Prediction markets have spent years getting better at one thing: getting the bet on-chain. But there's a harder problem nobody can solve with a simple price feed: What exactly happened? Imagine a market opens on FUD Markets on Solana: “Did this creator's campaign reach 1M verified views before Friday?” The market is fast. The transaction is cheap. The position is on-chain. Friday arrives. Now settle it. Which platform counts? What qualifies as a verified view? What if the number changes after the deadline? What if the evidence is spread across several websites? What if the answer depends on the meaning of the announcement itself? Suddenly, the hard part isn't executing the market. It's interpreting the outcome. That's where deterministic smart contracts hit a boundary. If the question is: “Was ETH above $5,000 at 12:00 UTC?” A deterministic data feed can provide the value. But: “Did this event happen as described?” requires evidence, context and interpretation. And if one administrator makes that final call, the market's resolution ultimately depends on that administrator's interpretation of the evidence. This is where @GenLayer powers market resolution for FUD Markets. GenLayer's Intelligent Contracts can work with natural-language criteria and live web information, while decentralized validators evaluate proposed results against the contract's rules. So the architecture becomes: Define the question → define the evidence → observe the event → evaluate the evidence → reach consensus → settle. That's a fundamentally different way to think about an oracle. The interesting part isn't simply: “AI can read the web.” It's: “Can the resolution of a subjective question become part of decentralized consensus?” And that matters because prediction markets on Solana don't have to stay limited to clean numerical questions. They can extend toward: • creator milestones • product launches • public announcements • internet culture • social trends • real-world outcomes that don't arrive as a neat number The more markets move in that direction, the more important the wording of the market becomes. A badly written market can still produce a bad outcome, no matter how sophisticated the settlement layer is. So GenLayer isn't a substitute for good market design. It addresses a different problem: how to reason about evidence when the answer isn't already sitting inside a deterministic data feed. That's why the FUD Markets + Solana + GenLayer combination catches my attention. Solana handles fast, permissionless execution. GenLayer brings evidence-based interpretation into the resolution layer. One gets the market moving. The other tackles the question waiting at the finish line: “Did the thing we agreed to bet on actually happen?” And that's the interesting frontier for prediction markets. Not just: “Can we put the bet on-chain?” But: “Can we make the resolution process worthy of being on-chain too?” If you were writing the rules for a subjective prediction market, what would you lock down first: the evidence source, the exact wording of the outcome, or the resolution process itself?