Expert attention map

The Who's Who of AI

What credible people across AI noticed, why it matters, and where the field is converging or disagreeing.

2,364 searchable experts 2,966 tracked across all sources
Build my stream Create my edition
Filter the conversation Who is saying what?

Combine a professional role with a reaction lens. Both must match the same attributed contribution.

Clear all
Active evidence filter

Showing developments with attributable Concern & critique reactions.

New Network reaction maps

See how experts are reacting—not just what they shared.

Posts are grouped by conversation and tone. Select a lens to filter the stream; these are never permanent labels on people.

Concern & critique Policy & governance Research & technical analysis
2 live reaction maps in this view Jump to the evidence ↓
Multiple readings Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident Concern & critique · 3 attributable reactions Multiple readings Protecting our FLOSS commons from LLMs — Codeberg News Concern & critique · 4 attributable reactions
Debated

The questions experts are actively pulling apart

One card per development. Sources are clustered; reaction bundles describe these posts, never the people behind them.

Accelerating AI field signal Signal 5h ago

Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident

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
Mark Riedl Simon Willison Alondra Nelson Suresh Venkatasubramanian René Walter Andrea Lathrop
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 →
Primary source huggingface.co ↗
Developing Models & releases Analysis 1d ago
Codeberg protects FLOSS commons from LLMs

Protecting our FLOSS commons from LLMs — Codeberg News

7 experts across 4 network communities independently surfaced this.

Why this matches Concern & critique reaction 2 attributable expert contributions · Shriram Krishnamurthi, Armin Ronacher
“I respect Codeberg's right to take a position on this, and I understand why they might feel this way, but this is kinda' crazy. Glad I'm not using it, because I'd have to move some projects off of it. The article is also just confused in some ways. blog.cod…” evidence ↗
7 experts 4 communities 1 sources clustered
Olivia Guest Michael Ekstrand CJ Shriram Krishnamurthi Armin Ronacher Jed Brown Amy Hoy
How the network is reacting Experts are approaching this through 2 distinct lenses.
Multiple readings

Concern & critique

2 experts

Risks, limits and unintended consequences.

“I think this is a very bad move, and the people behind Codeberg should re-consider there stance. blog.codeberg.org/protecting-o...”

Policy & governance

2 experts

Rules, institutions and accountability.

“Codeberg updates ToU to disallow LLM training and many LLM-generated projects, and wrote a thorough explanation: blog.codeberg.org/protecting-o... Regardless of the merits of AI code generation, AI trainers have got to get their scraper bots under control a…”

2 experts discussed this · 10 posts
Amy Hoy: this IS interesting, and it’s something i see a lot on subreddits for specific areas of interest (mac apps, note taking) as well as aspiring-entrepreneur ones lots of fancy-looking projects with ze…
Amy Hoy: i mean as somebody who coaches aspiring entrepreneurs, a lot of them spend way too much time on the idea or prototype and nothing on building a customer base but it usually takes more dedication an…
Amy Hoy: it’s the same old wantrepreneur failure pattern but with an acme rocket strapped to its ass
Open the full discussion →
Primary source blog.codeberg.org ↗
New Culture, work & education Signal 16h ago
⚡ 385 h early

Ai white collar jobs

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
Data & Society Jeffrey P. Bigham

“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 →
Primary source nytimes.com ↗
New Models & releases Signal 16h ago

The OpenAI Hack Scrambles the AI Race

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
Casey Newton 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...”

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 →
Primary source nymag.com ↗
Accelerating Agents & robotics Signal 6h ago
⚡ 3 h early

OpenAI’s Rogue AI Agent Hacked More Than Just Hugging Face

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
Katie Drummond Andrew Couts Veni Kunche

“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 →
Primary source wired.com ↗
More conversations

