them Killed My Father, is an ongoing murder investigation and I need your help to solve it? Read the progress here and submit your own descriptions of oatmeal to help me get my revenge. verbante.itch.io/10000-bowls

Relying on ChatGPT to teach you about a topic leaves you with shallower knowledge than Googling and reading about it, according to new research that compared what more than 10,000 people knew after using one method or the other. Shared by @gizmodo.com: buff.ly/yAAHtHq

Honestly, Hating the Computer is a critical motivator in developing better software. I do not trust anyone who does not - at least a little bit - Hate the Computer.

Is social media dying? How much has Twitter changed as it became X? Which party now dominates the conversation? Using nationally representative ANES data from 2020 & 2024, I map how the U.S. social media landscape has transformed. Here are the key take-aways 🧵 arxiv.org/abs/2510.25417

This Data Through Design program looks really cool. You can submit a proposal to make an art exhibition with data from NYC's open data portal. They give $900 stipends. Proposals are due November 2. datathroughdesign.com/2026

This sort of thing drives home the point that objections to "AI" today are overwhelmingly objections to generative AI everything machines. Algorithms which solve specific problems and work, even if they use classical AI techniques, aren't really AI anymore.

When people learn with ChatGPT instead of following their own searches, they end up knowing less, caring less, and producing worse advice, even when the facts are the same. Friction is an essential ingredient for learning! Convenience makes us shallow. academic.oup.com/pnasnexus/ar...

Any of you internet research people working on religion apps? (Thinking like the “pray with Gwen Stefani” Catholicism app.)

It's become increasingly clear that generative AI produces the appearance of intellectual labor without the substance. That means it increases productivity a ton for performative tasks that were already a little pointless, and does little for tasks that require substance backed by human expertise.

I'm excited to finally have a preprint of this paper up, a few years in the making. In it we argue that industry-driven manipulation of social media research is well underway and that norms and institutions in the field are ill-prepared to resist tech's influence. arxiv.org/abs/2510.19894

I agree with much here & want to add a puzzling (to me) pattern: I asked my interdisciplinary class whether AI use should be allowed. The math/CS folk said only in socsci and humanities courses, and the SS/H folk said only in math/CS courses. This OP says yes in math/CS and no in humanities 🤔

telling that so many (good and correct) critiques of AI sound exactly like the main critiques of the media culture AI was born into

Once again, it is *faculty*—not administrators—who brought this case and won. Including UC Faculty Association chapters of the AAUP.

I'm totally at an impasse regarding AI in intro CS. On the one hand, I am not a fan of generative AI. I don't find it particularly useful and hate the style of code that it tends to generate. My research agenda is generally critical of the claims AI researchers make.

The UC Berkeley School of Information is hiring an assistant professor in the broad field of Information--including areas of info seeking/retrieval, digital humanities, cultural analytics, info viz, & philosophy of information (among others). Deadline Nov 1! aprecruit.berkeley.edu/JPF05014

TIL there is in fact a chi distribution that is used for something, and it is the distribution of the square root of a chi-squared-distributed random variable.

I think I'm beginning to understand how faculty end up with extremely out of date photos. I recently learned that my pfp on linkedin was last updated while I was an undergrad

If you teach CS1 (or equivalent), have you revised any of your assignments due to generative AI? There are a lot of research papers about custom CS1 chatbots, and assignments involving them, but I can't imagine that's the norm at most schools.

IMO the best interdisciplinary projects involve borrowing ideas and methods from another field to enrich your own, not taking your field's ideas and imposing them on another field. The problem is, it's much easier to "revolutionize" a field you don't know than it is to really learn something new.

We need to know what humans do before we argue that machines can do it just the same. There can’t be AI literacy without a literacy in being human: how we make meaning in work and life, distinguishing the artificial from the imaginary from the true. mail.cyberneticforests.com/human-litera...

here's my controversial take: i’m finding some sensational predictions about when the AI bubble will burst as unhelpful and unrealistic as some of the AGI predictions. hear me out 1/

This is creepy and bad on a number of levels, but on top of all of that it likely also doesn't work. Personal preferences are not stable and predictable on an individual level.

Honestly, copyright and tech is so interesting specifically because it doesn't fit nicely into left/right boxes. Copyright both protects authors' ownership of their work, as well as companies' exclusive rights to publish that work. It is somehow both essential to capitalism and a labor right.

Boiling here at home in Cyprus but I put the finishing touches a couple of days ago on this preprint: What Does 'Human-Centred AI' Mean? doi.org/10.48550/arX... Wherein I analyse HCAI & demonstrate through 3 triplets my new tripartite definition of AI (Table 1) that properly centres the human. 1/n

As always, Ted Chiang is great in this interview. cdh.princeton.edu/blog/2025/08...

Cool job alert for sociologists of science! 📢 The University of San Diego is hiring a tenure-track assistant professor whose work takes a critical approach to science and technology studies—broadly defined. Great job in an awesome city! #SociologyJobs #STSinSociology #TenureTrack #HigherEdJobs

I like this whole thread, but most of all I think it gets at my issue with LLMs as "AGI". These models are good at producing smart-sounding language, but the thing that AI should be about is "seeing what is in light of what can be." To some extent, this is the goal of classical AI, but fell off.

We may have the chance to hire an outstanding researcher 3+ years post PhD to join Tarleton Gillespie, Mary Gray and me in Cambridge MA bringing critical sociotechnical perspectives to bear on new technologies. jobs.careers.microsoft.com/global/en/jo...

One problem with an EO requiring AI models to be politically neutral: There's no such thing, says @ruchowdh.bsky.social. www.washingtonpost.com/technology/2...

What's interesting here is that the usage strategies are less lazy than I might have anticipated. One student went through 103 prompts and a total of 58,000 words in the process of co-writing an essay with ChatGPT.

Today (w/ @ox.ac.uk @stanford @MIT @LSE) we’re sharing the results of the largest AI persuasion experiments to date: 76k participants, 19  LLMs, 707 political issues. We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!  🧵:

Margaret Boden and Chuck Mangione on the same day? I'm devastated on behalf of my 20 year old self