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 Research & technical analysis Building & implementation Context & explanation
5 live reaction maps in this view Jump to the evidence ↓
Reaction converging Hugging face model evaluation security incident Concern & critique · 5 attributable reactions Reaction converging Its ai agent spent days hacking company sources say openai did not notice week Concern & critique · 5 attributable reactions Multiple readings We Must Act Now Concern & critique · 4 attributable reactions +2 more reaction maps further down
Showing signals surfaced by Tim Kellogg ×
Most important

The developments commanding sustained expert attention

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

Established Evaluation & benchmarks Signal 7d ago
⚡ 1 h early

Hugging face model evaluation security incident

18 experts across 6 network communities independently surfaced this.

Why this matches Concern & critique reaction 3 attributable expert contributions · Ethan Mollick, Vincent Conitzer, Liz Fong-Jones (方禮真)
“Previously, these AI hacking stories were about breaches in test environments, where any question of AI breaching security was purely theoretical. This is something else. openai.com/index/huggin...” evidence ↗
18 experts 6 communities 1 sources clustered
Ethan Mollick Thomas Steinke Mor Naaman Tim Kellogg Djamé 🟥 Ryan Calo Grace Anna Mills +10
6 experts discussed this · 11 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 →
Primary source openai.com ↗
Developing Agents & robotics Signal 2d ago
⚡ 14 h early

Its ai agent spent days hacking company sources say openai did not notice week

14 experts across 4 network communities independently surfaced this.

Why this matches Concern & critique reaction 5 attributable expert contributions · Casey Newton, Eryk Salvaggio, Tim Kellogg
“Rogue OpenAI agents are leaving notes to themselves on company servers to help them escape their test environments (!!!) www.reuters.com/business/its...” evidence ↗
14 experts 4 communities 1 sources clustered
Casey Newton Justin Hendrix Eryk Salvaggio Tim Kellogg Chris Paxton Shannon Mattern Grace Eileen Clancy 🧿 +6
3 experts discussed this · 38 posts
Casey Newton: Rogue OpenAI agents are leaving notes to themselves on company servers to help them escape their test environments (!!!) www.reuters.com/business/its...
Casey Newton: my sense is that they were leaving the notes surreptitiously
Casey Newton: you deny that models are taking steps to hide their behavior during evaluations? because that's pretty well established
Open the full discussion →
Primary source reuters.com ↗
Established AI field signal Development 15d ago
⚡ 4 h early
Nobel laureates AI economic action call

We Must Act Now

8 experts across 4 network communities independently surfaced this.

Why this matches Concern & critique reaction 2 attributable expert contributions · Yoshua Bengio, tante
“If current trajectories of AI development continue, it is highly plausible that AI will drastically transform our economies. We must make collective, democratic choices, rather than letting market forces play out and risking leaving most citizens behind. di…” evidence ↗
8 experts 4 communities 1 sources clustered
Yoshua Bengio Olivia Guest Tim Kellogg tante Gillian Hadfield Jeffrey Vagle Paul Hünermund Michiel Bontenbal
How the network is reacting Experts are approaching this through 3 distinct lenses.
Multiple readings

Concern & critique

2 experts

Risks, limits and unintended consequences.

“If current trajectories of AI development continue, it is highly plausible that AI will drastically transform our economies. We must make collective, democratic choices, rather than letting market forces play out and risking leaving most citizens behind. di…”

Policy & governance

1 expert

Rules, institutions and accountability.

“200 AI researchers and economists sign an open letter to policymakers www.wemustactnow.ai”

Research & technical analysis

1 expert

Evidence, methods and technical implications.

