One developer. One AI session. A full native iOS app—
auth, offline, CarPlay, sleep timer—from expo init to done.
Ridecast turns articles, PDFs, and URLs into podcast episodes for your commute. The web version—Next.js, Prisma, PostgreSQL, AI scripting and TTS pipeline—was working. What it didn’t have was a native app.
expo init. The native build starts.Screens are the visible output. The real work is the infrastructure underneath them.
Not because the tools got easier. Because the bottleneck moved. The constraint used to be typing speed, API familiarity, boilerplate. Now it’s taste—knowing what to build and when to stop. The mechanical work compresses to nothing. The decisions are what take time.
Data as of: March 15, 2026
Feature status: Active
Research performed:
git log --oneline --since="2026-03-11" --until="2026-03-12" --reverse (54 commits)git log --format="%ai %s" — first commit 12:48 AM, last commit 9:44 PM PSTgit diff --stat across day range (118 files, +37,313/−5,211)git log --grep="feat(native)" (20 commits with feat(native): prefix)ls native/__tests__/ (14 test files created on Mar 11)git log --oneline | wc -l (292 commits); git log --reverse --format="%ai" | head -1 (Feb 25, 2026)Gaps: Feature count of 49 and session count of 18 are from the AI session’s own summary and were not independently verified per-session. The “afternoon” native build start is inferred from commit ordering (design docs precede expo init) — exact clock time not captured. Line counts include package-lock.json (19,723 lines).
Primary contributor: Chris Park — 292/292 commits (100%)
It used to be typing. Now it’s taste.