“Hack reveals that Suno (AI music generator) was trained on copyrighted music www.404media.co/hack-reveals...”
“The lawsuits have made clear that Suno did train on huge amounts of copyrighted works, but the hacked data shared with 404 Media sheds more light on how Suno scraped songs from the internet and where it took them from. @jasonkoebler.bsky.social reports. Rea…”
2 experts discussed this · 25 posts
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. Decades worth of music, …
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 songs on YouTube for vo…
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
“The UN Independent International Scientific Panel on AI's assessment is a reference point for AI governance worldwide, writes Eryk Salvaggio. His concern: a definition that obscures human accountability "will carry that blind spot into the policies it infor…”
“Great job of revising and updating my 2023 Wired article, @eryk.bsky.social ! I especially prefer the title (though I never chose a Wired article title, those come from higher powers.) But also the more detailed account of the definition drift. #AIEthics @t…”
2 experts discussed this · 4 posts
Joanna Bryson: Great job of revising and updating my 2023 Wired article, @eryk.bsky.social ! I especially prefer the title (though I never chose a Wired article title, those come from higher powers.) But also the…
Joanna Bryson: This one was great in its time though – I've been told it helped with the #AIAct by some of the people who were holding the "pens" www.wired.com/story/artifi... 2/2
Eryk Salvaggio: Apologies! I hadn’t seen this excellent 2022 accounting of the process when I was reviewing for this — my analysis’ “lit review” was more mired in the source docs. Yours is an excellent accounting …
“If you look at all of Biden’s bullets you see this LLM pattern show up in some form or another — LLM templates aren’t a Mad Libs-style template, but they tend to repeat the same form of sentence. I wrote more here, if it’s of interest: mail.cyberneticforest…”
“Not using AI? You must pay an AI company $$$ to certify you, and then accept their AI-generated recommendations for how to not sound like AI.”
2 experts discussed this · 13 posts
Eryk Salvaggio: Not picking on OP or endorsing the trend, but the number of people sharing and being genuinely moved by Hunter Biden’s Claude output is a good window into how audiences see and respond to LLM-gener…
Eryk Salvaggio: A microcosm in the quote-posts, too. And look, I’m not using Pangram (I don’t have to) or calling for anyone to shun anybody. Just noticing something.
Eryk Salvaggio: Can’t say it’s Claude specifically but it has lots of telltale signs. It’s bullets, for one. Two: the terse “The noun verbed” opener. 3, the negative parallelism (“not x, not y. Z.”) repeated twice…
“2/2 bsky.app/profile/oliv... pdf here: doi.org/10.5281/zeno...”
3 experts discussed this · 24 posts
Olivia Guest: New preprint! So honoured to have worked with Ishani Ray on providing vital critique on "AI literacy" & correctives for those interested in understanding why "just use the newest version of an LLM"…
Olivia Guest: Maybe many mid-career and above academics have my experience, and so can relate, but it never ever gets any less impressive, overwhelming, or old when a student appears our of nowhere (notice our g…
Olivia Guest: Anyway, back to the work itself (from the abstract): > Critical discourse on large language models (LLMs) has bifurcated between epistemic dismissal [...] and pragmatic accommodation that treats LL…
“Returning to this thread after a busy week, now that I've had time to read your "stochastic flocks" piece. Link here, just in case anyone else is following: www.techpolicy.press/stochastic-f... >>”
“This article managed to annoy all sides, but there are good technical explanations of why this kind of system would be capable of finding vulnerabilities & writing functional code. Rigorous critiques of AI are systemic, it holds up even when it “does things…”
“What is Cultural AI and the objects of its study? My paper addresses literary tools that can be borrowed to address cultural AI issues. The paper is on now ArXiv and forthcoming in Modern Fiction Studies in Spring 2027. arxiv.org/abs/2607.02369”
“Nina Begus World Wide Models: Literary Tools for Cultural AI https://arxiv.org/abs/2607.02369”
“This pre-print lays out a method for analyzing these systems (here, the latent diffusion model) by looking at its individual structural components to understand “the question” each part of the process was meant to “answer” and how that influences the output…”
“For decades we regarded the Internet as infrastructure used primarily by people. Following the OpenAI/Hugging Face incident, Konstantinos Komaitis argues AI agents challenge that assumption—and AI governance can no longer stay separate from Internet governa…”
“Jill Lepore wrote about automated art machines, not quite what you're looking for but part of the lineage, I'd say! www.newyorker.com/magazine/202...”
Dan Greene : Yeah I think also of people like Radway, writing a book about romance readers at a moment no one gave two shits
Eryk Salvaggio: Jill Lepore wrote about automated art machines, not quite what you're looking for but part of the lineage, I'd say! www.newyorker.com/magazine/202...
“This one was great in its time though – I've been told it helped with the #AIAct by some of the people who were holding the "pens" www.wired.com/story/artifi... 2/2”
“Well this reads as quite the combo breaker, at least with agents powered by LLMs ~3 months ago, developer-submitted and LLM-created AGENT md files don't seem to improve task performance for python projects, but do burn ~20% more tokens arxiv.org/abs/2602.11988”
““The dashboard known as the U.S. State AI Policy Tracker is free to the public, collecting information on the more than one million total bills being discussed by state lawmakers and compiling a list of AI-related legislation.””
“Gonna reference @lukestark.bsky.social like I always do. Facial recognition is the “plutonium of ai.” There’s no “benign” use, particularly where Facebook is involved.”
“💯 I am all in favor of critiquing *ChatGPT Search* and what gets obscured by the design decisions — I have written about this and I think I coined the term “shadow prompting!” www.techpolicy.press/shining-a-li...”
“Another recommended paper on the "Rogue AI" frame and the assumptions it makes: interdisciplinary perspectives that consider issues of infrastructure, social agency, Big Tech power position and politics remain scarce." link.springer.com/article/10.1...”
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 available given how muc…
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 to a relatively simp…
Eryk Salvaggio: Good thread if you want some context on the OpenAI “hacking” event you may have heard about.
Eryk Salvaggio: Key point re: mainstream AI reporting about “breaking containment” or whatever — the failure to isolate the model *is* the *incident.* OpenAI wants to sell you safe LLMs but they left their chicken…
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 in upstate New York…
Ted Underwood: American analogy, I guess; Europeans can substitute ... Lebensreform? followers of Tolstoy eating only brown bread?
Eryk Salvaggio: We really must acknowledge that when you see “AI” making half-assed work it’s because someone looked at it, said “good enough” and shared that half-ass work just to stand by it, point out that it w…
Eryk Salvaggio: Like yes Gen AI UX is designed to encourage half-assery but Netflix is designed to keep you binge watching and we would not tolerate an academic staying up all night watching the Great British Bake…
We use essential cookies to keep the site working (login, form security). With your permission, we also use analytics cookies to understand how you use the site.
Privacy policy
