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
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.

Showing signals surfaced by Luis Villa ×

The developments commanding sustained expert attention

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

Established Models & releases Release 14d ago
⚡ 4 h early
Anthropic open source vulnerability service

An opt-in vulnerability-finding service for open-source software

6 experts across 3 network communities independently surfaced this.

6 experts 3 communities 1 sources clustered
How the network is reacting Experts are approaching this through 2 distinct lenses.
Multiple readings

Research & technical analysis

2 experts

Evidence, methods and technical implications.

“I understand that Anthropic thinks they're doing open source projects a favor here, but what they're actually doing is creating an enormous amount of work for people that they are not paying. www.anthropic.com/research/lau...”

Concern & critique

1 expert

Risks, limits and unintended consequences.

“Claude for critical infrastructure Anthropic has a new program where open source projects can apply to use a new service for code review to spot vulnerabilities Users must prove that they’re actually maintainers, and projects must be critical infrastructure…”

2 experts discussed this · 11 posts
evacide: I understand that Anthropic thinks they're doing open source projects a favor here, but what they're actually doing is creating an enormous amount of work for people that they are not paying. www.a…
evacide: What do you do if you are the core maintainer on a project and your options are "more vuln reports than you can ever hope to triage" or "someone finds a critical 0day in your project and either exp…
evacide: I assume that the reason requests are considered on a case-by-case basis is that they think they will be able to catch all of these.
Open the full discussion →
Developing AI field signal Analysis 1d ago
AI bot midterm fact-checking lessons

I Built an AI Bot to Fact-Check the Midterms. Here's What I've Learned.

3 experts across 2 network communities independently surfaced this.

3 experts 2 communities 1 sources clustered

“I built an AI Community Notewriter. And here's why I think X's Community Notes is going to fail when election rumors about "rigged machines" and "illegal voters" start flying. agentsofinfluence.substack.com/p/i-built-an...”

“Really interesting hands-on digging from @noupside.bsky.social on some ways Notes has gone sideways on X: agentsofinfluence.substack.com/p/i-built-an...”

Established AI field signal Development 13d ago
AI developer open source Adobe clone

Recreating Adobe: AI and the limits of software copyright

5 experts across 2 network communities independently surfaced this.

5 experts 2 communities 1 sources clustered
How the network is reacting 2 experts are emphasizing policy & governance.
Shared emphasis

Policy & governance

2 experts

Rules, institutions and accountability.

“Recreating Adobe: AI and the limits of software copyright www.technollama.co.uk/recreating-ado… #AI #software #copyright”

Established Evaluation & benchmarks Signal 3d ago
⚡ 5 h early

GitHub - pardissz/fairstress: Code for "Asymmetrically Unfair Decision Making and Reasoning in LLMs" (FairStress benchmark, scoring, debiasing methods, figures)

2 directory members surfaced this signal.

2 experts 1 community 1 sources clustered

“Today we’re introducing FairStress, a benchmark for testing demographic fairness beyond a model’s final answer: how readily it changes a well-supported decision & how it explains its choice. 🧵 buff.ly/lWsW77r”

Established Evaluation & benchmarks Signal 3d ago

GitHub - pardissz/fairstress: Code for "Asymmetrically Unfair Decision Making and Reasoning in LLMs" (FairStress benchmark, scoring, debiasing methods, figures)

2 directory members surfaced this signal.

2 experts 1 community 1 sources clustered

“Today we’re introducing FairStress, a benchmark for testing demographic fairness beyond a model’s final answer: how readily it changes a well-supported decision & how it explains its choice. 🧵 buff.ly/lWsW77r”

Established Policy & governance Research 16d ago
Catala law programming language paper

Catala: A Programming Language for the Law

7 experts are actively discussing the implications.

1 expert 1 community 1 sources clustered

“Yeah overall the vast majority of law won’t be usefully provable in this way, only specific things (like tax law) that have highly structured logics. So I bet a lot of admin law? Anyway here is one of the key papers: arxiv.org/abs/2103.03198”

7 experts discussed this · 21 posts
D. Hicks: ffs
Grace: They should’ve compressed it into a 3 minute slop music video obviously
Luis Villa: gonna solve constitutional law this weekend
Open the full discussion →
Established Safety & security Resource 19d ago
AI safety whistleblower legal fund

AI Safety Whistleblower Legal Defense Fund - LASST

2 experts are actively discussing the implications.

1 expert 1 community 1 sources clustered

“eg this (which I just learned about and is great!) should have a multi-million dollar rainy day fund and a large PR budget on top of a legal budget. lasst.org/ai-safety-wh...”

2 experts discussed this · 4 posts
Luis Villa: Would love for some folks who’ve benefited materially from shitty companies to fund serious, organized, ongoing whistleblowing protection efforts rather than just tooting their own horns.
James Grimmelmann: Sorry to bother @riana.bsky.social about this twice in one day, but at least LASST does have a whistleblower legal defense fund.
Luis Villa: eg this (which I just learned about and is great!) should have a multi-million dollar rainy day fund and a large PR budget on top of a legal budget. lasst.org/ai-safety-wh...
Open the full discussion →

What experts are discussing without an anchoring article

4 experts · 27 posts · 15d ago
Multiple readings Opportunity & adoption · 2 Markets & investment · 1
Luis Villa: I would like this to be right, but I don't think it is. Once upon a time I looked semi-seriously on what it would take to do self-hosting on an Android base: apps fully sandboxed; installable from …
Luis Villa: Really deserves a whole essay, but: capitalist (really, Venture Capitalist) control of software creates enshittification in a very specific, predictable dynamic; third-party control of our data has…
Open the thread →
5 experts · 10 posts · 18d ago
Multiple readings Questions & unknowns · 2 Building & implementation · 1
Sung Kim: Look at what Meta did. They acquired Manus because Manus understood agents. Meta later had to unwind it, but the underlying point still stands: product instinct matters.
Sung Kim: OpenAI’s Dev Day 2026 was an unmitigated disaster. Now they’re trying to recover over the next 28 days. We’ll see. If it were up to me, I’d fire the cool, hip, highly educated people and hire more …
Open the thread →
4 experts · 14 posts · 11d ago
Multiple readings Building & implementation · 1 Research & technical analysis · 1 Context & explanation · 1
Amy Hoy: summarizing text is actually an incredibly difficult task (for people) and i’m continuously surprised that it’s positioned as being trivial
Amy Hoy: i assumed so. but yeah diff models will work better for diff stuff as jake says, but imo none of them are great at it except the frontier models
Open the thread →