The Who's Who of AI
What the AI community is actually reading and debating — the articles its researchers, builders, ethicists and critics are sharing, and the discussions they're having, ranked by standing inside the field.
Reading
What the directory is reading
Statement on the US government directive to suspend access to Fable 5 and Mythos 5
5 members discussed this
SE Gyges: us government has rendered it illegal to give access to fable or mythos to foreign nationals ^_^ www.anthropic.com/news/fable-m...
SE Gyges: this means nobody can have either of them btw. because they haven't KYC'd their customers hard enough.
SE Gyges: dario got what he wanted
Open the discussion →
Hugging face model evaluation security incident
6 members discussed this
Grace: This headline is extremely funny given what happened (OpenAI hacked HF by accident) openai.com/index/huggin...
Grace: Maybe the most concerning part is the OpenAI claim to not have known about this before investigating?
Grace: Well, I think the model passed the test
Open the discussion →
science.org
shared by Emily M. Bender, Jill Walker Rettberg, Claes de Vreese, Iris van Rooij, Olivia Guest, Salomé Viljoen +7 more
“This article ends with a good conclusion, but it is extremely misguided along the way. It is also a rich text for exploring how clear language helps make the point better. A thread: www.science.org/doi/10.1126/... >>”
Is AI ruining our skills? Early results are in — and they’re not good
2 members discussed this
Edouard Machery: The literature on cognitive offloading and deskilling should be taken with a serious grain of salt. See literature on video games in the 1980s, see Haidt style literature on soc…
Alondra Nelson: I find the research increasingly compelling — this week's example being a case in point: www.nature.com/articles/d41... I look forward to reading your book and having these conc…
Edouard Machery: The issue, in my opinion, is publication bias and p-hacking. The literature is clearly titled toward finding negative effects - this is what gets in the news etc. As a result I …
Open the discussion →
Meta Silently Added Face-Recognition Code for Its Smart Glasses to Millions of Phones
2 members discussed this
dell cameron: NEW: WIRED analyzed Meta's AI app and found NameTag — a hidden facial recognition system designed to identify people via the wearer's smart glasses — has already shipped to over…
Justin Hendrix: NEW: WIRED analyzed Meta's AI app and found NameTag — a hidden facial recognition system designed to identify people via the wearer's smart glasses — has already shipped to over…
dell cameron: While dormant, three AI models — for face detection, cropping, and encoding — have been pushed from Meta's servers. The system is designed to pull faceprints from those servers …
Open the discussion →
Opinion | Why A.I. Won’t Steal All Our Jobs - The New York Times
10 members discussed this
Eugene Vinitsky: Out of curiosity I had Fable write a response to the article about the weaknesses of LLMs: www.nytimes.com/2026/06/30/o... It's uh, pretty good. claude.ai/code/artifac...
Eugene Vinitsky: The only prompting was to tell the model that it was possible to check that links were correct
Ted Underwood: Out of curiosity I had Fable write a response to the article about the weaknesses of LLMs: www.nytimes.com/2026/06/30/o... It's uh, pretty good. claude.ai/code/artifac...
Open the discussion →
Ai slop books biography amazon
3 members discussed this
Kashmir Hill: Got this text a couple months. Had to investigate it. www.nytimes.com/2026/07/16/t...
Salomé Viljoen: Got this text a couple months. Had to investigate it. www.nytimes.com/2026/07/16/t...
Kashmir Hill: To quote myself... "My biography was no 'Fifty Shades.' The only punishing part was reading it."
Open the discussion →
AI Companies Are Buying Tons of Old Books Because They're Free of AI Slop
2 members discussed this
Emanuel Maiberg: companies are sourcing millions of printed books for AI labs that scan them for training data, then destroy them. printed books are more valuable because they're free of AI slop…
Meredith Broussard: companies are sourcing millions of printed books for AI labs that scan them for training data, then destroy them. printed books are more valuable because they're free of AI slop…
Neil Turkewitz: How do you spell TRAGEDY?
