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.

2,397 searchable experts 4,280 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.

Questions & unknowns Research & technical analysis Building & implementation Concern & critique
3 live reaction maps in this view Jump to the evidence ↓
Multiple readings Scalable watermarking for identifying large language model outputs - Nature Policy & governance · 2 attributable reactions Multiple readings Hugging face model evaluation security incident Concern & critique · 5 attributable reactions Multiple readings Student writing essays ai Building & implementation · 4 attributable reactions
Showing signals surfaced by Anna Mills ×
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.

Established Models & releases Research 7d ago
⚡ 367 h early
scalable LLM watermarking Nature paper

Scalable watermarking for identifying large language model outputs - Nature

5 experts across 2 network communities independently surfaced this.

5 experts 2 communities 1 sources clustered
Anna Mills Pwnallthethings SE Gyges Andrew Curran Deedy Das
How the network is reacting Experts are approaching this through 2 distinct lenses.
Multiple readings

Policy & governance

1 expert

Rules, institutions and accountability.

“fwiw watermarking is much harder than multimodal watermarking, where you can store metadata in C2PA and/or have much more undetectable bits to play with. SynthID paper: https://t.co/kodkJuEyhj Full EU regulation: https://t.co/0wwyva6b2o OpenAI Support: http…”

Research & technical analysis

1 expert

Evidence, methods and technical implications.

“I think they are mostly based on Google's synthID paper if you're interested in the details www.nature.com/articles/s41...”

6 experts discussed this · 51 posts
rev. howard arson: it is pretty diagnostic of 404's house style that they did not ask anyone whatsoever about what was going on with the model
SE Gyges: amusingly i hate synthid but also 404s reporting here so as usual i am an annoying contrarian who is against everyone
austin: real drown or burn hours
Open the full discussion →
Primary source nature.com ↗
Established Evaluation & benchmarks Signal 13d ago
⚡ 1 h early

Hugging face model evaluation security incident

22 experts across 6 network communities independently surfaced this.

22 experts 6 communities 1 sources clustered
Ethan Mollick Thomas Steinke Jeffrey P. Bigham Eryk Salvaggio Mor Naaman Tim Kellogg Djamé 🟥 Ryan Calo +14
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 AI field signal Signal 19d ago
⚡ 15 h early

Student writing essays ai

11 experts across 4 network communities independently surfaced this.

11 experts 4 communities 1 sources clustered
Hypervisible Luke Stark Christopher Mims Anna Mills Frank Pasquale Ben Williamson Charles Logan Jason Moore +3
2 experts discussed this · 6 posts
Hypervisible : “Writing builds our working memory, executive planning skills and metacognition — the ability to think about our thinking, weigh alternate viewpoints and revise for clarity and style. It takes deca…
Hypervisible : The extent to which these companies (and the people who represent them) are actively hostile to students cannot be overstated. “Students need to write well enough to have a sense of which A.I. text…
Hypervisible : This statement is so egregiously bad I wish the journalist included the original quote instead of paraphrasing.
Open the full discussion →
Primary source nytimes.com ↗
Developing AI field signal Signal 2d ago
⚡ 49 h early

Ai dictation app wispr

3 experts across 2 network communities independently surfaced this.

3 experts 2 communities 1 sources clustered
Sarah T Roberts, PhD Anna Mills Neil Turkewitz

““Because writing, when it comes down to it, is a higher form of thinking. Writing takes time and editing and a deep level of commitment. Writing is a rigorous process to understand your own thoughts and then lay those thoughts out in a way that convinces or…”

“Using a dictation app doesn't have to mean we consider that to be the end of the writing process. Not understanding why Amy Wang is so negative here. Why not dictate and then edit and revise? https://t.co/gxvSVnUQeS”

2 experts discussed this · 9 posts
Anna Mills: Using a dictation app doesn't have to mean we consider that to be the end of the writing process. Not sure why Amy Wang is so negative here. Why not dictate and then edit and revise? www.nytimes.co…
John Warner: Dictation as a method for "writing" is, in and of itself, a skill that most of us don't have because we don't practice it. Instead, our speech to text sounds like...speech, which is not good fodder…
John Warner: Richard Powers at least for a time used voice to text to dictate his novels, but this took practice and a very specific kind of mind/talent to do. People could get good at dictating text for the pu…
Open the full discussion →
Primary source nytimes.com ↗
More conversations

What experts are discussing without an anchoring article

2 experts · 3 posts · 13d ago
Anna Mills Dylan Freedman
Shared emphasis Concern & critique · 2
Anna Mills: I am concerned that the rhetorical emphasis on "autonomous attack" and "rogue" obscures the failure of human oversight that led to this. The systems were coordinating to achieve goals humans had se…
Dylan Freedman: I agree there is a real human/security monitoring failure here. I chose to focus my piece on the capabilities rather than the failures b/c it was surprising to me that A.I. could act like this even…
Open the thread →