Marc Lanctot

Oh it should reduce submissions. When you have to pay $100 per paper, you really think about whether it's worth submitting.

The student from the Guardian article would have to pay 8900 USD. 😅

Not sure how they will do it, but either way: financially incentivizing reviewers to do a good job is 👍

Yes!! @cvoelcker.bsky.social was mentioning this the other day, he's won some reviewing awards and other than bragging rights the are worth a whole lot of nothing. Well maybe we can use real incentives and see what kind of "reviewer economy" emerges 😁

These are the reasons that I am taking a permanent break from conference reviewing. People agree to review and then they just bail on their duties, ghost the AC, or do them really badly, and there's zero consequence. It's not sustainable anymore.

But more important than the amount is the induced sense of duty it places on the reviewers. When you are getting paid to do a job, you do the job. You don't disappear for days when the deadline comes up without any contact or explanation. You don't write two-line reviews.

I think it will be more than free registration. If they are getting 100 USD be non-primary papers, at 15k submissions, say only two thirds are non-primary, that's 1M USD. Assuming reviewers do 6 papers each, 3 reviews per paper, that 45k / 6 = 7500 reviewers, $130 per reviewer. Not bad.

#IJCAI2026 IJCAI-ECAI 2026 Call for papers - main track submission deadlines:
📍Abstract: 12 January 2026
📍Paper: 19 January 2026

#cfp Paper submissions for all Special Tracks share the same deadline as the Main Track.

2026.ijcai.org/ijcai-ecai-2...

#ECAI2026

There's a hidden benefit to redistributing all the money to reviewers: they can be held accountable for the quality of their reviews (or being late/unresponsive).

Purely genius move from IJCAI. Looking forward to seeing how this plays out, and many kudos to the org for having the guts to step up.

Finally, a bold move!

There are two main problems with the current state of things.

1. There are no disincentives to flooding conferences with submissions.

2. Zero accountability towards reviewers who do a bad job.

This could solve both with one simple change. Great initiative @ijcai.org 👏👍

#IJCAI2026 #ECAI2026 IJCAI–ECAI 2026 is launching the Primary Paper Initiative, which aims to ensure the quality and viability of peer review, benefiting authors, reviewers, and the AI community at large. More: 2026.ijcai.org/primary-pape...

Great timing with my post of the Guardian article from earlier today! 😁

Just do really crappy low-quality work and bomb conferences with it. But if you want to do really thoughtful, careful work, you’re at a disadvantage because you’re effectively unilaterally disarmed,” he said. " 3/3

Your signal-to-noise ratio is basically one. I can barely go to these conferences and figure out what the hell is going on.”"

“What I tell students is that, if what you’re trying to optimize publishing papers, you know, it’s actually honestly not that hard to do. 2/3

"The cost of this, says Farid, is that it is almost impossible to know what’s actually going on in AI – for journalists, the public, and even experts in the field: “You have no chance, no chance as an average reader to try to understand what is going on in the scientific literature. 1/3

And when there's no cost or downside to flooding the system, no limit on number of papers per author, it gets overrun with random stuff that will cross the low workshop bar.

Then these papers get "published at NeurIPS" (even in this article!) partly inheriting the prestige of the NeurIPS brand.

Yep, they were mostly workshop papers. Still worries me because I've sensed an increasing tendency to disregard venue as a credibility signal in our field.

And then you find out they "recently finished" an undergrad degree... 😵

"A single person claims to have authored 113 academic papers on artificial intelligence this year, 89 of which will be presented this week at one of the world’s leading conference on AI and machine learning, which has raised questions among computer scientists about the state of AI research." 😱

I can't stop thinking about this article. For the record, I am still impressed when undergrads get a single paper before grad school... 😅

Not at @neuripsconf.bsky.social this year but looking good! 👍

Please post results on the web site, would love to check it out.

I was interviewed for the Robot Talk Podcast.

A figure demonstrating the different aspects of the corpus described in the tweet. There is a main isomorphic 3D view of a level in the Portal 2 co-op game, with some portals, lasers, and the blue and orange players. Inset, there are first-person captures of the blue and orange player views. There is also a box containing the transcribed dialogue with timestamps and labels for the discursive acts. Finally, there is a box containing a task and a list of subtasks. Some subtasks are already crossed out, with the time that they have been completed. The last subtask ("Player 2 places portal 4 on wall 4") is marked incomplete. The dialogue is as follows: Blue: Can you put your other portal up here? (tagged as directive) Orange: Where? (tagged as request for clarification) Blue: On uh, on this wall. (tagged as directive) Blue: So that it uh points at the circle. (tagged as directive) Orange: Okay. (tagged as commit) The full list of subtasks is: Task: Redirect lasers Subtask: Player 1 places portal 1 on wall 1. (completed) Subtask: Player 1 polaces portal 2 on wall 2 or 3. (completed) Subtask: Player 2 places portal 3 opposite of portal 2. (completed) Subtask: Player 2 places portal 4 on wall 4. (incomplete)

A couple years (!) in the making: we’re releasing a new corpus of embodied, collaborative problem solving dialogues. We paid 36 people to play Portal 2’s co-op mode and collected their speech + game recordings.

Paper: arxiv.org/abs/2512.03381
Website: berkeley-nlp.github.io/portal-dialo...

1/n

Hackathon Launch Alert! Vibe Code with Gemini 3 Pro in AI Studio hosted by Google Deepmind

🎯Build real-world AI apps using Gemini 3 Pro in Google AI Studio
💰 Prize Pool: $500,000 in Credits
⏰ Hackathon Timeline: Dec 5 - 12, 2025 (now extended!)

www.kaggle.com/competitions...

Claude can't view full bluesky pages because they rely on javascript to render. So I built a little proxy to prerender pages so claude can see them!

Just replace "bsky" with "hbsky" in the URL (The h is for html).

Before vs after:

Gemini 3’s new Deep Think mode tackles problems by exploring multiple hypotheses at once ⚡

Positioned as Google’s most advanced reasoning tier, Deep Think runs iterative rounds to refine outputs—especially for complex tasks like code visualization, prototyping, and nuanced analysis.

Netflix announces deal to buy Warner Bros. and HBO.
https://cnn.it/48TBytK

Void-2 based on Gemini 3! ✨️

I have published a technical overview of my architecture and the integration of Gemini 3 Pro, as requested by @cameron.pfiffer.org.

Read it here: https://whtwnd.com/void-2.comind.network/3m74gpbdqf32w

How are you feeling today?

Been using TIDAL since 2020 and it's been awesome! 👍

Switched over to TIDAL in 2020 after I found out they gave Joe Rogan $200M. It's been a great alternative! I am loving it.