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 Research & technical analysis 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.

Research & technical analysis
1 live reaction map in this view Jump to the evidence ↓
Reaction converging Navier-Stokes lost in translation: Why Lean verification of AI autoformalisation does not guarantee correct natural language proofs Research & technical analysis · 3 attributable reactions
Showing signals surfaced by Vikram Saraph ×
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 NLP & language Research 15d ago
⚡ 14 h early
Lean AI autoformalisation verification limits paper

Navier-Stokes lost in translation: Why Lean verification of AI autoformalisation does not guarantee correct natural language proofs

5 experts across 3 network communities independently surfaced this.

Why this matches Research & technical analysis reaction 3 attributable expert contributions · Maxim Raginsky, steven t. piantadosi, Colin
“Looks like others have made the same point as well: arxiv.org/abs/2610.08144” evidence ↗
5 experts 3 communities 1 sources clustered
Maxim Raginsky steven t. piantadosi Colin Matthew Guzdial Vikram Saraph
How the network is reacting 3 experts are independently emphasizing research & technical analysis.
Reaction converging

Research & technical analysis

3 experts

Evidence, methods and technical implications.

“But mathematicians reading these proofs seem to often have a hard time, meaning that they're not great at aligning the formal stuff with our intuitions (yet). I'd also take this as a little evidence arxiv.org/abs/2610.08144”

Primary source arxiv.org ↗
Debates gaining attention

What experts are discussing without an anchoring article

5 experts · 12 posts · 20d ago
Mark Riedl SE Gyges critical slop studies rev. howard arson Vikram Saraph
Matches Research & technical analysis
“an "independent" ai researcher closely tied to MIRI and lightcone infrastructure” evidence ↗
Multiple readings Markets & investment · 1 Questions & unknowns · 1
SE Gyges: the "bad misstatement" in question is to the effect of "yes, I commit sexual harassment regularly and utilize the funds made available to me via nonprofit grants to commit sexual harassment and to …
SE Gyges: i don't wanna be bitter or whatever but the fact that people on bluesky are a meaningful force here is insane and should tell you that this entire area is deeply fucking ill. i don't know or like t…
Open the thread →
2 experts · 26 posts · 18d ago
Pwnallthethings Vikram Saraph
Matches Research & technical analysis
“But complexity only tells you how quickly a problem gets harder in theory, not how well it runs on real hardware. An algorithm that is less complex might be a major improvement on paper, but have v…” evidence ↗
Shared emphasis Opportunity & adoption · 2
Pwnallthethings: OK so this is a pretty diverse set of proofs and unfortunately probably most discourse will be dominated by the scale and method of the release, and the more foundational Qs of AI in math itself. B…
Pwnallthethings: Top one is probably a *faster* fourier transform. FFTs are a class of algorithms widely and diversely used in signal processing and digital video/audio compression. FFTs have had complexity ~O(n lo…
Open the thread →