Show HN: Finalrun – Spec-driven testing using English and vision for mobile apps

FinalRun just dropped a command‑line tool that lets you write phone‑app tests in plain English and store them next to your code, and the internet immediately turned into a courtroom drama. Fans swooned over the “finally, no more brittle test scripts” promise, while others side‑eyed the AI magic. One user asked the obvious: how does this thing actually understand English? Another threw roses for the choice to open‑source it, calling it “agentic testing” done right. A third voice went full “industry therapy session,” praising the move to keep tests with the app code and use vision (screen understanding) instead of those break‑every‑update locators.

Then came the twist: a sharp skeptic quoted the team’s own words about keeping tests in sync and asked if the bot’s generated test code actually gets saved in the project. If not, he warned, we’re just kicking the can down the road — today’s demo becomes tomorrow’s flaky mystery. Cue memes about robots “finally finding the Login button” and jokes about brittle XPaths sounding like brittle exes. Meanwhile, the team’s docs stress sane basics (don’t commit secrets; use .env files), and there’s a slick demo video plus the site. Verdict from the crowd? Hopeful hype with a side of healthy paranoia — and they want receipts on test persistence.