What Was I Watching - Remember every show. Share the good ones.

Hey Product Hunt! 👋 I built What Was I Watching because I kept having the same conversation: "What was that show we started on Netflix? Or was it Hulu?" — and nobody could remember. So I wanted a dead-simple app to just capture what you're watching, when, and where. But honestly, the app isn't the most interesting part of this launch. The process is. Before writing a single line of Swift, I spent 3 full days doing nothing but writing specifications — with Claude Code as my co-pilot and speckit a the tooling. We wrote a project constitution (9 principles, amended 16 times). We created 28 feature specs. Then the AI asked 60+ adversarial clarification questions about edge cases I hadn't considered: "What happens when a universal link arrives during onboarding?" "How does data actually flow from iPhone to Apple Watch?" Only after all that did we start coding. The result? 81 commits in a single day. When you've already made every decision, implementation becomes almost mechanical. The full app — camera OCR, Siri integration, Apple Watch companion, iCloud sync, home screen widgets, Spotlight search — shipped in 17 days. 582 commits. ~18,000 lines of Swift. It wasn't all smooth. We killed 3 features that looked great on paper: Voice dictation had 3-4 second latency — felt broken, removed entirely A 7-step onboarding flow — deleted and replaced with a single welcome card Background auto-enrichment — users wanted control, not magic Those failures were caught fast because the specs made expectations explicit. When reality didn't match the spec, we didn't patch — we cut. I've been writing up the entire journey as a blog series if you want the deep technical dive. Would love to hear from other makers: have you tried front-loading decisions before implementation? Did it help or just feel like overhead? Happy to answer any questions about the app, the workflow, or what it's like having an AI co-pilot that argues with your specs before you write code. 😄