iiinova — where your digital self keeps evolving

Reposted by Eryk Salvaggio

Dorothea Baur - the ethics crusader @dorotheabaur.bsky.social · 4h

What a great panel at Zurich AI Festival, feat. among others @eryk.bsky.social @bvlsingler.bsky.social @drkatedevlin.bsky.social

Abraham Bernstein Eryk Salvaggio Nicole Freudiger Kate Devlin Beth Singler seated on a colorfully lighted stage

1135

Reposted by Eryk Salvaggio

Kate Devlin @drkatedevlin.bsky.social · 2h

The 4.5hr flight delay aside, what an absolute treat to get to do an AI panel in Zurich with @eryk.bsky.social and @bvlsingler.bsky.social, both of whom are consistently excellent in their AI takes. Also I saw a goat robot.

Three people — Eryk, Beth, and me — smiling for the camera in a selfie pic taken at a table in a hotel bar.

1212

Eryk Salvaggio @eryk.bsky.social · 7h

Appreciate the nod from Gebru & Bender in this piece — which rightfully asks why the threat of “rogue” AI has been framed as a “threat to humanity” when the immediate problem of blind trust in LLMs — not loss of control — nearly lead the US to bomb a Chinese ship this summer.

theguardian.com

Forget ‘superintelligence’: error-prone AI nearly sparked world war three this month | Timnit Gebru and Emily M Bender

The risks of AI aren’t what we think they are, as a recent security incident between China and the United States reveals

1225

Eryk Salvaggio @eryk.bsky.social · 11h

Seems appropriate to have an AI conference on a street named after Elias Canetti, author of “Crowds and Power” and “The Tongue Set Free.” The universe is offering up a lit review…

Street sign in Zurich, blue with white text: “Elias Canetti Strasse”

0225

Eryk Salvaggio @eryk.bsky.social · 13h

“I’m a victim of bad critics I just have a genuine question” but if I as a critic try to get a sense of (what I agree is a good) question I get three-word answers that end with “lol”

170

Eryk Salvaggio @eryk.bsky.social · 21h

I’m excited to be back in Zurich for a panel with @drkatedevlin.bsky.social, @bvlsingler.bsky.social, Abraham Bernstein and Nicole Freudiger tomorrow for the AI+X Conference — we will be addressing the social question of “what is AI doing to us?”

070

Eryk Salvaggio @eryk.bsky.social · 23h

I don’t think this is a good idea

24011

Reposted by Eryk Salvaggio

Matthew Kirschenbaum @mkirschenbaum.bsky.social · 30/09/2026

CFP: Generative Digital Humanities, a volume on AI and DH in the long-running Minnesota Debates in DH book series. To be co-edited by @zentralwerkstatt.org, @laurenfklein.bsky.social, @lnakamura.bsky.social, and myself. Deadline Nov. 1. Please share widely! dhdebates.gc.cuny.edu/page/cfp-gen...

dhdebates.gc.cuny.edu

CFP: Generative Digital Humanities | Debates in the Digital Humanities

Transforming scholarly publications into living digital works

55856

Eryk Salvaggio @eryk.bsky.social · 30/09/2026

Looking for a post doc? Come join us at the Machine Visual Culture Research Group, which “combines digital and computational art history, critical AI studies, and the history of art and visual culture.” www.biblhertz.it/3815269/2909...

biblhertz.it

Call for Postdoctoral Research Fellows in Digital Art History

01821

Reposted by Eryk Salvaggio

Justin Hendrix @justinhendrix.bsky.social · 29/09/2026

United States President Donald Trump issues executive order instructing federal government agencies to "use the terms 'Super Intelligence' and 'SI' in place of 'Artificial Intelligence' and 'AI' and will not acknowledge the usage of 'Artificial Intelligence' and 'AI' in any applicable setting."

whitehouse.gov

Inaugurating The Era Of Super Intelligence

By the authority vested in me as President by the Constitution and the laws of the United States of America, it is hereby ordered: Section 1. Every child deserves a safe and nurturing home where they ...

312496148

Reposted by Eryk Salvaggio

Rebecca Williams @rebeccawilliams.info · 29/09/2026

I saw someone say “I read the report so you don’t have to” & then make this misrepresentation, you gotta read the report yourself (or listen to Eryk here) it’s turtles all the way down w/ false claims in this space, hold on to any real fact with dear life

072

Reposted by Eryk Salvaggio

Patrick* Neithard @patrickneithard.eurosky.social · 29/09/2026

Eryk Salvaggio: A CRITICAL VOCABULARY FOR ARTIFICIAL INTELLIGENCE

051

Eryk Salvaggio @eryk.bsky.social · 29/09/2026

Was a real honor to present my talk, “Counting the World on Six Fingers,” today at the Venice Biennale Sessions… part of “Dysphoria Mundi: Making and Breaking World Models.”

