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
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.

Showing signals surfaced by David Picard ×
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 Evaluation & benchmarks Research 17d ago
⚡ 10 h early
hyperparameter tuning on test set paper

On Hyperparameter Tuning on the Test Set

3 experts are actively discussing the implications.

Why this matches Research & technical analysis reaction 1 attributable expert contribution · jvgemert.bsky.social
“In addition to the research findings, this is the type of research I really enjoyed doing 😊 "On Hyperparameter Tuning on the Test Set" arxiv.org/abs/2610.05902” evidence ↗
2 experts 1 community 1 sources clustered
David Picard jvgemert.bsky.social

“In addition to the research findings, this is the type of research I really enjoyed doing 😊 "On Hyperparameter Tuning on the Test Set" arxiv.org/abs/2610.05902”

3 experts discussed this · 6 posts
jvgemert.bsky.social: In addition to the research findings, this is the type of research I really enjoyed doing 😊 "On Hyperparameter Tuning on the Test Set" arxiv.org/abs/2610.05902
David Picard: In addition to the research findings, this is the type of research I really enjoyed doing 😊 "On Hyperparameter Tuning on the Test Set" arxiv.org/abs/2610.05902
David Picard: Joining the heresy, I see... But in all seriousness, I think the observation you make that test-tuning is just a uniform boost (fig 3) is an important lesson of deep learning (interpolation thresho…
Open the full discussion →
Primary source arxiv.org ↗
More conversations

What experts are discussing without an anchoring article

3 experts · 5 posts · 3d ago
Clément Canonne David Picard A. Feder Cooper
Matches Research & technical analysis
“One solution: normalize desk-rejections. After a fast first assessment, papers are either desk-rejected, or go to full review. The current conf peer-review system has desk rejects as the exception,…” evidence ↗
Shared emphasis Research & technical analysis · 2
Clément Canonne: One solution: normalize desk-rejections. After a fast first assessment, papers are either desk-rejected, or go to full review. The current conf peer-review system has desk rejects as the exception,…
Clément Canonne: It does sting to get a desk rejection after one week. What else does? Waiting 4+ months for the same rejection, with a 2-line review (or, worse, an entirely AI-generated one). If you're not going t…
Open the thread →
3 experts · 5 posts · 17d ago
Marc Lanctot Christian Wolf David Picard
Matches Research & technical analysis
“6 papers to review for ICLR. Fair game. It's a lot, but I understand I have to contribute, sure. But 16 days to do it is just bad. I won't be able to deliver the best review I can do in such a shor…” evidence ↗
Reaction converging Research & technical analysis · 3
David Picard: 6 papers to review for ICLR. Fair game. It's a lot, but I understand I have to contribute, sure. But 16 days to do it is just bad. I won't be able to deliver the best review I can do in such a shor…
Christian Wolf: I have only 3 but now I receive invites from ACs to receive additional papers.... ?!?
Open the thread →
2 experts · 9 posts · 20d ago
Pierre Alquier David Picard
Matches Research & technical analysis
“Ça fait quelque temps (2/3 ans) que je fais de la modération pour arXiv sur la catégorie stat.ML. L'été a été tranquille comme d'hab mais cette rentrée est y'a effectivement trois fois plus de taff…” evidence ↗
Multiple readings Policy & governance · 1 Research & technical analysis · 1
David Picard: Il faut regarder par catégorie. Le gros de l'exp vient de CS.LG, CS.CV et CS.CL c'est à dire l'IA moderne.
David Picard: Ce sont les deadlines des conférences de ML
Open the thread →
3 experts · 4 posts · 17d ago
David Picard David Marx Eric Brachmann
Matches Research & technical analysis
“It appears that coding agents have been trained on too many SOTA chasing papers, because they latch onto tiny, meaningless improvements in metrics. That's what you get...” evidence ↗
Multiple readings Research & technical analysis · 1 Questions & unknowns · 1
Eric Brachmann: It appears that coding agents have been trained on too many SOTA chasing papers, because they latch onto tiny, meaningless improvements in metrics. That's what you get...
David Marx: maybe try adding a prompt indicating a research phase, like "We are early in design exploration. There is a lot of uncaptured value to be leveraged. Don't get caught up on small incremental improve…
Open the thread →
2 experts · 3 posts · 11d ago
David Picard Michiel Bontenbal
Matches Research & technical analysis
“She might be right but she is definitely AI. #irony hints: pl*net, handwriting too perfect, sign not behaving like paper, car in background is blurring etc.” evidence ↗
Michiel Bontenbal: She might be right but she is definitely AI. #irony hints: pl*net, handwriting too perfect, sign not behaving like paper, car in background is blurring etc.
David Picard: The house wall and roof on the left has this Escherian geometry that is really hard to capture on a real camera 🙃
Open the thread →