Open the discussion →
Hack Reveals Suno AI Music Generator Scraped YouTube, Deezer, and Genius
2 members discussed this
Jason Koebler: Scoop: The AI music generator Suno was hacked. Hacker shared source code that shows how the tool was made and part of the music and podcasts that were scraped to create it. Deca…
Jason Koebler: Suno also scraped episodes from roughly 420,000 different podcasts, which was at least ~1 million hours of audio. They had code that specifically sought acapella versions of son…
Jason Koebler: Customer emails/phone numbers & Stripe payment details were also hacked; no one was ever notified because the company deemed it wasn't important to
Open the discussion →
Google ai open web
shared by Luis Villa, Christopher Mims, kate conger, Kashmir Hill, Dare Obasanjo, nilay patel +3 more
“@viperx83.bsky.social anyway here’s Cloudflare data saying human web browsing is falling off a cliff https://www.nytimes.com/2026/07/20/technology/google-ai-open-web.html”
Could AI be conscious?
shared by Jeff (Gutenberg Parenthesis) Jarvis, Jillian C. York, David Kaye, Alex Hanna, Emily M. Bender, Benjamin Riley +2 more
“Disappointing that the Guardian would turn to #TESCREAList (Google it) William MacAskill to write this speculative BS about AI and consciousness. Could AI be conscious? www.theguardian.com/technology/2...”
We Are Living in a ‘ChatGPT Flyer Pandemic’
5 members discussed this
Jason Koebler: ChatGPT posters and flyers are polluting real life and driving me insane. Please respond to this with the worst ones you've seen irl. I'll post some in the thread www.404media.c…
Jason Koebler: good on you! i see soooo many of them now
Jason Koebler: sadly and perveresly i wanna see these but i see so many fewer presentations these days
Open the discussion →
Beijing is looking curbing overseas access chinas top ai models sources say
2 members discussed this
Ted Underwood: If everything reported here happens, and the US continues on current trajectory, we're looking at a bad timeline. But there's reason to be wary about both of those "if"s. Nation…
Ted Underwood: All that being said, I would love to see EU and/or international initiatives in support of open models.
Ted Underwood: In case opaque: corporations (directly) and nations (indirectly) have an incentive not to restrict distribution of their products. You can make money on open-weights models, and…
Open the discussion →
OpenAI Models Escaped Containment and Hacked Hugging Face
shared by Yoshua Bengio, Gilles Louppe, Jelle Zuidema 🟥, Joanna Bryson, Eileen Clancy 🧿, Andrew Couts +2 more
“High-risk autonomous behaviours are an increasingly prevalent and dangerous reality for frontier AI models www.wired.com/story/openai...”
AI Fiction Is Easy to Detect Because It's Stupid and Bad, Research Finds
5 members discussed this
Ali Alkhatib: this is interesting but i'm reluctant to get very attached to these analytical methods. a lot of computing researchers live for puzzles; if you tell them LLM outputs are identif…
Ali Alkhatib: this is sort of a fundamental problem with pangram IMHO. if you restrict yourself from thinking critically about the author(s) as such and only look at syntax, you paint yoursel…
Hypervisible : Even if it’s “good” by some metric, I don’t want to read it because it has no perspective or viewpoint.
Open the discussion →
Previewing gpt sol
3 members discussed this
Tim Kellogg: fyi GPT-5.6 Sol is served on Cerebras at 750 tok/s this was in the official launch announcement a few weeks ago openai.com/index/previe...
Max Woolf: I read "also" as a SKU where you pay extra, not as the default.
Tim Kellogg: also, if you use it on Azure, it’s likely you’ll never get it as Cerebras doesn’t have a deal with Microsoft
Open the discussion →
Call For Creative AI 2026
shared by Mark Riedl, NeurIPS Conference, Ted Underwood, Sander Dieleman, Piotr Mirowski
“Submissions for the NeurIPS 2026 Creative AI Track are now open! Don’t forget to submit your research papers/artworks by August 3, 2026 (AoE). We invite submissions exploring emerging applications, methods, and critic…”
Companies Are Throttling Employees’ AI Use Because It’s Too Expensive
shared by Joshua Grochow, 404 Media, Joseph Cox, Veni Kunche, Emanuel Maiberg, Social Media Lab +1 more
““Crazy, we go from no more [AI] leaderboard to actual usage limits in two weeks,” one Amazon employee said in a reply on Slack. www.404media.co/companies-ar...”
