MCG x EasyA Ecosystem Day $FIT | @fittercircle w/@shaolin_vjk Fitter Circle is an AI powered cashback app that helps busy shoppers find relevant products within seconds Key topics: 00:43 - Meet the founder 01:32 - The thesis 03:55 - 500+ brand partnerships in the UK, 100+ in the US already signed 05:08 - The Groupon learning 06:09 - ChatGPT has replaced Google as the first search destination for product research 07:30 - The mechanic 08:55 - No brand logos in the app (purposely hides them so users describe what they want, not which brand) 10:14 - iOS app live in the UK, but intentionally not marketing yet 14:18 - New brand partnerships signed 15:55 - Economics 16:35 - $FIT loyalty tiers 18:38 - Considering an equity round prioritized for $FIT token holders first 24:01 - Roadmap

Launchpads & capital formation When you're ready to launch a token or raise from the people who already use your product: → @echodotxyz + @legiondotcc — community rounds that turn your users into your first investors → @Pumpfun, @bonkfun, @BagsApp — launchpad infra → @LaunchOnSoar — non-dilutive token raises, keep your equity & most of your supply → @EasyA_Kickstart — fair-launch idea coins on Solana

Every time I've shipped AI features at a startup, I've ended up building the same thing: an eval system to A/B test prompts properly. One-shot and few-shot testing both failed under real conditions. Eyeballing diffs between prompt versions doesn't scale and isn't reproducible. So I built one again at Fitter Circle, a quick way to simulate user queries at scale, instead of manual testing. A fixed query set, with scorers on the things that matter and regressions visible at a glance. Best decision I've made on the AI side this year. A few posts coming on how it works, what I've learned, and the methodology.