CONFESSION TIME: after 7 days and $35K+ in volume, our current version of Breakout, our latest product, is not meeting our expectations. In build-in-public fashion, here’s my honest post-mortem around why First, why shut it down this early? Because we can. With our new AI-native stack, shipping a completely new product now takes weeks instead of months. Lower iteration cost raises the quality bar. Showtime (our company) is "in the business of hits", not sunk-cost coping. We’re already building the next one and we're extremely excited about it, which gives us confidence to share the learnings below without hard feelings. WHAT WE GOT WRONG (product lessons) ↓ Great prediction markets are moments, not trends. The best markets concentrate attention into a sharp, live spike: elections, finals, breaking events. That urgency drives volume. And vice-versa, the worst markets reflect continuous trends (e.g. “BTC price in 15 min” or "weather"), exactly the kind to get arbitraged by Claude bot hustlers. Our CT mindshare markets were closer to trends than moments, structurally capped in excitement and flow. Long-tail prediction markets are a bad business. Unless you capture an absurd number of micro-markets (which we don’t want to), the math just doesn’t work. Even at an optimistic $1M/week in volume with a 2% fee, you’re looking at just ~$1M/year in revenue. It’s far better to own the major markets (Polymarket / Kalshi style) than to bet on the long tail. There are deeper technical learnings too, but these two product truths explain most of the outcome. New product coming very soon, being cooked as we speak, AI x crypto too 🫡 https://x.com/BreakoutApp/status/2018416324893094052?s=20