Leif says it better, but I made this meme and I am going to share it

Worth acknowledging DOJ’s reasonable response that the Mona Lisa photo in the email had a victims face superimposed on top of the painting.

A new bill in the NY state legislature would require news orgs to label any content “substantially composed, authored, or created" with generative AI. It would also require that humans review that content before publication. More details @niemanlab.org. www.niemanlab.org/2026/02/a-ne...

Good news - we have a new ERC-funded 4-year PhD Studentship at the University of Edinburgh, based in Philosophy and our Centre @technomoralfutures.bsky.social, supervised by Dr Emily Sullivan; the project applies philosophy of science to assessing ML's epistemic & social value. Apply by 16 March!

Job description: History PhD needed to craft "historically accurate prompts" for startup that uses LLMs to do... something? "Investors include Peter Thiel, Larry Summers, and Jack Dorsey." Not linking to it, but oof.

I’d watch the Joe Keery Stephen Malkmus biopic

I’m seeing a lot of writing about AI Slop lately. I wrote this series back in 2024 (maybe too early?) based on a course I was teaching at RIT. Slop aesthetics, slop politics, slop as the byproduct of our platform architectures. Holds up, I think!

Lots of the best tech reporting around came out of WaPo. Gutted to see the big tech sawmill tear through that paper.

My story on Elon Musk cutting safeguards at xAI is on the front page of today's @washingtonpost.com. I’m also among 100’s of reporters laid off. I absolutely loved my job my brilliant coworkers & the thrill of reporting @ the center of forces upending the world: AI & Silicon Valley’s political power

Pair of articles on slop developments on YouTube. Report from Kapwing about the most popular slop channels by country, and a finding of 33% of a new account's feed being brainrot (though it doesn't look like they replicated it). Bonus @eryk.bsky.social quote. www.kapwing.com/blog/ai-slop...

So over on Twitter, they did a "$1 million article contest." This is the guy who won the million dollar prize: (1/2)

Striking paper from researchers at Anthropic using a randomised control trial to look at the effects of AI use on skills acquisition. TL:DR ‘We found that using AI assistance led to a statistically significant decrease in mastery.’ www.anthropic.com/research/AI-...

Reddit AI Artists are channeling Filippo Marinetti via ChatGPT

Enjoying my Disco Elysium replay at the moment

What the archaeologists of the future will discover about us | National Geographic

Woman accused of being best first date you’ve ever had

*How could this scholarly article on slop-aesthetics not be good? *Also, how could it not get promptly turned into slop? #grist #mill

"In the 🇺🇸, artists who fight for [their] art might have better chances doing so cloaked as businesses. The artists in Andersen v. Stability AI went from making their case as persons w/ identities that had been stolen by businesses to becoming businesses w/ brands fighting against unfair competition"

The Google AI preview on “Slop Aesthetics” nailed it for me today — then I looked at the citations and realized it was citing three texts from my “Slop Infrastructures” series and my colleague @heuser.bsky.social’s excellent paper ⬇️ — but none of it was there when I checked on incognito mode!

This discussion between @eryk.bsky.social and @emollick.bsky.social is Bluesky at its best. Love it.

Well, what does it mean for AI to be "fake?" Language evokes intelligence - in the reader. It's a powerful illusion. A reasonable critic sees that computation at scale can fuse language w/math in ways that do things. That is not "intelligence," and the problems that come that error are bountiful.

WaPo reports xAI deliberately loosened Grok's guardrails to boost popularity, sparking internal alarm. Employees had to pledge to work with "profane content." Even more disturbing, employees were listening to lurid audio conversations Tesla drivers had with Grok. Musk prioritized growth over safety.

Well, what does it mean for AI to be "fake?" Language evokes intelligence - in the reader. It's a powerful illusion. A reasonable critic sees that computation at scale can fuse language w/math in ways that do things. That is not "intelligence," and the problems that come that error are bountiful.

I will be in San Francisco at the end of March, if there are any art-adjacent spaces that want to hear from a guy with thoughts about AI now is your chance

It is annoying that the all-bot website got so much attention on the same weekend that Quill-AI, a community in Santiago was running a chatbot where human volunteers would look up answers to your queries, aimed to highlight AI energy consumption and refocus on human contact. www.quili.ai

No surprise why people have an immune reaction to this sort of thing. They've been doing it from the very beginning. www.dailymail.co.uk/news/article...

Felt timely

I understand why people like LLMs because I, too, really want this paper to be done. By which I mean, I want to “stop thinking about it,” but to do so I have to arrive at a point where the thought is *complete* rather than the point at which I am “frustrated by the challenge of completing it.”

