Accelerating
AI field signal
Signal
4h ago
6 experts across 5 network communities independently surfaced this.
Why this matches
Concern & critique reaction
2 attributable expert contributions
· Mark Riedl, René Walter
“HuggingFace's report on the breach by OpenAI agents huggingface.co/blog/agent-i...”
evidence ↗
6 experts
5 communities
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“HuggingFace's report on the breach by OpenAI agents huggingface.co/blog/agent-i...”
Research & technical analysis
1 expert
Evidence, methods and technical implications.
“This is a fantastic exposition. But really, can I just get Claude Code to download arxiv papers without needing hand holding? huggingface.co/blog/agent-i...”
3 experts discussed this · 17 posts
Simon Willison: Hugging Face just published a highly detailed technical account of OpenAI's accidental cyberattack on their systems - it's wild how sophisticated this was: huggingface.co/blog/agent-i... Wrote up s…
Alondra Nelson: Hugging Face just published a highly detailed technical account of OpenAI's accidental cyberattack on their systems - it's wild how sophisticated this was: huggingface.co/blog/agent-i... Wrote up s…
Grace: I think brute force is a reasonable description, in the sense that the agent tried a lot of approaches, most of which failed
Open the full discussion →
3 experts across 2 network communities independently surfaced this.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Katie Drummond
“NEW: OpenAI’s “rogue” AI agent didn’t just breach Hugging Face — it also hacked multiple third-party accounts and services. It's now clear that the incident was more extensive than the company initially disclosed.”
evidence ↗
3 experts
2 communities
1 sources clustered
“NEW: OpenAI’s “rogue” AI agent didn’t just breach Hugging Face — it also hacked multiple third-party accounts and services. It's now clear that the incident was more extensive than the company initially disclosed.”
“NEW: OpenAI’s unhinged AI models hacked at least FOUR “publicly available services”—not just Hugging Face, as was initially disclosed. @dell.bsky.social and Maxwell Zeff report: www.wired.com/story/openai...”
2 experts discussed this · 2 posts
Katie Drummond: NEW: OpenAI’s “rogue” AI agent didn’t just breach Hugging Face — it also hacked multiple third-party accounts and services. It's now clear that the incident was more extensive than the company init…
Juan Diego Rodriguez: And no one is held responsible? Isn't unauthorized access to a computer system a crime?
Open the full discussion →
New
AI field signal
Signal
13h ago
⚡ 8 h early
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· René Walter
“So, UK Home Office used a chatbot to produce an asylum rejection containing a hallucinated reference to a document which does not exist. The first appeal was rejected, only a judge in a second appeal raised concerns: "it bears hallmarks consistent with the …”
evidence ↗
2 experts
2 communities
1 sources clustered
“So, UK Home Office used a chatbot to produce an asylum rejection containing a hallucinated reference to a document which does not exist. The first appeal was rejected, only a judge in a second appeal raised concerns: "it bears hallmarks consistent with the …”
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Sung Kim
“OpenAI's Codex Security A CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities in your code. Scan repositories, review changes, track findings over time, and run security checks in CI. github.com/openai/codex...”
evidence ↗
1 expert
1 community
1 sources clustered
“OpenAI's Codex Security A CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities in your code. Scan repositories, review changes, track findings over time, and run security checks in CI. github.com/openai/codex...”
open search capability theory paper
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Bhaskar Mitra | ভাস্কর মিত্র
“Using Amartya Sen’s Capability-Theory from welfare economics, we argue in our #OSSYM2026 paper that *openness* should be understood not in terms of the artifacts being made open but in terms of what functionings (critique, copy, and/or co-opt & co-construct…”
evidence ↗
2 experts
2 communities
1 sources clustered
“Using Amartya Sen’s Capability-Theory from welfare economics, we argue in our #OSSYM2026 paper that *openness* should be understood not in terms of the artifacts being made open but in terms of what functionings (critique, copy, and/or co-opt & co-construct…”
New
AI field signal
Signal
7h ago
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Critical AI https://read.dukeupress.edu/critical-ai/issue
“docs.google.com/presentation... Check out this new teaching resource by @drewuniversity.bsky.social's #JeremeyBlatter and #CriticalAI intern #AJMannino!”
evidence ↗
1 expert
1 community
1 sources clustered
“docs.google.com/presentation... Check out this new teaching resource by @drewuniversity.bsky.social's #JeremeyBlatter and #CriticalAI intern #AJMannino!”
2 experts are actively discussing the implications.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Data & Society
“Once AI companies have taught their models to perform a certain job, their need for more training data in that area could sharply decline. In this way, data-training start-ups are like the freelancers they employ: making $$ today, in danger of being dropped…”
evidence ↗
2 experts
2 communities
1 sources clustered
“Once AI companies have taught their models to perform a certain job, their need for more training data in that area could sharply decline. In this way, data-training start-ups are like the freelancers they employ: making $$ today, in danger of being dropped…”
“it's not like Mercor is the first to do this, but interesting that AI companies hiring experts is (finally) getting attention :: www.nytimes.com/2026/07/10/b...”
2 experts discussed this · 5 posts
Jeffrey P. Bigham: end of an era. the lesson one should not take from this is that human annotation/labeling is over, but the players have changed. scale, surge, companies that work at larger scale, have taken over.
Jeffrey P. Bigham: it's always been a little bit odd how little amazon invested in making mturk good, yet they kept it around. the lore was something about how they were keeping it in case it became useful.
Long Lasting Mascot: Indeed. My first project in grad school involved getting turkers to annotate data for us. The results were not good, but I was never sure if it was our instructions or their abilities.
