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
Active evidence filter

Showing developments with attributable Building & implementation 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.

Building & implementation
0 live reaction maps in this view

No development in this filter has enough independent reactions yet. The maps appear as soon as the evidence supports them.

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.

New AI field signal Analysis 41m ago
⚡ 6 h early
essay how to talk AI

How to talk about

11 experts across 2 network communities independently surfaced this.

Why this matches Building & implementation reaction 1 attributable expert contribution · Charles Logan
“I’m working with a librarian to develop a module for undergraduates about conducting research, and I’m once again grateful for this resource from @emilymbender.bsky.social and @nannainie.bsky.social:” evidence ↗
11 experts 2 communities 1 sources clustered
Emily M. Bender Timnit Gebru Hypervisible Luke Stark Eileen Clancy 🧿 Frank Pasquale Charles Logan Jeffrey Vagle +3

“I’m working with a librarian to develop a module for undergraduates about conducting research, and I’m once again grateful for this resource from @emilymbender.bsky.social and @nannainie.bsky.social:”

“3. Any press releases & news articles examined for anthropomorphizing language, especially language that locates agency with software rather than people & companies. See: buttondown.com/maiht3k/arch... >>”

2 experts discussed this · 16 posts
Emily M. Bender: So many conversations where otherwise serious people throw around the term "AI" without defining it, despite it being notoriously ill-defined. Thinking about why this might be >>
Emily M. Bender: In some cases, no doubt, the speaker fully believes that autonomous thinking machines are among us. (Though I said "otherwise serious people" in the prev post, so.) >>
Emily M. Bender: In other cases, I suspect that people feel like they lack the expertise to call bullshit and are repeating what they hear others say. >>
Open the full discussion →
Primary source buttondown.com ↗
More conversations

What experts are discussing without an anchoring article

5 experts · 39 posts · 21h ago
Ed Zitron SE Gyges Doll Nafnlaus 🇮🇸 🇺🇦 mackuba.eu
Matches Building & implementation
“it's actually just that my posting history for the last hour right now is a litany of horrors that depresses me and i'm trying to avoid building the pile higher” evidence ↗
Multiple readings Policy & governance · 1 Opportunity & adoption · 1
SE Gyges: trying to deduce why one would quit a good job at anthropic
SE Gyges: what her boyfriend is up to today. they really do get around
Open the thread →
5 experts · 33 posts · 23h ago
joao Grace Colin coty.bsky.social david-p-reichert.bsky.social
Matches Building & implementation
“The basic idea of the instrumental convergence thesis is that an intelligent being with, as Bostrom puts it, a "final goal", will develop certain goals that are instrumental to the final goal. The …” evidence ↗
Colin: There are indeed thousands of LessWrong posts, as well as papers and books, which make a particular theoretical case for why AI will kill us all that many people find convincing. I'm going to try t…
Colin: First of all, we must just agree off the bat that a superintelligent AI will be able to kill us all. After all it is superintelligent. As this random dude who became famous 3 days ago because of a …
Open the thread →
3 experts · 14 posts · 1d ago
SE Gyges tweety fish austin
Matches Building & implementation
“the thing is that the skills that make you a “good developer” are arguably the same irrespective of how the code is extruded it is also the case that they are very rare skills” evidence ↗
Multiple readings Building & implementation · 1 Opportunity & adoption · 1
tweety fish: parsimoniously, we will relatively soon see junior devs that are native to agentic coding with harnesses in ways that allow them to blow the doors off what senior devs can do. there'll probably be …
tweety fish: I didn't mean in terms of writing code faster, particularly. I mean in terms of knowing how to steer them and what is important and what's not to have an end result of reliable extruded software.
Open the thread →
3 experts · 3 posts · 1d ago
Melanie Mitchell Alondra Nelson Benjamin Riley
Matches Building & implementation
“Hey @ericlipton.nytimes.com 👋 AI is not “evolving” on its own. *People* are choosing how to build it. *People* are choosing how to train it. *People* are choosing how to run it. Maybe the problem i…” evidence ↗
Melanie Mitchell: Hey @ericlipton.nytimes.com 👋 AI is not “evolving” on its own. *People* are choosing how to build it. *People* are choosing how to train it. *People* are choosing how to run it. Maybe the problem i…
Alondra Nelson: Hey @ericlipton.nytimes.com 👋 AI is not “evolving” on its own. *People* are choosing how to build it. *People* are choosing how to train it. *People* are choosing how to run it. Maybe the problem i…
Open the thread →
2 experts · 5 posts · 1d ago
Ramon Astudillo Marco Z
Matches Building & implementation
“I am missing a good KV-cache-friendly workflow that goes from Specs to closed PRs. Yeah it should be Specs -> Commits/Tags <-> PR but in reality it is Specs -> C/T <-> improve Specs -> C/T <-> PR R…” evidence ↗
Ramon Astudillo: I am missing a good KV-cache-friendly workflow that goes from Specs to closed PRs. Yeah it should be Specs -> Commits/Tags <-> PR but in reality it is Specs -> C/T <-> improve Specs -> C/T <-> PR R…
Marco Z: so you either reuse the cache or use current data. Can't have both 🥶
Open the thread →
2 experts · 4 posts · 1d ago
Ryan Moulton tachikoma
Matches Building & implementation
“yeah, i think they can dive into the details as needed to build a better understanding to make decisions. i think that's where we're at now with AI tools, and the human users that steer them.” evidence ↗
Ryan Moulton: I'm not sure about this, because figuring out what makes sense to do next usually hinges on the details, often very fine details. You gotta look at examples.
tachikoma: i can see what you mean, but at the same time clearly leaders of organizations have to make big decisions about what to do next without having the full, fine-grained details. they have to delegate …
Open the thread →
2 experts · 4 posts · 1d ago
jake Singularity's Bounty e/cc
Matches Building & implementation
“I particularly love the "AI will simply deploy nanobots" takes 🙄” evidence ↗
Multiple readings Concern & critique · 1 Building & implementation · 1
jake: I don’t think there’s *no* vector of attack, but I think there’s basically no realistic vector that doesn’t have a pretty obvious escalation ladder of precursor mini-dooms first. If I were to back …
jake: Agreed that surveillance state, automated warfare, cybercrime (inc video phishing our parents retirement accts) are more near-term pressing. And that countering most steps on vectors of doom are do…
Open the thread →