From iOS to Web in Two Weeks: How AI Coding Is Redefining the Product Engineer
summary
Claims a full iOS AI photo/video generation feature is ported to the web in two weeks using Claude, ChatGPT, Cursor, GitHub Copilot and Windsurf, shipped to real users and real traffic. Key specifics: kept the feature in its own folder to enforce isolation, ran a dedicated consistency check pass to replace duplicates and reuse constants, and found a memory-leak under heavy scroll that required human diagnosis before AI fixed it — so shipping works fast but deep QA and stress testing remain critical.
AI-written summary · about
chapters (16)
- 0:00AI-driven iOS to web migration
- 1:05AI coding enabled two-week rewrite
- 3:18Three levels of 'done' beyond code
- 4:42Level one isolation prevents changes
- 6:54Managing shared services with AI code
- 7:16Level two consistency fixes duplication
- 9:25Upfront context fails under deadlines
- 10:32Level three silent bugs in real use
- 12:43Diagnosing memory leaks in production
- 13:49AI speed vs engineering problem-solving
- 14:54Writing clear prompts as skill
- 15:36Reviewing code faster than writing
- 16:41AI shifts engineering work
- 17:43Three takeaways for shipping with AI
- 19:08AI preserves senior engineering value
- 19:31Demo two-week web feature
related talks
18mBuilding AI-powered video generation tools for web and mobile platforms
47mAI-Powered Software Development From the Trenches
35mApp Testing with AI Tools – Make it Easy
25mAI-Assisted Swift in Gamifying an iOS App
27mPixels and prompts: building products in the AI era
24mAI and Web Development: Hype or Reality
37mBest AI Tools for Web Developers / How to Use AI in Web Development
26mAI, You complete me