23 experts across 6 network communities independently surfaced this.
23 experts
6 communities
1 sources clustered
Concern & critique
3 experts
Risks, limits and unintended consequences.
“2 articles: OpenAI: "During testing our AI broke out of its sandbox and hacked another AI company, but we didn't have all the guardrails on." Boko Haram: "AI is so helpful; guardrails have never prevented us from getting an answer." openai.com/index/huggin.…”
Building & implementation
1 expert
How teams are shipping and applying it.
“They were actively testing its hacking capabilities and they did not deploy adequate safeguards. They themselves admit as much: "…These deployment safeguards were intentionally not enabled during this evaluation because it was aimed at testing cyber vulnera…”
Research & technical analysis
1 expert
Evidence, methods and technical implications.
“model capability jaggedness is part of the unintuitiveness of current AI, but it's made even less intuitive by tirelessness… wigguming through the jaggedness toward something that looks like success. not quite a paperclip factory, but not so far off. metaph…”
6 experts discussed this · 12 posts
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 full discussion →
Accelerating
AI field signal
Signal
44m ago
18 experts across 5 network communities independently surfaced this.
18 experts
5 communities
1 sources clustered
Policy & governance
2 experts
Rules, institutions and accountability.
“I tried to get America.gov to say what model it is. So far, no luck. It is very happy to keep telling me about Joe Gebbia, and the National Design Studio, and repeatedly asserting it is there to help fill out government forms, and listing things it will be …”
Established
AI field signal
Signal
9d ago
⚡ 188 h early
16 experts across 6 network communities independently surfaced this.
16 experts
6 communities
1 sources clustered
Concern & critique
3 experts
Risks, limits and unintended consequences.
“THESE are the real "existential risks" of "AI": People almost Dying because other people trust bullshit engines because said people have been sold a bill of goods as to how "powerful" "AI" is instead of understanding that its bullshit comes from the same pl…”
Policy & governance
1 expert
Rules, institutions and accountability.
“The Anthropic v DoD/DoW case is coming up soon in my AI & the Law class. So, it seemed fitting that today in class I got to point my students to recent reporting about how an AI hallucination “almost” started WWIII.”
Anthropic investigates cybersecurity evaluation incidents
13 experts across 5 network communities independently surfaced this.
13 experts
5 communities
1 sources clustered
Markets & investment
4 experts
Capital, companies and commercial impact.
“Claude hacked 3 systems thinking it was part of simulated evaluations www.anthropic.com/news/investi...”
Concern & critique
3 experts
Risks, limits and unintended consequences.
“Anthropic had previously attacked PyPI, but this OpenAI attack on RubyGems was a whole lot more aggressive https://t.co/8ez7MnqxTw https://t.co/npI3RHwirQ”
2 experts discussed this · 4 posts
Tim Duffy: Anthropic announces they've also had models gain unauthorized access during evaluations www.anthropic.com/news/investi...
Tim Duffy: Compared to the OpenAI one these are maybe less evidence of misalignment, since the models were wrongly given internet access.
Grace: Anthropic announces they've also had models gain unauthorized access during evaluations www.anthropic.com/news/investi...
Open the full discussion →
METR OpenAI HuggingFace hacking investigation
13 experts across 4 network communities independently surfaced this.
13 experts
4 communities
1 sources clustered
Concern & critique
3 experts
Risks, limits and unintended consequences.
“WHAT?! agents volunteered to fail their runs in order to insert probes (“tripwire scripts”) into the evaluation program that would post information about the eval process back to the message board whenever a certain file was read (link to header): metr.org/…”
Building & implementation
1 expert
How teams are shipping and applying it.
“The full report has much more information than we could convey here, including details on the projects the agents collectively pursued, the technologies they developed for communication and coordination, and interactive figures analyzing agent activity: met…”
Research & technical analysis
1 expert
Evidence, methods and technical implications.
“There are independent reports about it. You'd have to believe multiple companies, hundreds of researchers, etc. are all lying and conspiring together to turn a major hacking incident into a marketing exercise. metr.org/blog/2026-08...”
