Marc Lanctot reposted
As Meta employees brace for layoffs next Wednesday, May 20, many say the vibes are horrifically, historically low.
Who's Who of AI
Marc Lanctot
Research Scientist at Google DeepMind, multiagent RL
Why they matter
Research Scientist at Google DeepMind, multiagent RL with public evidence across Agents & robotics, AI research, Models & releases.
- AI signals
- 110 past 30d
- Sources
- 8 distinct domains
- Discussions
- 19 past 30d
- Latest signal
- 4d ago
Research Scientist at Google DeepMind, interested in multiagent reinforcement learning, game theory, games, and search/planning.
Lover of Linux 🐧, coffee ☕, and retro gaming. Big fan of open-source. #gohabsgo 🇨🇦
For more info: https://linktr.ee/sharky6000
What they're sharing
Meta’s New Reality: Record High Profits. Record Low Morale | WIRED wired.com
View on Bluesky →
Meta CTO Andrew Bosworth Admits the Company’s AI Reorg Was ‘Atrocious’ wired.com
View on Bluesky →
Scalable decision-making for games of imperfect information - Nature nature.com
View on Bluesky →
Declaration — Math and AI mathandai.org
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Scalable decision-making for games of imperfect information - Nature nature.com
Articles & links
Marc Lanctot reposted
In an internal memo seen by WIRED, Bosworth promised employees more stability, better communication, and the return of workplace perks as the company seeks to improve morale. www.wired.com/story/andrew...
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
25 Fields Medalists.. 😱 mathandai.org
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos beat Pim Niemeijer, the most decorated Stratego player, by 15 wins, 1 loss and 4 draws over 20 games.
- Training cost was under $8,000, versus an estimated $3–4.5 million for DeepMind's prior Stratego system DeepNash.
- The same self-play plus test-time search recipe also produced a superhuman Barrage Stratego AI and state-of-the-art bots for Hanabi and dou dizhu.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos beat top Stratego player Pim Niemeijer 15 wins to 1, with 4 draws over 20 games, which the authors call the first superhuman result in Stratego's history.
- The system was trained for roughly $8,000 on 16 NVIDIA H100 GPUs for a week, against an estimated $3-4.5 million for the prior DeepNash effort.
- The same approach set a new state of the art in cooperative Hanabi across 2-5 player variants and defeated PerfectDou and DouZero at Dou Dizhu.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos defeated Stratego's most decorated player Pim Niemeijer 15-1-4 in a 20-game official series, an 85% effective win rate.
- The system trained on 16 NVIDIA H100 GPUs for one week at a cost of a few thousand dollars, roughly 1/500th of DeepMind's DeepNash.
- The same method set new state-of-the-art on Barrage Stratego, 2-to-5-player Hanabi, and Dou dizhu, not just Stratego.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos beat the most decorated human Stratego player 15 wins to 1 loss with 4 draws, which the authors call the first superhuman Stratego result.
- The system ran on 16 NVIDIA H100 GPUs for a week, about $8,000 of compute, versus the roughly $3-4.5 million spent on prior Stratego AI DeepNash.
- The same recipe produced a superhuman Barrage Stratego AI and state-of-the-art results on cooperative Hanabi and the Chinese card game dou dizhu.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos defeated Pim Niemeijer, the most decorated human Stratego player of all time, 15 games to 1 with 4 draws over a 20-game series.
- Training cost under $8,000 at 2025 pricing versus roughly $3-$4.5 million for DeepNash: 1/500th the compute, 1/30th the self-play games.
- Same recipe hit a new state-of-the-art in Hanabi across 2-5 players and beat PerfectDou and DouZero in dou dizhu.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos beat Pim Niemeijer, the most decorated Stratego player in the game's history, 15 wins to 1 loss with 4 draws over 20 games.
- Training ran roughly $8,000 on 16 NVIDIA H100s for a week, about 1/500th the compute cost of DeepMind's prior DeepNash system.
- The same recipe set new state-of-the-art marks on Hanabi for 2 to 5 players, Barrage Stratego, and Dou Dizhu.
Read full analysis →
View on Bluesky →
Marc Lanctot reposted
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
AI Weekly's analysis
→
- Ataraxos defeated Pim Niemeijer, Stratego's most decorated player, with 15 wins, 1 loss and 4 draws across a 20-game series.
- The system was trained for 'a few thousand dollars' on 16 H100 GPUs for one week, versus $3-4.5 million for DeepMind's DeepNash.
- It also beat top Barrage Stratego players, PerfectDou and DouZero at Dou Dizhu, and set new Hanabi scores across 2-5 players.
Read full analysis →
View on Bluesky →
Their own posts
Recent commentary
I'm delighted to announce the release of OpenSpiel 2.0! ♟️🎲♦️🎉 Structured types for states, observations, and actions, standard trajectories (based on JSON), 19 new games, AlphaZero ported to JAX, Windows PyPI support, language model fine-tuning examples and an MCP server (demo below 🤩👇)! 🧵 1/N
The For You feed is getting amazingly good. Lately I feel like it's Twitter circa 2014. I have found more relevant stuff in my AI research meets retro games meets cat & dog memes niche than ever before, despite zero or negative DAU growth. Well done @spacecowboy17.bsky.social .. keep this up! 🫵👍🙌
Had a great time @coop-ai.bsky.social summer school yesterday! Learned many new things, was made aware of some great work, and met some people! I got all nostalgic being on UofT campus.. somehow haven't returned to the campus since we left Toronto 14 years ago. Saw some old faves downtown, too 😁🍁🦫🇨🇦
These days you're not a serious AI startup unless you're using "in-" as a prefix for your name: - Inflection - Ineffable - Inherent Let's look for a few more!
Their network
In Marc Lanctot's orbit
Center = Marc Lanctot. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.
Are you Marc Lanctot? Show it.
Add the Who’s Who of AI badge to your site or bio. It links back to this profile.
Markdown: [](https://aiweekly.co/whos-who/person/marc-lanctot)