Claude Formanek

Hey @dwarkesh_sp! @NomadsVagabonds @gilg_oscar @TomTaylor256 and I already did this and had similar findings. teapaw.org/NomadsVagabonds/status…

Lecture recordings and a blogpost recapping the Summer School will be released soon. Sign up to our mailing list to learn about future events: cooperativeai.us1.list-manag…

We also announced the Best Poster Awards from our Summer School poster sessions. Congratulations to @ClaudeFormanek, whose poster on ‘Molten Pot: Evaluations & Datasets for Social Offline Reinforcement Learning’ received the top prize.

Today we're releasing Orbit v0: a framework for multi-agent safety and security research. If you've tried running multi-agent experiments, you know the drill - weeks of infra before your first result. Built on Inspect, Orbit lets you swap environments, threat models, and topologies without rebuilding anything.

Can LLM agents coordinate in long-horizon, open-ended worlds? We evaluate 13 modern LLMs in a new benchmark where agents must work together to explore, communicate, trade resources, craft tools, build structures, and fight mobs. TL;DR: Most agents struggle, averaging only ~6% normalised return. Yet on the hardest setting, zero-shot Gemini 3.1 Pro performs comparably to the best MARL agent trained for 1 billion environment steps. More broadly, we find coordination is a distinct bottleneck beyond long-horizon task competence, with communication having the largest effect in our harness ablations. 🧵👇

ARC-AGI-3 is built different, it has dumbfounded almost all regular attempts so far because it's so much harder than anything that came before. It has no rules, it's agentic and has no explicit goals, they need to be discovered. @tufalabs won the first milestone of @arcprize > There is no language built into the benchmark, but these guys "put the language back in", because in their view - it's the best way to climb up the notional "abstraction mountain" and effectively use many of the abstractions which have evolved over millions of years of language evolution. > They built a novel harness "The Duck" around a 27B open weights model (Qwen 3.6) to solve extremely challenging and novel reasoning problems that require abstraction. > This is the launch video of their winning agentic harness, "The Duck". We have also released an exclusive interview with them on MLST, just dropped. > The million dollar question is: what will @fchollet think about how they've done it, and is this a step towards AGI?

𝗦𝗶𝗹𝘃𝗲𝗿 𝗺𝗲𝗱𝗮𝗹 An incredible week for a team that was not ranked in the top 3 before the tournament begins to fall just short of the gold. We are proud of this team, and proud of the support they received. Its always about #MoreThanHockey

Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇

It's been a long hard journey for Bafana Bafana and now We Are Here📍Let's have a look back at the road to the 2026 FIFA World Cup Qualification. Are you ready Mzansi? 🇿🇦 💛💚🔥 ​#BafanaPride @adidasfootball @adidasZA @REXONA_SA @StandardBankZA @10bet_ZA @flysaa @CastleLagerSA @CocaCola_ZA @SABC_Sport @NUTRITECHFIT @sociosdotcom @SouthernSunGrp @HonorAfrica

IT'S OFFICIAL: After many years of hard work the Sanlam Cape Town Marathon has officially become the 8th Abbott World Marathon Majors race in the world. On 23 May 2027, we take our place alongside Tokyo, Boston, London, Sydney, Berlin, Chicago and New York. This is only possible because of the thousands of runners who kept on running and believing in the Cape Town Marathon. 🇿🇦 22 - 23 May 2027, here we come 🎉🇿🇦🇿🇦

The energy is unmatched 🇿🇦🕺 #FIFAWorldCup

If you’d like to master the foundations of cooperative AI and gain confidence to work in the field, apply for our next ‘Introduction to Cooperative AI’ course. Running online from 6 July - 28 August 2026. Link below.

We’re transforming Google Antigravity into a scientific workbench. The new Science Skills bundle allows researchers to run complex workflows like protein analysis in minutes using specialized Alpha* models and 30+ major scientific databases.

Our beautiful 'Great African Seaforest' featured in @TIME 🌊😍 "At the edge of Cape Town’s city lights, one of the ocean’s most intricate ecosystems is held together by creatures most people have never seen." 📲 Full article: time.com/article/2026/05/28/…

We're delighted to announce the launch of the Africa Jobs Fund, housed at @RenPhilanthropy. A new initiative aiming to mobilize $100m to create high-productivity jobs for Africa’s next generation, led by Daniel Yu, founder of @wasoko_official, and Ben Hyman, with @SamanthaJPower and Iyinoluwa Aboyeji advising. Read more here: Africajobsfund.com Get in touch at [email protected]

We’re rolling out 3.5 Flash to everyone in the @GeminiApp and AI Mode in @Google Search. Developers can start building in @Antigravity and via the Gemini API in @GoogleAIStudio. #GoogleIO

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

Congratulations to Professor Benjamin Rosman, Director of Wits University’s MIND Institute, who was appointed by Communications and Digital Technologies Minister Solly Malatsi to lead South Africa’s AI policy advisory panel and help shape the future of AI governance and innovation in the country. 💙🏛️ Read More: ow.ly/H4BO50YZtto #WitsForGood #InnovationForGood