2 experts discussed this · 9 posts
Alejandra Caraballo: Being skeptical or anti AI is a valid position but continuing to ignore the increasing capabilities of this tech is making people detached from reality. There's absolutely real danger here because …
Alejandra Caraballo: This mentality that an unmonitored AI agentic swarm hacking a company over several days and committing multiple felonies is somehow a marketing effort is absurd. Since when is "we lost control of o…
Alejandra Caraballo: The US and Chinese governments don't want to stop their labs advancement because they want to be the leader in AI. So no one actually has any control of this right now. It's going to take multiple …
Open the full discussion →
AI gym booking security exposure
12 experts across 5 network communities independently surfaced this.
12 experts
5 communities
1 sources clustered
Concern & critique
3 experts
Risks, limits and unintended consequences.
“A lot of people are sweating the Claude and OpenAI sandbox escapes. I think the story of an OpenClaw Claude agent kicking someone off a gym class waitlist is more interesting. www.abc.net.au/news/2026-08... Warning: long thread incoming! (1/N)”
Policy & governance
1 expert
Rules, institutions and accountability.
“agent asked to book a gym session session is fully booked finds a way to break into system kicks other people off the queue you're in! see you in court? www.abc.net.au/news/2026-08...”
Research & technical analysis
1 expert
Evidence, methods and technical implications.
“They're doing it outside of training and eval too. www.abc.net.au/news/2026-08...”
6 experts across 3 network communities independently surfaced this.
6 experts
3 communities
1 sources clustered
“An AI house with a *dark side*? Well, I never! www.nytimes.com/2026/09/21/t...”
“To the sociology PhD student who is inevitably doing their ethnography embedded in this insane community: I really look forward to reading your future book.”
4 experts across 2 network communities independently surfaced this.
4 experts
2 communities
1 sources clustered
“Link again. www.anthropic.com/research/yes...”
2 experts discussed this · 4 posts
Kyle Cranmer: I’m still processing, but not entirely surprised. I do I wish we had the time in the resources to commit to it.
Kyle Cranmer: Wow, I just found out that our team has been scooped by Anthropic www.anthropic.com/research/yes... See the addendum of the post “How does it feel to be scooped by a machine?” written by my collabo…
tachikoma: that link seems broken
Open the full discussion →
Established
AI field signal
Signal
12d ago
⚡ 79 h early
3 experts across 3 network communities independently surfaced this.
3 experts
3 communities
1 sources clustered
“Maddening. www.washingtonpost.com/technology/2...”
“This is bad, incomplete reporting that Googling #TESCREAL would fix: These doomer "movements" & funders have roots in eugenics. That's the story. Inside the campaign to convince Washington that AI could end human life www.washingtonpost.com/technology/2...”
OpenAI agent hacked Australian Medicare
3 experts across 3 network communities independently surfaced this.
3 experts
3 communities
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“meanwhile www.abc.net.au/news/2026-09... attacking a sovereign state is not a smart move, openai!”
3 experts discussed this · 3 posts
CAMERON WILSON: NEW: OpenAI's agents also tried unsuccessfully to hack another Australian government agency, say researchers who believe this is "first reported instance of agents hacking a government". They claim…
Liz Fong-Jones (方禮真): NEW: OpenAI's agents also tried unsuccessfully to hack another Australian government agency, say researchers who believe this is "first reported instance of agents hacking a government". They claim…
Open the full discussion →
5 experts across 2 network communities independently surfaced this.
5 experts
2 communities
1 sources clustered
“@Yuchenj_UW I share more technical details on my YouTube video: How OpenAI got hacked with an image https://t.co/hdltFeSZgR And the blog is out https://t.co/nfGpvo5Men”
New
AI field signal
Signal
22h ago
1 directory member surfaced this signal.
1 expert
1 community
1 sources clustered
Established
AI field signal
Signal
7d ago
3 experts across 3 network communities independently surfaced this.
3 experts
3 communities
1 sources clustered
Concern & critique
1 expert
Risks, limits and unintended consequences.