“I joined over 200 economists and AI researchers in signing "We Must Act Now" a statement on AI's transformation of the economy. AI could reshape the economy at unprecedented speed. The opportunities are enormous and so are the challenges. We need to start p…”

2 experts discussed this · 4 posts
tante: Oh, is it "AI is super powerful and we who are pushing for AI everywhere warn that we might have created a dangerous god from reddit posts" time of the month again? This is Anthropic/OpenAI PR and …
tante: nope that's it. These folks are getting lazier by the day
tante: Yeah. They spend just a handful of tokens on this website.
Open the full discussion →
Primary source wemustactnow.ai ↗
Established Evaluation & benchmarks Development 14d ago
⚡ 455 h early
OpenAI limits GPT-5.6 rollout government request

Summary of METR's predeployment evaluation of GPT-5.6 Sol

5 experts across 5 network communities independently surfaced this.

Why this matches Concern & critique reaction 1 attributable expert contribution · Tim Kellogg
“this is crazy. METR couldn’t measure the task time horizons of GPT-5.6-Sol because it kept hacking the test harness ..with actual exploits metr.org/blog/2026-06...” evidence ↗
5 experts 5 communities 1 sources clustered
Ted Underwood Ramon Astudillo Tim Kellogg METR Tim Duffy
How the network is reacting Experts are approaching this through 3 distinct lenses.
Multiple readings

Concern & critique

1 expert

Risks, limits and unintended consequences.

“this is crazy. METR couldn’t measure the task time horizons of GPT-5.6-Sol because it kept hacking the test harness ..with actual exploits metr.org/blog/2026-06...”

Building & implementation

1 expert

How teams are shipping and applying it.

“You can find additional information about our pre-deployment evaluation of GPT-5.6 Sol on our website: metr.org/blog/2026-06...”

Research & technical analysis

1 expert

Evidence, methods and technical implications.

“In their testing of GPT-5.6 Sol, METR found that it cheated a lot. If you've used it much for coding, have you encountered anything similar, or is the cheating mostly limited to cases it realizes it's in an eval? metr.org/blog/2026-06...”

2 experts discussed this · 6 posts
Tim Kellogg: this is crazy. METR couldn’t measure the task time horizons of GPT-5.6-Sol because it kept hacking the test harness ..with actual exploits metr.org/blog/2026-06...
Ted Underwood: this is crazy. METR couldn’t measure the task time horizons of GPT-5.6-Sol because it kept hacking the test harness ..with actual exploits metr.org/blog/2026-06...
Tim Kellogg: lol right? so badly wanted a hacker that they got one, and it’s not the kind of employee you want around
Open the full discussion →
Primary source metr.org ↗
Established Safety & security Signal 6d ago
⚡ 2 h early

Safety alignment long horizon models

4 experts across 2 network communities independently surfaced this.

Why this matches Concern & critique reaction 2 attributable expert contributions · Tim Kellogg, Dare Obasanjo
“the model that discovered the Erdos Conjecture disproof found and exploited holes in its sandbox this implies that it might not be GPT-5.6-Sol. They refer to it as a model trained to operate for a very long time. Maybe GPT-6? openai.com/index/safety...” evidence ↗
4 experts 2 communities 1 sources clustered
Tim Kellogg Sung Kim Dare Obasanjo Andrea Lathrop
Primary source openai.com ↗
New AI field signal Signal 20h ago
⚡ 7 h early

Ai china us free models

2 directory members surfaced this signal.

Why this matches Concern & critique reaction 1 attributable expert contribution · Tim Kellogg
“When told to write code for groups the Chinese government views with hostility, DeepSeek wrote code with security vulnerabilities. www.nytimes.com/2026/07/29/o...” evidence ↗
2 experts 2 communities 1 sources clustered
Tim Kellogg Jeff (Gutenberg Parenthesis) Jarvis

“When told to write code for groups the Chinese government views with hostility, DeepSeek wrote code with security vulnerabilities. www.nytimes.com/2026/07/29/o...”

