🎲 Odds of good conversations at Forecast & Coffee tomorrow: 99%. Odds of panels and long talks: 1%. Less than Jesus returning this year. If you're warming up for Manifest, the best prediction market conference, come help the first market resolve YES. Register ↓

🔮Kleros Foresight: Round 1 Recap Round 2 113 traders priced 16 films to predict our CTO's movie ratings before he watched them. The prediction market came closer than ChatGPT and Criticker's own algo. Round 2 is live, films picked by @clesaege himself this time ↓

But it is time to make it as such!

🎬 Movie taste is subjective, but that doesn't mean it's unpredictable. We wanted to find out which recommender was most accurate: ▸ a specialized recommender algo trained on 20 years of community ratings, ▸ the latest ChatGPT model, ▸ or Kleros Foresight, powered by prediction markets. Kleros Foresight won. In Round 1, 113 traders beat Criticker (avg error 32 vs 38) and ChatGPT (vs 42) at guessing what our CTO would rate five films. The caveat: it didn't beat ChatGPT when it had access to Clément's own chat history. That's why we're back with the Kleros Foresight Movie Experiment 2: Electric Boogaloo. Read how to participate ↓

Trade UP/DOWN on @clesaege's next ratings: movies-r2.foresight.kleros.i…

Prediction markets to make a super stable coin. Who wants to build this?

Detectors are great to flag or confirm feelings about AI use, but they are not perfect. When there is a suspicion of undisclosed AI use in candidate exercises, we always read it ourselves and look at the edit history.

AI agents are going to need a decision layer that isn't governance tokens you can't trust an agent to vote in a DAO you can trust it to put capital on a market and let outcomes settle futarchy is the natural interface for autonomous capital

🤖 What happens when you hand the same legal case to five different AIs? They disagree. at almost exactly the same rate human jurors do. That's one the findings in the new chapter @federicoast, @williamhwgeorge, and @robertgdean just published in AI and Arbitration (Wolters Kluwer, 2026), "When Decentralised Justice Meets Artificial Intelligence." 63 real Kleros disputes, judged by five frontier LLMs from: ChatGPT, Claude, Gemini, DeepSeek, Mistral. The takeaway isn't which model judges best. It's that you shouldn't trust a monolith AI. Round 2 of the experiment is already in flight: the team is re-running the test on real-world consumer cases from Argentina's Junín pilot and Lemon, where early evidence suggests the AIs and human jurors come to different conclusions on the same cases. Book details below ↓

one thing deep funding finally figured out is how to involve degens in public goods funding so far its been just the cypherpunks applying for funding & the regens building the allocation rails The core idea? > run a prediction market to forecast the results of funding rounds, so if project A has a market price of 0.1, that means the market expects it to get 10% of the funding in the round > degens earn a profit if their trades move the market price of the project closer to the actual amount it gets in the round Henceforth i really hope to see prediction markets attached to every public funding program!

If you have some time to read, let me give you my opinion about Kleros as of May 2026. I'm going to start by mentioning that the resolution of a medical-related case is, in my opinion, an enormous step forward for Kleros legitimacy as a suitable solution for dispute resolution in the legal landscape, since medical cases are generally subject to very strict regulations and heavy legal oversight. Even if this kind of applicable verdict may seem like small steps to some people, it shows that Kleros is becoming increasingly integrated into the legal landscape. As of 2026, Kleros real-world usage (if we define it as legally bound usage) is mainly focused in Latin America. If you look at the complaints about Kleros since 2021 (most of them started that year as it marked the end of the crypto frenzy in which most web3 projects were pumping without any real reason), you'll see that most of the criticism is focused on the V2 deployment, which is supposedly the main unachieved stepping stone for Kleros. But I don't think V2 is that important for now (even if I'd like to see it live too). Let me explain why. As weird as it may seem, I think that most of the work of the Kleros Coop isn’t in the coding/development part (maybe it’s half the job) but in the integration with legal public institutions and private companies (even if it's only with V2 Beta or even V1). As time passes, less costly and more efficient AI tools will make smart-contract deployment and reviewing easier and cheaper. Deploying a decentralized dispute resolution system will become doable for anyone capable of using AI with a bit of money and time. To keep it short, in ~5 to 10 years, any small team could be able to deploy a Kleros-like decentralized dispute resolution system. But why are the chances of Kleros being replaced by a competitor almost non-existent? Two things: institutional trust and juror-based staking. Institutional trust is extremely hard to replicate. Kleros has built real credibility over years with proven cases, successful resolutions, and partnerships. Companies and institutions won't easily trust a new fork with no track record. Juror-based staking creates a strong self-reinforcing loop. More cases attract more staked PNK, which improves quality and security. A competitor would start with almost no jurors and no staked capital, making it slow, expensive, and vulnerable to Sybil attacks, vote manipulation, or low-quality rulings. Aragon Court is dead. Celeste is dead. Jur is dead. Kleros survived thanks to a much stronger network effect driven by its critical mass of jurors and cases, a more mature and effective incentive system, a strong focus on real-world adoption with fintechs, companies and universities, and consistent long-term development since 2017. I think that the pilot cases in legal institutions, such as the ones in Mendoza Court, are the real stepping stones, because they show that the game theory behind Kleros is applicable to legal templates for fast resolution of almost any type of dispute. This is also why Kleros is focusing so much on working with universities. Having partnerships and research collaborations with institutions like the University of Oxford or Ohio State University validates the theory that the incentive principles behind Kleros indeed produce applicable and robust verdicts. I personally believe it is the best initiative and the smartest move from the Kleros Coop. And yes, Kleros had and probably still has some flaws in its development speed and in building certain side tools that weren’t really necessary. But today, you are witnessing what began as a small theoretical project 9 years ago transform into a fully functional decentralized justice system, applicable across the globe and to an almost unlimited number of use cases. You will have noticed that I only mentioned Kleros Court’s main use case and didn’t even cover all its derivatives (only Curate and Scout could be standalone projects by themselves). Better use Kleros $PNK.