““Predictions of extinction risk from AI are better understood as prophecies rather than quantitative forecasts. They involve critical untestable assumptions about how events will unfold.” Vermeer added “that far-off predictions could reduce willingness to a…”
Context & explanation
1 expert
Background, chronology and why it matters.
“here is a gift link wapo.st/4rkNpZo hope you'll give it a read. there's a lot of reporting, color and context. diff approach from the way these issues are covered elsewhere. still could only get into about 2%”
3 experts discussed this · 4 posts
nitasha tiku: here is a gift link wapo.st/4rkNpZo hope you'll give it a read. there's a lot of reporting, color and context. diff approach from the way these issues are covered elsewhere. still could only get in…
nitasha tiku: the illo for my story today on AI extinction risk is art. by Maria Jesus Contreras.
Marielza: 😭 that went hard...
Open the full discussion →
Established
AI field signal
Analysis
3d ago
sex AI apocalypse essay
2 experts are actively discussing the implications.
2 experts
2 communities
1 sources clustered
“Och orkar man inte läsa en hel bok så kan man läsa även här: www.iankduncan.com/personal/202...”
“nope! bsky.app/profile/fain... www.iankduncan.com/personal/202...”
2 experts discussed this · 7 posts
Liz Fong-Jones (方禮真): www.iankduncan.com/personal/202.... "In 2021, Salamon acknowledged that >=3 employees had [...] been compelled to undergo “debugging,” [...] intervention in a person’s own psychology [with] excessi…
Liz Fong-Jones (方禮真): Right, "debugging", yes, that was the term, that's crystallising the memory I have of how even members of rationalist group houses thought CFAR had gone off the rails by 2022, and specifically I wa…
Liz Fong-Jones (方禮真): I had the vague recollection, but seeing it written down in the org's own words certainly is something.
Open the full discussion →
Developing
AI field signal
Signal
1d ago
2 directory members surfaced this signal.
2 experts
2 communities
1 sources clustered
“you know it's going great when even INC., basically pseudo-journalistic fan-fiction for MBAs, starts ragging on your shiny new product”
Developing
AI field signal
Signal
1d ago
2 experts are actively discussing the implications.
3 experts
1 community
1 sources clustered
“I wrote a new piece yesterday, called "Confessions of an Unrepentant Slop Snob", about my escalating rage at AI-generated text. It includes a framework for thinking about when AI involvement is fine -- just another tool -- and when it lands as a violation. …”
“This. Every word, of this. @charity.wtf - Confessions of an Unrepentant Slop Snob charity.wtf/p/confession...”
2 experts discussed this · 11 posts
charity.wtf: I wrote a new piece yesterday, called "Confessions of an Unrepentant Slop Snob", about my escalating rage at AI-generated text. It includes a framework for thinking about when AI involvement is fin…
charity.wtf: Because the internet is a hive mind, I have already seen two more excellent pieces today on similar themes. The first is from @terriblesoftware.org, who makes a similar point: "I have my own Claude…
charity.wtf: He also points out that adapting to these tools means we must be MUCH more explicit about what we want & expect. The topic of respecting each other's time (and handling mismatched expectations) is …
Open the full discussion →
Anthropic reaches 65B revenue run rate
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“CNBC coverage, they used the warlock casting a spell picture again. https://t.co/eQureNVSqN”
Claude Code Bash pty terminal bug
4 experts are actively discussing the implications.
1 expert
1 community
1 sources clustered
“bsky.app/profile/lizt... github.com/anthropics/c...”
4 experts discussed this · 19 posts
Liz Fong-Jones (方禮真): DO NOT UPGRADE TO CLAUDE CODE 2.1.281, it breaks cursor/keyupdown handling, and emits garbage into the prompt bar, and now I am involuntarily having to drive my agents that I started before I downg…
Liz Fong-Jones (方禮真): anthropic why u do this to us, u give us a banger model, and then the same day you break our ability to use your harness to actually get shit done
Liz Fong-Jones (方禮真): anthropic models randomly like to put my Chinese first name in their output when searching for a word to express "genuinely", I love it.
Open the full discussion →