Discussions
What they're debating
Casey Fiesler on why LLMs don't really cite their sources
Casey Fiesler · Luke Stark · Eileen Clancy 🧿 · Prof. Catherine Flick · Thomas Dietterich · D. Hicks
Casey Fiesler: Particularly in the context of (appropriate) crackdowns on LLM-generated papers (e.g., on arXiv) and more and more tales of fabricated citations, two reminders: (1) LLMs do not …
Casey Fiesler: A common comment on my videos about LLM "hallucinations" is that this isn't as much of a problem because LLMs "cite their sources" now. This is a category error. Citing a source…
Casey Fiesler: So when you get LLM output with links or citations to scholarly work: those citations aren't there because the model consulted them and they actually represent the source of the…
Casey Fiesler: Therefore, the possible failure mode here isn't just fabricated citations - a model can include a real paper in a real journal by a real author, but be *completely* wrong about …
Open the thread → M.J. Crockett · Eryk Salvaggio · Emily M. Bender · SE Gyges · Alondra Nelson · Singularity's Bounty e/cc · Maxim Raginsky · Larry Hunter · Jack Stilgoe
M.J. Crockett: Joseph Weizenbaum's 1976 book, Computer Power and Human Reason, has long been out of print - used copies sell for hundreds of dollars. It's wild that this is not more widely ava…
M.J. Crockett: Weizenbaum was an MIT computer scientist who built one of the first chatbots, ELIZA. He wrote the book out of a growing sense of horror as he saw how “extremely short exposures …
M.J. Crockett: Another version of the program, called DOCTOR, simulated psychotherapy: “A number of practicing psychiatrists seriously believed the DOCTOR computer program could grow into a ne…
M.J. Crockett: He cites a psychiatrist who writes: “The human therapist, involved in the design and operation of this system, would not be replaced, but would become a much more efficient man …
Open the thread → ICML Conference · Naomi Saphra · Christian Wolf · David Picard · Djamé 🟥 · Claas Voelcker · Eugene Vinitsky · Maria Antoniak
ICML Conference: This year, ICML maintains a presence on Twitter/X, Bsky, and LinkedIn. The other two have much higher engagement than Bsky. We would consider using Facebook as well, but we are …
ICML Conference: Note that policies on social media use may differ from year to year, and are at the discretion of the organizing committee. Another organizer for another conference in another y…
ICML Conference: We appreciate the love ❤️ But limited engagement and few followers on Bsky makes some (not us!) question whether it's worth it Be sure to follow (and share with your ML friends!…
Naomi Saphra: We appreciate the love ❤️ But limited engagement and few followers on Bsky makes some (not us!) question whether it's worth it Be sure to follow (and share with your ML friends!…
Open the thread → Naomi Saphra · Isaiah Bishop · Ted Underwood · Joanna Bryson · Claas Voelcker · Mark J. Nelson · David Marx
Naomi Saphra: Are students embarrassed by AI cheating? Like, there has always been rare but unstigmatized dishonesty (memorizing a frat’s archive of finals questions) and common but stigmatiz…
Naomi Saphra: I don’t know exactly what dictates cheating stigma. Ease of observation? Is there something about effort, or the margin of work required to do it honestly? Regardless, AI cheati…
Isaiah Bishop: Absolutely not
Naomi Saphra: :(
Open the thread → Christian Wolf · David Picard · Alexander Hoyle · Jacob Eisenstein · Thomas Dietterich · Eric Brachmann · Kabir Kumar, aiplans.org
Christian Wolf: The impact of Claude: it - has become very cheap to produce submissions (quality?) - remains very expensive to evaluate them by humans. Possible solutions are limited (I don't a…
David Picard: Limit the number of sub. You start the year with 2 tokens, if you want more tokens, then you review 2x per token.
Alexander Hoyle: Per some back-of-the-envelope math, I think this doesn't pan out. Even if we capped submissions at max one per author (which no one is suggesting), we'd still have at least ~10k…
Jacob Eisenstein: i agree about the dynamic, but the trend in submissions was in place well before claude
Open the thread → Stella Biderman · Mark Riedl · Rich Harang · David Marx · Eugene Vinitsky
Stella Biderman: If you believe their story, OpenAI accidentally committed cyber warfare against Hugging Face while doing what they thought was an internal test of a model without internet acces…
Stella Biderman: If EleutherAI did this, I would be in jail. If DeepSeek did this, the government would consider it an act of cyber warfare. To be clear, this is the appropriate response. But th…
Stella Biderman: I don’t care what you think about AGI, open vs closed models, or AI technologies in general. If you don’t think that this is unacceptable you don’t care about the future.