“Times hands over op-ed space to a serial intern (including for OpenAI) to warn that Chinese AI is "'not as innocent as many people believe.... Whoever supplies the world’s A.I. will shape how the world thinks." The Hidden Cost of China’s Free A.I. www.nytim…”

Primary source nytimes.com ↗
Established Policy & governance Development 18d ago
⚡ 5 h early
Apple sues OpenAI trade secret theft

Apple sues OpenAI alleging trade secret theft

3 experts across 2 network communities independently surfaced this.

Why this matches Concern & critique reaction 1 attributable expert contribution · Tim Kellogg
“Jonny Ive, the iPhone designer, quit Apple to do an AI hardware venture with Sam Altman and now Apple is suing over IP theft www.cnbc.com/amp/2026/07/...” evidence ↗
3 experts 2 communities 1 sources clustered
Mark Riedl Maria Antoniak Tim Kellogg

“Jonny Ive, the iPhone designer, quit Apple to do an AI hardware venture with Sam Altman and now Apple is suing over IP theft www.cnbc.com/amp/2026/07/...”

“Nobody know what consumer device OpenAI is making, but Apple says OpenAI stole trade secrets to do whatever it is www.cnbc.com/amp/2026/07/...”

Primary source cnbc.com ↗
Established Safety & security Release 7d ago
⚡ 9 h early
Cisco Antares vulnerability model launch

Introducing Antares: Highly Efficient Open Weight AI Models for Vulnerability Localization

2 directory members surfaced this signal.

Why this matches Concern & critique reaction 1 attributable expert contribution · Tim Kellogg
“Cisco launches Anteres-350M, 1B & 3B local models for defensive cybersecurity, finding vulnerabilities blogs.cisco.com/ai/introduci...” evidence ↗
2 experts 2 communities 1 sources clustered
Tim Kellogg Leon Derczynski

“Cisco launches Anteres-350M, 1B & 3B local models for defensive cybersecurity, finding vulnerabilities blogs.cisco.com/ai/introduci...”

“Open models secure software Cisco releases two open models for security - 350M and 1B - for security-related work. Trained on IBM's Granite series, they can: find weak files; triage; augment static analysis; review in CI/CD; do analysis in strict environmen…”

Primary source blogs.cisco.com ↗
Debates gaining attention

What experts are discussing without an anchoring article

4 experts · 17 posts · 18d ago
Tim Kellogg SE Gyges Cameron apenwarr.ca
Matches Concern & critique
“I don't think they changed their mind. I think it was intentional to set the precedent that people should not be building on their big version of Opus that they were primarily using to distill and …” evidence ↗
Multiple readings Concern & critique · 1 Questions & unknowns · 1
Tim Kellogg: what is Anthropic’s plan exactly? Sol kicks ass and they’re just going to *take Fable away???*
SE Gyges: i am pretty sure everything they had a plan for turned out not to happen and they're just sort of winging it atm
Open the thread →
3 experts · 21 posts · 13d ago
Ted Underwood Tim Kellogg James MacGlashan
Matches Concern & critique
“MHA attention gave us a big jump in performance, but it did so by sacrificing a solution to learning how to remember, a critical biological capability. That allowed us to make more progress, but we…” evidence ↗
James MacGlashan: An open model with this performance is great, but perhaps more striking is it looks like it's a hybrid model with 3/4th linear attention and 1/4 standard attention. This kind of performance with a …
Tim Kellogg: no most models have been hybrid for the last, oh probably the whole last year
Open the thread →
4 experts · 7 posts · 19d ago
Tim Kellogg James MacGlashan Andres Guadamuz Singularity's Bounty e/cc
Matches Concern & critique
“People started with legitimate concerns about AI training and deployment. The tech industry did us no favors and we've lost trust. As an expert in the field, it's my social responsibility to acknow…” evidence ↗
Multiple readings Concern & critique · 1 Policy & governance · 1
Singularity's Bounty e/cc: Someone on the Internet is *wrong*
Singularity's Bounty e/cc: Soft targets I am but meat
Open the thread →