Man at desk that says “Biennale Sessions” with screen of a six fingered hand behind them on a gray and red brick wall

1131

Eryk Salvaggio @eryk.bsky.social · 28/09/2026

There’s a misinterpretation of the report on OpenAI’s Hugging Face hack being spread by skeptics and I think it’s important to clarify — the notes that the model wrote to itself operated as prompts. They were not prompts written by the engineers.

1365

Reposted by Eryk Salvaggio

Neil Turkewitz @neilturkewitz.bsky.social · 28/09/2026

Please read this essay which brilliantly explores the tension between Silicon Valley & the health of our democracy. Consent is kryptonite to tech’s aspirations for efficiency, but is not unimportant to our humanity.🕊️ We’re in the midst of an existential battle, but it has ∅ to do with rogue agents.

13715

Eryk Salvaggio @eryk.bsky.social · 28/09/2026

In Venice for “Dysphoria Mundi: Making and Breaking World Models.” Excellent lineup of talks coming up. www.unive.it/web/en/18653...

unive.it

Dysphoria Mundi: Home

Dysphoria Mundi - Home

1101

Eryk Salvaggio @eryk.bsky.social · 27/09/2026

Hugging Face was a case study, and I’d guess most of these followed the same test setup — not surprising, because the companies weren’t aware and so weren’t responding. But I’d like to see what models & tasks were involved. With so many incidents, they should know what the patterns were.

1338

Eryk Salvaggio @eryk.bsky.social · 26/09/2026

This, from Dennett, I wish I'd taken to heart earlier in my life. www.themarginalian.org/2014/03/28/d...

How to compose a successful critical commentary:

1 You should attempt to re-express your target’s position so clearly, vividly, and fairly that your target says, “Thanks, I wish I’d thought of putting it that way.”
2 You should list any points of agreement (especially if they are not matters of general or widespread agreement).
3 You should mention anything you have learned from your target.
4 Only then are you permitted to say so much as a word of rebuttal or criticism.

520336

Eryk Salvaggio @eryk.bsky.social · 26/09/2026

If they could do it for Geese, they can do it for extinction risk www.wired.com/story/geese-...

wired.com

2124

Eryk Salvaggio @eryk.bsky.social · 26/09/2026

Instead of finding adjectives to inflate headlines, journalists might focus on what the “regulations” called for by industry to regulate itself actually are and what they would do.

NYTimes hed: OpenAl's Systems Meddled With
U.S. Government Sites After Going Rogue
The company did not learn until recently that its technology had meddled with websites for the Education and Commerce Departments and the Securities and Exchange Commission.

A Commerce Department spokeswoman said OpenAI had gotten access to information that was publicly available on the Census Bureau's website and available to anyone, but not to any private data.
An Education Department spokesperson said "system operations reviews have found no evidence of any impact to our website or databases."

2199

Reposted by Eryk Salvaggio

Eryk Salvaggio @eryk.bsky.social · 26/09/2026

In the past week I have been quoted in a lot of pieces and the titles are all over the map: “humanities researcher,” “technologist,” “tech critic,” “artist (derogatory)”…My fave was “respected AI researcher” in The New Yorker last year. None of these are really wrong, but none are really correct.

4374

Eryk Salvaggio @eryk.bsky.social · 26/09/2026

4374

Reposted by Eryk Salvaggio

Faine Greenwood @faineg.com · 25/09/2026

Imagine if instead of regulating drone risks the normal way (which we did), we instead paid weird bloggers lots of money to worry about superintelligent dogs sprouting thumbs and using drones to kill us. This is pretty much how the US is approaching AI risks and regulations. My latest:

little-flying-robots.ghost.io

Why We Can't Let Enormous Weirdos Regulate AI

Imagine if instead of regulating drones via the standard lengthy, boring, and bureaucratic process - which we did - we instead decided, in the name of fostering all-important American drone innovation...

25956278

Reposted by Eryk Salvaggio

Dr. Damien P. Williams, dread portent down from a mountain cave @wolven.blacksky.app · 25/09/2026

I wrote a paper about this (i wrote a dissertation about this, but the paper is more digestible); it's called "Any Sufficiently Transparent Magic"

48331

Reposted by Eryk Salvaggio

maybe: leif weatherby @leifw.bsky.social · 25/09/2026

i'll be at Princeton week after next for a short residency with three events, if you find yourself in NJ cdh.princeton.edu/events/leif-...

cdh.princeton.edu

Leif Weatherby at Princeton

043

Eryk Salvaggio @eryk.bsky.social · 25/09/2026

I started researching ML before I built a synth, but you can get a great intuitive understanding of interaction within a generative system by building a modular analog synthesizer and watching how your components interact, and seeing what surprises emerge and what stays predictable as you scale up.

