There are no words for how evil this is

Write code to make being human easier. Rip code out if it makes being human harder. Write code to make caring for each other easier. Rip it out if it makes caring for each other harder. Empathy driven development has literally never failed me. It’s done me better than everything else combined

Find myself wanting a blue check here, it’s truly a disease

Getting nervous for the talk I'm about to give at a workshop about "using AI to drive impact" which features slides such as these.

I have an image in the latest issue of Limn, a still from Human Movie, here illustrating an excellent article by Wendy Sung on "Digital misrecognition and the making of the 'Asian face.'" limn.press/article/iden...

I have an image in the latest issue of Limn, a still from Human Movie, here illustrating an excellent article by Wendy Sung on "Digital misrecognition and the making of the 'Asian face.'" limn.press/article/iden...

Digital Artist: In my work I invented the Gowl Detector as a cautionary tale Tech Company: At long last, we have created the Gowl Detector from classic art installation Don't Create The Gowl Detector

“Who gets to be seen as fully human in this world? Our job is to expand that and to push and make that as wide as it can possibly be.” Artist Mimi Ọnụọha's work questions the contradictions of technological progress. Anika Collier Navaroli spoke to her for a special podcast series:

okay, now let's hear from people who don't have millions of dollars in tech company stocks still vesting

There are certain theorists who I'd say were wrong about digital art but are right about AI slop, mostly because they saw digital arts as an extreme parody of themselves, and AI slop is, ultimately, an extreme parody of digital art (Slop being the raw & unchallenged output).

“authors & publishers who filed a lawsuit against the Sam Altman-led firm have secured access to internal Slack messages… discussing the mass deletion of a pirated books dataset… A NY district court ordered OpenAI to hand over the communications regarding data deletion” futurism.com/artificial-i...

As a pre-emptive self-defense, I have been hard dashing for ages. I hope one of the many annoyed copy editors will testify to this under oath when the revolution comes for me.

If the sublime is the gap between the Idea and its material presentation (Rancière), then those who see the uncanny valley as a gap between their mind and ~something~ set up the conditions to be convinced that AI is sublime rather than over-fitted representations

New art video just dropped! The 7 Trillion Dollar Seance 🔮 This is meant to be performed live, recited in front of the video at conferences, museums, public events. It got a standing ovation at Tech Together in NYC. Your phone works but best on your largest screen. vimeo.com/1116289621 🧵

Hoping this helps our colleagues across the industry

Kind of a leap, if you ask me. "Common Crawl’s executive director, Rich Skrenta, has publicly made the case that AI models should be able to access anything on the internet. “The robots are people too,” he told me, and should therefore be allowed to “read the books” for free."

In Venture Beat today, a startup claims its LLM can write scientific research papers. In "a striking proof of concept," the author writes, a paper was accepted to a scientific conference. The conference? Agents4Science, "the first open conference where AI serves as primary authors and reviewers."

"Many doomsday anxieties around AI likely reflect the broader epistemological shift away from language as a source of truth. The anxieties we place on them now reflect a deep unease and unfamiliarity with how we navigate a world where words persuade without regard to reality." - @eryk.bsky.social

ah, perhaps instead of pumping millions into endless factional infighting Dem donors could invest in making local Dem organizations genuine civic spaces that can reach people during and between elections

Great interactive social maps for a bleak as hell premise: the tech billionaires behind the #aicoup in the US, and how they’ve set their eyes on Europe.

Great read on LLM “meta fiction” and the constant reversion to the mean lareviewofbooks.org/article/lite...

Great discussion at my Critical AI seminar at Elisava today. Laid out 12 quoted definitions of AI from 12 different fields, from cognitive science to venture capital to AI abolitionists and critics: AI as a field, as an apparatus, as a model, as a political project. What even is this AI thing?

Recently, Reinforcement Learning through Verifiable Rewards was promised to make LLMs better at information retrieval. A recent paper suggests it can help confirm verifiable (true/false) fact, but will also confirm flawed beliefs: www.nature.com/articles/s42...

Some context on Will’s post: Mitchell and Pistilli were being mass blocked not because they were “pro-AI” (they are fairly critical) but because they were falsely *labeled* as pro-AI by lazy blocklist creators. Opt out all you want — I do! But be careful about who/what you trust to filter for you.

Recently, Reinforcement Learning through Verifiable Rewards was promised to make LLMs better at information retrieval. A recent paper suggests it can help confirm verifiable (true/false) fact, but will also confirm flawed beliefs: www.nature.com/articles/s42...

Geoffrey Hinton warns us of the persuasive powers of superintelligence. But, "while AI won't “decide” to persuade people to do harmful things on a global scale any time soon; AI companies are persuading us to do harmful things on a global scale already." mail.cyberneticforests.com/thoughts-on-...

