By Hiram BarskyFounder & Principal Designer
Role
Design · Code · Art direction
Timeframe
2025
Published
July 2026
Outcome
Live, with a real player leaderboard
Built with
Lovable · Figma Design · Claude Code · Netlify — web, no install
The result
Shipped and playable at firelion.netlify.app, with a live leaderboard of strangers trading high scores — one designer, AI in the loop, taken from idea to played by total strangers.
Overview
Fire Lion is a word-casting arcade game — spell words, cast fire, clear the board. It exists to prove a point: one designer, working with AI, can take a game from idea to something strangers actually play.
The problem
Small games usually die in the gap between concept and playable — art needs an artist, code needs an engineer, tuning needs a team. That's the same gap that kills product ideas everywhere. We wanted to prove one designer with AI can close it.
Research
What kills small games isn't a lack of ideas — it's the specialist gap between concept and playable.
Three specialists, one budget of zero
Art needs an artist, code needs an engineer, tuning needs a team. Solo projects die waiting for skills they can't hire.
→ Drove: AI in the loop for all three — accelerating each craft while authorship stays with one designer.
Vague scope burns speed
AI acceleration compounds only when pointed at a sharp target. Spent exploring a fog, it just produces fog faster.
→ Drove: A core loop tight enough to state in one sentence before building anything.
Installs filter out curiosity
Casual players won't install to try. Every step between 'heard about it' and 'playing it' loses people.
→ Drove: Web-first distribution — a link is the entire onboarding.
The approach
- 1
Designed a core loop tight enough to explain in one sentence: type words, chain combos, survive the escalation.
- 2
Used AI to move at team speed without giving up authorship — art, copy, and gameplay tuning all accelerated, all still mine.
- 3
Shipped to the web, where playing costs one tap — no install, no app store, no excuse not to try it.
Our thought process
Fire Lion was the proof run. If we're going to tell clients the gap between idea and shipped is crossable solo-with-AI, we cross it in public — and then do it again.
What didn't work
Early versions grew features the way side projects do — power-ups, meta-progression, unlockables. Every addition made the build slower and the game fuzzier. Cutting back to the one-sentence loop is what un-stuck it. The discipline that shipped this game was subtraction.
The outcome
A finished, playable game — and a working demonstration of the solo-with-AI process behind every product on this site.
What this means for your project
Have an idea stuck in the gap between concept and shipped? This is the whole pitch. We take it from sketch to something real people actually use — fast, with AI in the loop and the craft still deliberate, not left to chance.
“Always innovating and looking to improve. I would suggest him for senior and leadership roles.”
