Overnight Dev: From Rage-Quit to Release in 72 Hours—With AI Riding Shotgun

Burned-out coder to live app in one weekend: two failed scrapers, one hidden JSON API, and AI tools that scaffolded the rest. How DevTools + GPT turned AES volleyball data into an MVP—and why your next Jira ticket might build itself.

Overnight Dev: From Rage-Quit to Release in 72 Hours—With AI Riding Shotgun
Photo by Andre Hunter / Unsplash

I. Friday Night — Friction

I’d rather shoot volleyball games than untangle legacy HTML—but AES (the platform tournaments rely on) makes schedule-hunting a pain, and planning my courtside visits becomes cumbersome.

Two quick shots at automation went nowhere:

  1. “GPT, build me a parser.” Generic prompt, brittle output.
  2. Classic HTML scraping. Race-condition hell; one div shift killed the script.

By midnight I was back to swearing at the editor, convinced I hated coding again.


II. Saturday Morning — The Breakthrough

During a frustration scroll through DevTools I hit the Network tab and finally asked the obvious question:

“What else is flying over the wire?”

AES wasn’t serving HTML at all—it was calling a tidy JSON endpoint that already held club names, matches, times, courts. No scraping required.

Pivot: Pipe that JSON straight into Supabase, seed a quick schema, build UI on clean data.

Lesson tattooed on the forearm: DevTools is still the best LLM.


III. Saturday Afternoon — Flow State

With data solved, I opened lovable.dev—AI pair partners that feel like senior devs who never tire of boilerplate. Fifteen minutes later I had:

  • A React/Vite scaffold with dark-mode toggles baked in.
  • Mobile friendly and responsive views.
  • Supabase row-level security policies I wouldn’t have written on a weekend.

Coding felt playful again—less “hunt the semicolon,” more “compose features.”


IV. Sunday — Epics as Prompts

Mid-flow a ridiculous thought landed: Could I ship an MVP from a single Jira ticket? The prompt was basically a beefed-up user story anyway.

Imagine a board where:

  • Gherkin = executable instruction set.
  • Grooming = prompt tuning.
  • Sprint Review = demo of AI-generated build.

The BA who writes the story becomes a vibe-coder overnight, pressing Run on their own acceptance criteria. It’s half joke, half forecast—and closer than it sounds.


V. What Changed (and What Didn’t)

  1. AI shrank the hate-zone. The gap between stuck and shipping is now a weekend.
  2. Engineering intuition still matters. The breakthrough was a human remembering to open DevTools.
  3. Prototype > speculate. Two discarded attempts exposed the real path faster than architecture diagrams.
  4. The premium shifts upstream. Clear, atomic stories out-value raw code hours.

VI. Closing Signal

I started the weekend dreading another brittle scraper. I ended it joking that Jira tickets might compile. The joke lands because the compiler is already here—and it’s fluent in Epics.

Ship early, scrap fast, keep the Network tab open.


(Feedback welcome—especially from any BAs ready to press the big green “Deploy Story” button.)