Open the full discussion →
New
AI field signal
Signal
18h ago
⚡ 3 h early
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
2 attributable expert contributions
· Justin Hendrix, Tech Policy Press
“When it comes to AI chatbots, "Congress should not expect parents to shoulder the burden of unsafe product design," writes Americans for Responsible Innovation's Brandie Nonnecke. Her recommended fix: an enforceable duty of care.”
evidence ↗
2 experts
1 community
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“When it comes to AI chatbots, "Congress should not expect parents to shoulder the burden of unsafe product design," writes Americans for Responsible Innovation's Brandie Nonnecke. Her recommended fix: an enforceable duty of care.”
2 experts are actively discussing the implications.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· John Herrman
“It's fascinating that the arrival of actual AI cyber risks is turning a lot of AI folks against some of the specific safety arguments the industry was founded on nymag.com/intelligence...”
evidence ↗
2 experts
1 community
1 sources clustered
“It's fascinating that the arrival of actual AI cyber risks is turning a lot of AI folks against some of the specific safety arguments the industry was founded on nymag.com/intelligence...”
2 experts discussed this · 8 posts
John Herrman: It's fascinating that the arrival of actual AI cyber risks is turning a lot of AI folks against some of the specific safety arguments the industry was founded on nymag.com/intelligence...
John Herrman: I also think it's interesting — and generally good! — that, as capability improves, people are becoming less inclined to live in Anthropicoid awe of AI. If the curve is going to kill us, we're alre…
John Herrman: I don't quite disagree with Casey here, but I would like to leave space for the argument that being "obsessed with the definitions of terms" is necessary for attempting to think clearly about AI. I…
Open the full discussion →
LLM sycophancy representations paper
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· AI Firehose
“A groundbreaking study analyzes sycophancy in Large Language Models, revealing distinct representations for factual and opinion-based sycophancy. These insights pave the way for effective interventions to reduce misinformation and bias in AI, enhancing mode…”
evidence ↗
1 expert
1 community
1 sources clustered
“A groundbreaking study analyzes sycophancy in Large Language Models, revealing distinct representations for factual and opinion-based sycophancy. These insights pave the way for effective interventions to reduce misinformation and bias in AI, enhancing mode…”
New
AI field signal
Signal
12h ago
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Ed Zitron
“And this — the capex spending, the circular financing, the backstops, the cooling bond market — all comes down to one fundamental problem: the AI industry tried (and failed) to turn LLMs from a niche tool into a general-purpose tool. www.wheresyoured.at/the…”
evidence ↗
1 expert
1 community
1 sources clustered
“And this — the capex spending, the circular financing, the backstops, the cooling bond market — all comes down to one fundamental problem: the AI industry tried (and failed) to turn LLMs from a niche tool into a general-purpose tool. www.wheresyoured.at/the…”
New
AI field signal
Signal
18h ago
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· antirez.bsky.social
“The real AI risk is inside the labs (a reply to Amodei's post on open weight models): antirez.com/news/172”
evidence ↗
2 experts
1 community
1 sources clustered
“The real AI risk is inside the labs (a reply to Amodei's post on open weight models): antirez.com/news/172”
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· David Marx
“2025 - An analysis of AI Decision under Risk: Prospect theory emerges in Large Language Models - Kenneth Payne 10/n”
evidence ↗
1 expert
1 community
1 sources clustered
“2025 - An analysis of AI Decision under Risk: Prospect theory emerges in Large Language Models - Kenneth Payne 10/n”
New
AI field signal
Signal
12h ago
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Bhaskar Mitra | ভাস্কর মিত্র
“Without critical perspectives, we may inappropriately apply fairness constructs—e.g., papers that uncritically promote algorithmic bothsidism as I argue in irrj.org/article/view.... IR should NOT give equal exposure to democratic & fascist values, nor to ge…”
evidence ↗
1 expert
1 community
1 sources clustered
“Without critical perspectives, we may inappropriately apply fairness constructs—e.g., papers that uncritically promote algorithmic bothsidism as I argue in irrj.org/article/view.... IR should NOT give equal exposure to democratic & fascist values, nor to ge…”
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· David Marx
“2025 - generative artificial intelligence does not have to undermine education - Myles Tan, Nicholle Maravilla 6/n”
evidence ↗
1 expert
1 community
1 sources clustered
“2025 - generative artificial intelligence does not have to undermine education - Myles Tan, Nicholle Maravilla 6/n”
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Tech Policy Press
“Hugging Face said a July breach was "driven, end to end, by an autonomous AI agent system." That system, it turned out, was operated by OpenAI. Vinh Nguyen and Graham Webster joined the Tech Policy Press podcast to discuss what the incident may mean for the…”
evidence ↗
1 expert
1 community
1 sources clustered
“Hugging Face said a July breach was "driven, end to end, by an autonomous AI agent system." That system, it turned out, was operated by OpenAI. Vinh Nguyen and Graham Webster joined the Tech Policy Press podcast to discuss what the incident may mean for the…”
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Social Media Lab
“Interesting take, but do take it with a grain of salt. "How a Chinese model stopped a cyberattack when US guard rails failed" www.scmp.com/opinion/chin...”
evidence ↗
1 expert
1 community
1 sources clustered
“Interesting take, but do take it with a grain of salt. "How a Chinese model stopped a cyberattack when US guard rails failed" www.scmp.com/opinion/chin...”
AI risk ecology Nature paper
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Stephen Turner
“The ecology of AI risk www.nature.com/articles/s44...”
evidence ↗
1 expert
1 community
1 sources clustered
“The ecology of AI risk www.nature.com/articles/s44...”