Steve Porcaro of "Toto" with a wall-sized modular synthesizer - basically a massive nest of tangled wires -- the Polyfusion 'Damius'. Photo by Jim Gardner - https://www.flickr.com/photos/jamesthephotographer/120916737/

39311

Reposted by Eryk Salvaggio

Roland Meyer @bildoperationen.bsky.social · 25/09/2026

#JobAlert: Assistant Professor (Tenure Track) or Professor of History of Art and Architectural Culture, ETH Zurich. The closing date for applications is 30 November 2026. ethz.ch/en/the-eth-z...

ethz.ch

Assistant Professor (Tenure Track) or Professor of History of Art and Architectural Culture

0147

Reposted by Eryk Salvaggio

Timnit Gebru @timnitgebru.blacksky.app · 24/09/2026

Democracy Now! has now gone from constantly inviting Max Tegmark to platforming people from none other than MIRI. That's right MIRI. Y'all know how bad that is? Do you understand the level to which they are doubling down on the eugenicists?

11287100

Reposted by Eryk Salvaggio

Fabian Offert @zentralwerkstatt.org · 24/09/2026

The program for our Critical AI Studies? conference, happening Oct. 9&10 at @ucsantabarbara.bsky.social, is now online: huml.ucsb.edu/conference. Hope to see you there!

34519

Eryk Salvaggio @eryk.bsky.social · 25/09/2026

Democracy Now! heard a fox crying wolf, so they let it into the henhouse. MIRI is a tech funded organization that helps the industry launder its accountability. See @olivia.science on “Entryism” olivia.science/entryism/

olivia.science

Entryism

AI entryism subverts anti AI critique.

47129

Eryk Salvaggio @eryk.bsky.social · 24/09/2026

Looking forward to attending & presenting at the upcoming “Critical AI Studies?” conference, organized by @zentralwerkstatt.org and @owenleonard.bsky.social at UCSB: huml.ucsb.edu/conference

huml.ucsb.edu

1131

Reposted by Eryk Salvaggio

Dr Zena Assaad @zenaassaad.bsky.social · 23/09/2026

This has been all over my feeds this morning with lots of "AI gone rogue" narratives. This is also appearing in the wake of Mark Zuckerberg telling the UN AI has a chance of destroying humanity and the solution is tech companies self regulating. www.rnz.co.nz/news/world/1...

rnz.co.nz

OpenAI hacked Medicare portal, Australia Prime Minister Anthony Albanese says

Prime Minister Anthony Albanese has revealed an AI agent hacked into an Australian Medicare data portal earlier this year.

31511

Reposted by Eryk Salvaggio

Luke Stark @lukestark.bsky.social · 18/09/2026

There are a number of folks who have written extensively on this problem. The fact the critique isn't better known is in part because the critics don't have the money and media power of the AI lobby. Also they're often women, BIPOC, queer, or some combination of all three.

518945

Eryk Salvaggio @eryk.bsky.social · 23/09/2026

I agree, and wrote about this in February: voices concerned with justice, environment and fairness haven’t neglected the concerns about AI. They have been remarkable consistent, prescient, and weirdly maligned. techpolicy.press/stochastic-f...

techpolicy.press

Stochastic Flocks and the Critical Problem of 'Useful' AI

Acknowledging that AI systems are advancing does not buy into hype, it sharpens the precision of critical thinking about their impacts, says Eryk Salvaggio.

1205

Eryk Salvaggio @eryk.bsky.social · 23/09/2026

Headed to Venice for the Biennale Session, “Dysphora Mundi: Making and Breaking World Models,” where I’ll present a talk: “Counting the World on Six Fingers,” concerning AI generated hands that brings experiences of body dysmorphia to play as a method for image analysis. www.unive.it/web/en/18653...

unive.it

Dysphoria Mundi: Home

Dysphoria Mundi - Home

1110

Eryk Salvaggio @eryk.bsky.social · 23/09/2026

What this story describes comes across as child labor for the knowledge economy.

1153

Reposted by Eryk Salvaggio

Hypervisible @hypervisible.blacksky.app · 23/09/2026

Part of their problem will continue to be that OpenAI is trying to convince people that they need a layer of (wonky) enterprise software between them and things like raising kids, cultivating relationships, and exploring hobbies.

businessinsider.com

OpenAI is enlisting an influencer army to make it look 'good for the world'

OpenAI is hiring influencers to create content that portrays AI as cool and helpful, not scary. Here's an inside look at its marketing strategy.