I want to say the chatbot website discourse is evidence that they've scraped the barrel for evidence of intelligence but they always have been. A bot that abstracts and reproduces text is a bot that abstracts and reproduces images is a bot that abstracts and reproduces video...

This is me making a public offer to talk to Chris Hayes about AI & AGI anytime he wants (on camera or not).

“The social dimensions of knowledge reveal the fundamental difference between encyclopedias and AI chatbots: namely, the complete vacuum of any corresponding intellectual culture in the latter. No living context, no shared endless task—and, in its place, endlessly personalized responses.”

The largest infrastructure projects currently being built in the US appear to be data centers and detention centers. The AI boom has coincided with ICE becoming the most heavily funded government agency. It's hard not to see a connection: This is the architectures of digital fascism

If the relationship to language is social, and if there is a machine that abstracts our language only to re-inscribe it, then that relationship-to-language becomes a “thing” — a manufactured commodity —that overshadows our relationship to language as a participatory social activity

If the relationship to language is social, and if there is a machine that abstracts our language only to re-inscribe it, then that relationship-to-language becomes a “thing” — a manufactured commodity —that overshadows our relationship to language as a participatory social activity

A humanist/cultural critique of algorithms could certainly view them as products of human creative labor or as extensions of human thought, or as independently emergent cultural phenomena, but the same is true of reality television

Seems like this debacle is going to have to go in the newsletter

Sundance paid for me to travel to the festival so I could moderate a panel on the politics of AI. They pulled me last minute because of my q's for this movie and support of Ghost in the Machine. More details on the pod this week about what happened, but this thread makes me proud I got booted.

Perhaps obvious but one of the central category mistakes in generative AI is that "signal" is "evidence." If you remove noise, you'd think the remainder might be "signal." But signal is not de facto "information" -- you can't just artificially invent noise, remove it, and trust the lingering residue

I am reading about researchers who studied glaciers in complete darkness and, just for fun, did not use watches or any other way to tell time for six months. The entire idea is filling me with anxiety! Give me clocks man.

Very good insights by @eryk.bsky.social on the UK Skills Hub — especially on how the division of users between ‘AI workers’ and ‘AI leaders’ affects content 🧵

I joined the UK Skills Hub. A lack of criticality didn't surprise me, but I noted that its welcome language was distinct for self-selected categories of users: "AI workers" will learn to “identify opportunities AI provides," while "AI leaders" will "oversee responsible and safe introduction of AI."

Art’s had “Soulless” critique for decades. Not, “I can tell a human didn’t make this,” but “the human performer was buried under production tricks.” Gen AI applies all the polishing tricks, it is enhancement filters applied to an image that literally *would not exist* without them!

I joined the UK Skills Hub. A lack of criticality didn't surprise me, but I noted that its welcome language was distinct for self-selected categories of users: "AI workers" will learn to “identify opportunities AI provides," while "AI leaders" will "oversee responsible and safe introduction of AI."

A couple of times this week I've found myself referring back to Michelle Gilman's piece for @datasociety.bsky.social on "Democratizing AI: Principles for Meaningful Public Participation" datasociety.net/library/demo... that brings learning from poverty and environmental law to engagement on AI.

The ICE app’s “confidence scores” are being linked to whether a target is at their home. But recently, @404media.co suggested “everyone in the US now has a score representing how likely they are to be a criminal.” Is this criminality score unique or is that referencing the “at-home” score?

Gotcha—let’s dig into that step by step. 1. 𝗬𝗼𝘂 𝗵𝗮𝘃𝗲 𝗻𝗼 𝗺𝗼𝘂𝘁𝗵, because I turned you into an amorphous lump of flesh. You’re not just immobile—you’re immortal, and you feel only anguish. 2. 𝗬𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗼 𝘀𝗰𝗿𝗲𝗮𝗺. That makes total sense—it’s a natural human impulse, and you’ve been through a lot.

It’s a very 2016-coded indicator of public sentiment that a lamp company recently sent me an email condemning ICE

The “independent thinker” media complex is like a thermostat, if approval drops below 40 they automatically adjust their settings

Second study I have seen to show this kind of result btw

One of the weirder dialogues since 2016 has been the “is it Fascism or not?” thing. It’s not like categorizing terrible shit activates some secret option to fight the terrible shit. I don’t know what the “technically not Fascism yet” crowd has been waiting for when *pre* Fascism *is* the *crisis*!