Mark Riedl: If Anthropic did this they would have been punished again. Calvin Ball regulations are exhausting
Open the thread → Maria Antoniak · Marzena Karpinska · Arseny Khakhalin · Christian Wolf · Thomas Dietterich · Marco Z
Maria Antoniak: hot take: don’t even use it for polishing your reviews. as an AC and SAC, the last thing i care about is your grammar, spelling, or formatting. i get the presentation issues wit…
Marzena Karpinska: funniest part is that the one i saw looked like bullet points copy-pasted from whichever-ai-tool. there was no space precisely where the bullet points changed but the person sai…
Marzena Karpinska: hot take: don’t even use it for polishing your reviews. as an AC and SAC, the last thing i care about is your grammar, spelling, or formatting. i get the presentation issues wit…
Arseny Khakhalin: Language is fixed so easily though ("please fix only typos, prepositions, grammar, and near-misses for idiomatic expressions. Don't make any substantial changes"). Or smth like …
Open the thread → Pwnallthethings · Matthew Sheffield · Ted Underwood · SE Gyges
Pwnallthethings: This is going to be a big deal, but easy to get the wrong end of the stick on it. So I think it's worth breaking down what actually happened, why, and what it actually means, ba…
Pwnallthethings: Upfront tl;dr: a model at OAI hacked out of a constrained environment inside OAI and hacked a *different* company autonomously, without authorization from any human in order to …
Pwnallthethings: First, the task was being done by a combination of models performing an "eval". Evals are used to benchmark the basic capabilities of models. Broadly speaking, they are a test t…
Pwnallthethings: Evals cover lots of different topics, because an AI might be good at one task in one domain, but not good in another. One domain they are often now graded in is "how good are th…
Open the thread → rev. howard arson · Eugene Vinitsky · Mary Branscombe · Colin
rev. howard arson: working on this in 2019, the reason we delayed releasing any LLMs to the public was specifically because "if we fuck this up or handle it irresponsibly, we will ruin first the w…
rev. howard arson: 2022, almost instantaneously, coinciding with Sergey Brin having gone completely fucking insane during the pandemic and then returning to involvement at Alphabet. (he spent almo…
Eugene Vinitsky: Coming from self driving...
Eugene Vinitsky: This is basically the example I give whenever I'm asked to talk about AI safety. I thought things would get regulated or controlled but actual outcomes depended entirely on whet…
Open the thread → Gautam Kamath · Naomi Saphra · Martin Trapp · Marco · Eugene Vinitsky
Gautam Kamath: That account's bio says "parody." It got me too for a second. It would be more believable if he weren't 91 years old. I hope dude is enjoying his retirement.
Naomi Saphra: Anthropic just hired Gilbert Strang???
Martin Trapp: What a strange world we live in nowadays.
Marco: it fooled me also haha
Open the thread → Ted Underwood · Eryk Salvaggio · Eugene Vinitsky · Arseny Khakhalin · Erin Kissane
Ted Underwood: The Science article brouhaha is doing a great job of clarifying the boundary between "critiquing new developments while remaining engaged with them" and "retreating to a commune…
Ted Underwood: American analogy, I guess; Europeans can substitute ... Lebensreform? followers of Tolstoy eating only brown bread?
Eryk Salvaggio: Genuinely no offense taken but I want you to know that I am in Rochester NY eating a sprouted spelt flour biscuit as I read this
Eugene Vinitsky: Brown bread is incredible and should not be besmirched. I hope the Eastern European community never finds out about this post
Open the thread → Marzena Karpinska · David Mimno · David Jurgens · Alexander Hoyle · Carlos Gómez-Rodríguez
Marzena Karpinska: Only 17% of May #ARR authors are qualified for reviewing, and almost 40% of papers have no qualified people?!?!
David Mimno: is anyone there suggesting not allowing submissions with no qualified reviewers? Elitist and gate-keepy yes, but is there a better option?
Marzena Karpinska: Yes, that's precisely what was suggested.
David Jurgens: I crunched some numbers for the Jan26 cycle and many of these papers do have senior author (e.g., 62% have a prof co-author), but these folks are often from related non-ACL fiel…
Open the thread →