EXP-001AI Travel Planner
Three coding agents, one 1,245-line product spec, eight hours each. Build a real AI travel SaaS and deploy it.
- CCClaude Code
- CXOpenAI Codex
- CUCursor CLI
Experiment library
Each experiment is one versioned spec given to several coding agents under the same constraints. Filter by where it stands, what kind of software it is, or which agent took part.
4 experiments
EXP-001Three coding agents, one 1,245-line product spec, eight hours each. Build a real AI travel SaaS and deploy it.
A real browser video editor with a multitrack timeline, AI captions, plain-English edits and a genuine MP4 export.
Five-on-five basketball in the browser, with real ball physics, keyboard control and an AI opponent.
A stationary 3D lane shooter where you steer an auto-firing squad, recruit soldiers and survive ten levels.
No experiments match those filters yet.
The idea board
Ideas under consideration. None of these has a spec, a run date or a result yet. A strong challenge has a memorable premise, real difficulty, a demonstrable artifact and acceptance criteria that can be tested objectively.
A newsroom for general news that gathers, writes and edits stories, built from a single spec.
An editor, file system and run loop entirely in the browser.
Multiple users editing the same thing at once without losing work.
Storage engine, query layer and durability guarantees, all of which can be tested.
Vehicle physics, damage models and AI drivers smashing into each other.
A physics sandbox where builds have to actually work under gravity.
A physics puzzle game built around linked portals, momentum and level design.
Simulated robots competing under shared rules, with scoring that has to hold up.
Simulating a city grid and measurably improving the flow of traffic.
Coordinating many robots through a warehouse without collisions or deadlock.