How the inventor of the printing press made souvenir mirrors that could "store" the holy essence of relics! Magical Persian goblets of yore! The Aztec god Tezcatlipoca "Smoking Mirror" and his connection to Elizabeth I's personal sorcerer! These all relate to drones, really - read my latest:

Hinton has been talking about “maternal instincts” in AI, but that’s a symptom of a larger issue: he worries about AI’s power to persuade us, even as he believes in what it tells him. mail.cyberneticforests.com/thoughts-on-...

In case you’re worried about the synthetic text extruders becoming conscious and farming us Matrix-style: this is an excellent and very accessible piece on why those concerns are probably overblown. 👇

Synchronicity again! I just published a long piece on substack culminating with almost your exact arguments on hinton (I even analyze the Stewart interview), but from more of a theological/poetic perspective. Here's the link if you're interested: narrascaping.substack.com/p/twin-spire...

🗣️📢 Call for short essays, poetry and short fiction: LOOMING ‘zine (issue 2, winter 2025). Deadline: Dec 1 LOOMING is a technocritical zine celebrating the human and the analog in the face of Big Tech. CC @kconrad.bsky.social Submit to: [email protected]

I get why it’s confusing, but “everything an LLM says is a lie, including the truth” is, I think, a good frame for dealing with them.

“Who gets to be seen as fully human in this world? Our job is to expand that and to push and make that as wide as it can possibly be.” Artist Mimi Ọnụọha's work questions the contradictions of technological progress. Anika Collier Navaroli spoke to her for a special podcast series:

Eryk is as usual onto the good stuff here. There’s a lot of yardage you can get in AI discourse by taking seriously that language production isn’t evidence of a mind. And you have to make that distinction over and over and over.

More in this vein from @eryk.bsky.social mail.cyberneticforests.com/thoughts-on-...

Doomer visions are fundamentally about language and persuasion, not maximizing paperclips. Very, very old challenges for humans, now wrapped in technical complexity that leads us to debate silly things like the subjective experience of bicycles instead of focusing on e.g. trust.

This is evident in the language around these much-touted industry, academic, or even governmental “partnerships.” The MOUs usually state that they are trying to “identify opportunities,” not that they have any clear sense of the value LLMs will add. (Though it’s even worse when they claim they do).

I’m often confused by the bad-faith readings people have of my writing, often in angry tones, that oversimplify my point or miss heavily signposted caveats. It only just occurred to me that they’re probably responding to AI generated summaries and then getting mad at me for what it skipped.

“AI won't persuade people to do harmful things any time soon; but AI companies are persuading us to do harmful things already.” @eryk.bsky.social on the danger of focusing exclusively on the hypothetical risks at the frontier of AI while ignoring present, tangible harms that don’t implicate AGI.

Hinton has been talking about “maternal instincts” in AI, but that’s a symptom of a larger issue: he worries about AI’s power to persuade us, even as he believes in what it tells him. mail.cyberneticforests.com/thoughts-on-...

Again, this precise use case is why these systems exist-to proliferate people’s racist (transphobic, misogynistic…) imaginary at scale. This should not be seen as a “misuse” but rather the product being used exactly as intended.

Netflix would think this helps bring the scrolling-viewer’s attention back to the show but it seems it barely works in real life

It’s The #WrongBiennial Season and I pleased to be in two Pavilions: “Radical Dreamers” curated by Laura Focarazzo and “Debox” curated by Jenna deBoisblanc & Jon Chambers (Nov 9). Radical Dreamers is open now: thewrong.org/RadicalDream...

If only every single critical "AI" researcher had warned about the likelihood of this danger for machine learning-based systems for the past decade or so and if only an acronym like "GIGO" had existed for decades Prior. Oh If Only 😐 "Too much social media gives AI chatbots ‘brain rot’"

Selten hab ich eine so intelligente Lösung gesehen, wie man den Eigentümer eines Geldbeutels kontaktieren kann. Chapeau! (Netzfundstück)

Extremely random question for anyone who has student loans (since I did my degree without them thanks to a fucking miracle): does your servicer allow third party payments to your account?

The Torment Nexus tweet shows us how artists can invent all kinds of dystopian ends for idealistic and well-intended technologies, but that tech companies can transform those same technologies into forces for social good with zero of the suggested downsides

hey i heard bluesky is improving reply quality by banning u

someone pointed me to a paper at AIES this year that introduces the term "street-level AI" and asks whether LLMs can make real-world judgments. surprisingly it doesn't cite my paper ("street-level algorithms") speaking to similar questions this paper tries to address... from 2019

Researchers claim their tool can use generated text to discover the prompt that generated it: “language models operate as maps on the *sequence* space rather than the embedding space … all information about the input sequence is almost surely preserved end-to-end.” arxiv.org/abs/2510.15511