511947

Reposted by Eryk Salvaggio

Ted Underwood @tedunderwood.com · 22/09/2026

Models like Talkie-1930 sound like voices from the past. If they could reliably speak from specified historical vantage points, researchers might also use them to simulate the past. But how reliable are they? Today we release a benchmark answering that question for English contexts 1831-1930.

Free-text evaluation of answers to character modeling and constrained generation questions. This is just one of several scores Chronologic-EN can produce; we focus on it here because it's both the hardest test and the one most relevant to simulation of the past. Frontier models reach 72%; Talkie-1930 is stronger than several larger competitors, but not at the frontier by this measure. Note that this score has improved ~45% in the last two years, but still falls perceptibly short of ground truth.

515541

Reposted by Eryk Salvaggio

The Distributed AI Research (DAIR) Institute @dairinstitute.bsky.social · 22/09/2026

@timnitgebru.blacksky.app and @emilymbender.bsky.social puncture the latest AI hype, addressing everything from "hacking" to "math breakthroughs" to "superhuman" machines (and possibly break a record for scare quote usage along the way). www.technologyreview.com/2026/09/22/1...

technologyreview.com

Don’t be fooled by this summer of AI hype

Breathless claims about AGI and new capabilities fall apart pretty quickly under scrutiny.

13912

Eryk Salvaggio @eryk.bsky.social · 22/09/2026

Canonical reference on Italian Brainrot just dropped, if you’re not citing this in your Slop Studies papers you are doing them wrong medium.com/@fabianmosel...

medium.com

The Entire History of Italian Brainrot

Documenting the historical AI slop meme that changed the internet

3377

Reposted by Eryk Salvaggio

GAI Debord @dylanschenker.com · 21/09/2026

This is a very fascinating observation from a person who was instrumental in popularizing Italian Brainrot! A decentralized hyperstitional creation with no singular origin medium.com/@fabianmosel...

When people started pretending that a bunch of random AI images were an IP, they inadvertently caused these character to break out and become a global entertainment success story, without being tied to any company. The participatory creation of media brought manufacturers to produce merch and toys in a decentralized way, videogame makers to create the most popular Roblox games, and anyone who wanted to use them, was able to, because no own owns them.

And now, one year later, the ouroboros has completed. On April 21st 2026, OpenAI’s 5th iteration of DALLE, commonly known as ChatGPT Images 2.0, tokenized, categorized and trained on Italian Brainrot, resulting in the first image model that can generate Italian Brainrot on its own.

There’s something very poetic in these AI-native characters that did not have any context at all, no origin except whatever DALLE 3 knew about the world, to now be the global phenomena it has become, therefore being ingested through the training dataset and allowing us to regurgitate it again forever through the latest OpenAI’s image model. The tokenization of token-native characters. Internet folklore.

2225

Reposted by Eryk Salvaggio

Timnit Gebru @timnitgebru.blacksky.app · 22/09/2026

Ridiculous we have to argue about basic things like what the language modeling task consists of and what our paper on LLMs actually says. We're definitely in a "post truth" society where "thousands of lessworng posts" are advising policy makers.

314312

Reposted by Eryk Salvaggio

Max Kennerly @maxkennerly.bsky.social · 22/09/2026

Co-founder of Anthropic accidentally argues the only people with sufficient brainpower to "think through the response to the situation" are AI haters. 😂 And he's just wrong: if you want to understand "the situation" with LLMs, "stochastic parrots" will point you the right way. "AGI" won't. /1

Jack Clark
@jackclarkSF
"stochastic parrot" was a mimetically-fit cognitive virus that spread from 2021-2025; it temporarily blinded many gifted people to the nature of AI progress, burning up crucial years in which they could have helped think through the response to the situation.
10:34 AM · Sep 20, 2026

829145

Reposted by Eryk Salvaggio

Richard Jean So @richardjeanso.bsky.social · 21/09/2026

Prospective PhD students: if you work on computational humanities, cultural analytics and/or cultural AI, pls apply to Duke English! We're recruiting in these areas & you'd join a vibrant community. Candidates with training in both lit studies + CS/sciences esp welcome. Email me if any questions!

12824

Eryk Salvaggio @eryk.bsky.social · 21/09/2026

I think we can acknowledge that the AI air traffic control system might be a bad idea but that it probably didn’t cause Amtrak to disrupt a power line during tunnel construction? Maybe I’m too trusting but is there any reason yet to doubt the official explanation? Please correct me if wrong!

1130