Baltimore

The little brother 🇸🇰 doesn't get salty. He gets motivated.
Congrats on the success. Truly.
But we're not here to clap forever. 🚀

Radim Sara @tomassvoboda.bsky.social Jakub Balada @davidcermak.bsky.social Lukas Burget Radim Burget Petr Sojka, Aleš Horák @jansedivy.bsky.social Pavel Kordik Tomáš Řehořek Pavel Matějka, Václav Mlejnský, Hynek Heřmanský, Jiří Boudal and many others, backed by KAYA, Václav, Tomas, Karel Zheng!

@stanislavfort.bsky.social Michal Pechoucek Jan Hajic Ondřej Bojar @jaroslavbeck.bsky.social Martin Schmid @matejmoravcik.bsky.social Rudolf Kadlec Tomas Gogar Petr Baudis Tomas Tunys Josef Sivic Jiri Matas Ondřej C. Tomas Pajdla Jan Černocký Filip Jurcicek @jindrahelcl.bsky.social Tomáš Pevný

"Any link with Knaster-Tarski there?" 🧠 Every time I go visit our big brother 🇨🇿 I am amazed by how fast their AI space grows! 🚀

Huge respect to the people building Czech AI: Tomas Mikolov Eva Hajičová Jakub Jurových Filip Stollár Lukas Toth Ondrej Vlcek

"And this is Mr. President 🇨🇿 how our neolab trains self-improving LLM agents using self-play and game-theory to push massive token throughput, and get new scaling laws..."

And then Petr comes back with the only acceptable follow-up: "Nádhera. But can you explain to me why this does not collapse?"

arXiv preprint coming soon → these are exactly the kinds of optimal algorithms we’re using our neolab that our pushing out multi-agent systems to the next level 🔥🔥🔥

w/ Côme Fiegel, Pierre Ménard, @Tdash_Koz @VianneyPerchet

We just closed this story for two-player zero-sum games, think rock–paper–scissors 🪨📄✂️, but harder. Even in the “no communication / only your own outcome” setup, we now have an algorithm that learns as fast as is theoretically possible, for the final strategy you actually deploy → not just an avg.

In game-like settings 🎮, the math is blunt: if agents can share rich feedback (what would have happened under different actions), they can learn fast ⚡. If they can’t, and each only sees the score from the move, they actually played → learning is inevitably slower 🐢.

“Just add more agents” is not a theory of learning. Communication is! 🤝

You can hear tons of AI bros say multi-agent systems are just parallelization: spawn a swarm, split the work, and coordination will magically emerge ✨.

Michal Valko on X: "Imagine... #AI! When Anita Hornikova recruited me to work with Frantisek Ruzicka and @DruckerTomas to organize AI Global summit in between AI Global Summits in Paris in Feb 2025 and AI Global Summit in Delphi in Feb 2026, the idea sounded crazy! Full: https://t.co/VrvHD6h72J https://t.co/PxPKiHtlTO" / X Imagine... #AI! When Anita Hornikova recruited me to work with Frantisek Ruzicka and @DruckerTomas to organize AI Global summit in between AI Global Summits in Paris in Feb 2025 and AI Global Summit in Delphi in Feb 2026, the idea sounded crazy! Full: https://t.co/VrvHD6h72J https://t.co/PxPKiHtlTO

Imagine... #AI! When Anita Hornikova recruited me to work with Frantisek Ruzicka and Tomáš Drucker to organize AI Global summit in between AI Global Summits in Paris in Feb 2025 and AI Global Summit in Delphi in Feb 2026, the idea sounded crazy!

x.com/misovalko/st...

@christinelagarde.bsky.social in FL series talks about the future. In the wake of indiaai summit, we met with her, TD, AS Gill, MC, KEE, RMfrom @oecd, @unitednations1945.bsky.social, @unesco.org, @ec.europa.eu, ECB to prepare AI future!

Whatch: x.com/misovalko/st...

(c) AK/ @bloomberg.com

Minister Drucker: Michal, you come from a world where people who move faster than anyone else does are actually building AI. Do you feel this community speaks with one voice about where AI should go, or is it still too divided between ambition, business, and science?

x.com/misovalko/st...

Closing remarks from the AI Global Summit, with the heads of states and general secretaries of @oecd-ocde.bsky.social @unitednations1945.bsky.social @unesco.org @ec.europa.eu @europeancentralbank@AmandeepSighGill @MathiasCormann
@roxanaminzatu.bsky.social @KhaledElEnany

x.com/misovalko/st...

See you in January 2026 in Abu Dhabi with @yann-lecun.bsky.social
@jpvert.bsky.social, Ivan Laptev, and other friends at Machine Can See!

Also, had a fun blackboard discussion with the great Sean Meyn, I share a few pictures 📷 playing with the blackboard signs 😄

Was very good to see Michal Valko again!
@misovalko.bsky.social

Journey of a lifetime to build something new ... #AI #RL #ReinforcementLearning #LLMs