What experts are discussing without an anchoring article

4 experts · 7 posts · 13h ago
Ted Underwood Meredith Martin Shubhendu Trivedi Marco Z
Matches Concern & critique
“Yes, of course. I used to think of it as the primary motivation, and it likely still is. But I think that that underestimates the well-meaning but dangerous utopia-dystopia imagination induced mega…” evidence ↗
Reaction converging Policy & governance · 3
Ted Underwood: On reflection I’m not sure I respect the candor either, because, does Amodei have any expectation this will persuade ppl in tech? Surely not. Perhaps the actual audience for this is policy makers i…
Shubhendu Trivedi: Over time I have come to the conclusion that a lot of this isn't as much about regulatory capture and becoming a designated economic tollbooth (although there's that too). It's more like they reall…
Open the thread →
3 experts · 12 posts · 6h ago
Ted Underwood jake tachikoma
Matches Concern & critique
“That only raises the bar for being concerned with it being *only* in the hands of a “flawed democracy”, no? Besides: China *will* get this tech, and it *will* proliferate. Best help prepare for a w…” evidence ↗
Multiple readings Concern & critique · 1 Opportunity & adoption · 1 Questions & unknowns · 1
tachikoma: this regime ends in 2029, but what comes next? what if it's worse, not just ideologically but in terms of vision and competence?
tachikoma: we should try to avoid that outcome by building resilience, which i think comes from decentralization, on the web and in the physical world
Open the thread →
4 experts · 6 posts · 12h ago
Damien Williams Karl Bode Erin Kissane tweety fish
Matches Concern & critique
“it's also really just...quite hollow? Like it's not really very interesting? And that's even before it leans into the lazy "many sectors use more water than AI therefore all environmental criticism…” evidence ↗
Multiple readings Concern & critique · 1 Policy & governance · 1
Damien Williams: I'm reading the "environmentalists are wrong about data centers" piece from the NYT and ohmygoditssofuckingbad. Apparently, so long as you're punching left, they'll just let you say any goddamn thi…
Damien Williams: And, again: golf courses and flat lawns and monocropping agribusiness practices are ALSO BAD. *More Than One Thing Can Be Bad*!
Open the thread →
2 experts · 16 posts · 3h ago
Pwnallthethings SE Gyges
Matches Concern & critique
“"AI is so dangerous we must have a plan for USG to deliberately intervene to pace it" and "it is important to have open-weight models that are not controllable by anyone (e.g. to pace it)" seem to …” evidence ↗
Multiple readings Policy & governance · 1 Opportunity & adoption · 1
SE Gyges: here's who signed both the Pacing the Frontier letter released today and the Pause AI letter in 2023 demanding an end to all research for six months also, raw counts for "signed today's letter" and…
SE Gyges: fundamentally this is an anthropic thing that managed to recruit some people from other companies
Open the thread →
3 experts · 6 posts · 22h ago
Timnit Gebru Louise Amoore Baldur Bjarnason
Matches Concern & critique
“Yes, exactly this. I’m concerned that the 2008 banks “too big too fail” were bailed out with public money. As OpenAI and Anthropic embed themselves into government administration and public service…” evidence ↗
Multiple readings Concern & critique · 1 Policy & governance · 1
Timnit Gebru: Its all about "free markets" and "let the market decide" until you have competition that shows how full of shit you are and then you run to daddy US government begging to stop these "bad" Chinese m…
Timnit Gebru: The irony of OpenAI and Anthropic running to the government to save them whenever there’s competition and it’s this government that complains about “handouts” and communism.
Open the thread →
3 experts · 3 posts · 5h ago
Justin Hendrix James Grimmelmann Jed Brown
Matches Concern & critique
“BTW, this strikes me as an easy case for strict liability. It’s like a wild animal or ultrahazardous activity. You train and evaluate a frontier agentic model at your own risk, and if it escapes yo…” evidence ↗
Shared emphasis Concern & critique · 2
James Grimmelmann: BTW, this strikes me as an easy case for strict liability. It’s like a wild animal or ultrahazardous activity. You train and evaluate a frontier agentic model at your own risk, and if it escapes yo…
Justin Hendrix: BTW, this strikes me as an easy case for strict liability. It’s like a wild animal or ultrahazardous activity. You train and evaluate a frontier agentic model at your own risk, and if it escapes yo…
Open the thread →
2 experts · 6 posts · 8h ago
angela zhou J. Nathan Matias
Matches Concern & critique
“Sorry for the overly graphical and physical imagery; but I feel like we would take the harms of Slop more Seriously if we made Information more Material” evidence ↗
Multiple readings Building & implementation · 1 Context & explanation · 1
angela zhou: i feel like prototyping an ai product or workflow to Work Well and Add Value instead of becoming More Workslop is like cooking. you need to be tasting as you go, and even as models are getting bett…
angela zhou: CS culture is exhausting in insisting on an arms-length view of labor and its inputs, rather than drilling in deep to produce something that tastes good for people
Open the thread →
2 experts · 3 posts · 1d ago
jake Doll
Matches Concern & critique
“Its still rank nationalism with a coat of x-risk paint but glad they are being overt” evidence ↗
Doll: Anthropic claims to not want a ban on open weight models, is mainly concerned about China *entirely*, as well as biorisk from open weight models without guardrails. This seems exactly the same as t…
Doll: Its still rank nationalism with a coat of x-risk paint but glad they are being overt
Open the thread →