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
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
4 live reaction maps in this view Jump to the evidence ↓
Reaction converging Its ai agent spent days hacking company sources say openai did not notice week Concern & critique · 5 attributable reactions Reaction converging Hugging face model evaluation security incident Concern & critique · 5 attributable reactions Multiple readings A global workspace in language models \ Anthropic Markets & investment · 3 attributable reactions +1 more reaction maps further down
Showing signals surfaced by Grace ×
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.

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.

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 Evaluation & benchmarks Signal 7d ago
⚡ 1 h early

Hugging face model evaluation security incident

18 experts across 6 network communities independently surfaced this.

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 ↗
Established Models & releases Research 10d ago
⚡ 3 h early
global workspace theory language models paper

A global workspace in language models \ Anthropic

6 experts across 3 network communities independently surfaced this.

6 experts 3 communities 1 sources clustered
Ethan Mollick Grace Tim Duffy Amy Hoy Michiel Bontenbal Stephan Geering
How the network is reacting Experts are approaching this through 3 distinct lenses.
Multiple readings

Markets & investment

1 expert

Capital, companies and commercial impact.

“Anthropic has a new paper out, alleging a global workspace in LLMs. The term comes from Global Workspace Theory, a leading theory of consciousness. The method they use to investigate this is a refinement of logit lens, which they call J-lens. www.anthropic.…”

Building & implementation

1 expert

How teams are shipping and applying it.

“Some really interesting research from Anthropic that AI models have spontaneously developed a workspace that "appears to support the functions associated with conscious access" Demo of how this works: www.neuronpedia.org/qwen3.6-27b/... Research: www.anthro…”

Research & technical analysis

1 expert

Evidence, methods and technical implications.

“On the second point see here: www.anthropic.com/research/glo...”

2 experts discussed this · 4 posts
Tim Duffy: Anthropic has a new paper out, alleging a global workspace in LLMs. The term comes from Global Workspace Theory, a leading theory of consciousness. The method they use to investigate this is a refi…
Tim Duffy: More detail on the method from the paper:
Grace: Anthropic has a new paper out, alleging a global workspace in LLMs. The term comes from Global Workspace Theory, a leading theory of consciousness. The method they use to investigate this is a refi…
Open the full discussion →
Primary source anthropic.com ↗
Developing Models & releases Research 2d ago
⚡ 17 h early
LLM story diversity diagnosis paper

Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories

4 experts across 2 network communities independently surfaced this.

4 experts 2 communities 1 sources clustered
Ted Underwood David Mimno Grace Sil Hamilton
How the network is reacting 4 experts are independently emphasizing research & technical analysis.
Reaction converging

Research & technical analysis

4 experts

Evidence, methods and technical implications.

“In the "Limitations" section they acknowledge that their study was monolingual, but say a multilingual generalization would be interesting. arxiv.org/abs/2605.26492”

2 experts discussed this · 2 posts
Ted Underwood: was reading this and enjoying it — and then enjoyed it more when it turned out to be a paper by Sil Hamilton and @dmimno.bsky.social
David Mimno: was reading this and enjoying it — and then enjoyed it more when it turned out to be a paper by Sil Hamilton and @dmimno.bsky.social
David Mimno: The preprint is arxiv.org/abs/2605.26492, and @srhm.ca is working on a post!
Open the full discussion →
Primary source arxiv.org ↗
Established AI field signal Signal 10d ago

arxiv.org

2 experts are actively discussing the implications.

2 experts 1 community 1 sources clustered
Grace Luis Villa

“Bluesky loves to Trust the Experts until it comes to this particular issue. Here's Prof. Eric Schwitzgebel on the subject: https://arxiv.org/pdf/2510.09858”

2 experts discussed this · 6 posts
Grace: Bluesky loves to Trust the Experts until it comes to this particular issue. Here's Prof. Eric Schwitzgebel on the subject: https://arxiv.org/pdf/2510.09858
Grace: The article overall is pretty measured and reasonable. I do have qualms with the "character-level" interventions mentioned here, as the relationship to subjective experience is unclear, and might d…
Grace: Thank you!
Open the full discussion →
Primary source arxiv.org ↗
More conversations

What experts are discussing without an anchoring article

5 experts · 23 posts · 4d ago
Grace Dustin Moskovitz SE Gyges Phillip Carter Doll
Multiple readings Policy & governance · 2 Questions & unknowns · 2
Dustin Moskovitz: Apparently the government is against moving the AI frontier until jailbreaks are solved. For the uninitiated, jailbreaks have existed since the beginning and as far as we know so far, cannot be sol…
Grace: I’m so curious what the actual jailbreak is
Open the thread →
2 experts · 6 posts · 8d ago
Grace futur.blue
Multiple readings Concern & critique · 1 Research & technical analysis · 1
Grace: Another funny aspect of this is that the model was smart enough to take all of these actions, but not smart enough to think "OpenAI would probably prefer that I fail this benchmark than hack the sy…
Grace: Weak orthogonality ++ imo
Open the thread →
2 experts · 5 posts · 8d ago
Grace Colin
Colin: I guess there’s going to be a gold rush on siccing Fable & co on conjectures that are widely believed to be true
Colin: If a language model proves the Collatz Conjecture I will become a rationalist-style extremist doomer
Open the thread →
2 experts · 3 posts · 16d ago
Grace brennan
brennan: ironically, my flow now is the least complex it's ever been - back in the day you needed these huge sprawling comfyUI setups to do things and now I mostly do two: - de novo generation in midjourney…
brennan: and my midjourney flow is: have a few moodboards filled with things I like, and then a bunch of style reference images on top of that (it seems to prioritize them slightly?) which don't even have t…
Open the thread →