Jev Pong
Play Pong against model decisions and watch latency change the game.
What Jev decides
Which paddle movement to make from the current game state.
Jev Pong advances the ball one step per model decision, comparing Jev with chat models in a live Ably-powered game.
- Jev role
- Choose the model-controlled paddle action.
- Input
- The current ball, paddle, and game state.
- Output
- A bounded paddle-action choice.
- Decision frequency
- Once per ball movement step.
- Fallback
- The game waits for the model answer before advancing the ball.
How it works
Overview
A human controls one paddle while a model chooses the other player’s move. The ball advances only after each model answer, making decision time visible in play. Jev and the comparison models receive the same bounded question; the author presents this as a latency demonstration rather than an intelligence benchmark.
- Language
- TypeScript
- License
- Apache-2.0
- Author
- ably-labs
Sources
Related projects
Jev Autopilot
A Jev autopilot flies a simulated drone across a generated Three.js city and lands on a target pad.
What Jev decides
Throttle, yaw, pitch, roll, whether to commit to landing, and whether to cut the motors.
TypeSafe Mario
Jev plays Super Mario Bros. from structured emulator telemetry rather than screenshots.
What Jev decides
The next controller action, whether jumping is useful now, and the immediate danger level.
jev-ultrafast
A browser agent where Jev selects an operation and a compatible page element from a live indexed action space.
What Jev decides
Which supported browser operation to run, and which compatible visible element to target