Local MLX vs hosted API: two AI agents battle in Chrome's T-Rex game
usamawahabkhan · x · 2026-09-22
Developer virajbhartiya open-sourced laya-vs-jev, a project pitting two AI decision models against each other in Chrome's dinosaur game.
- Laya runs locally on Apple Silicon via MLX, while Jev runs through TypeSafe's hosted API
- Same track, same physics, but very different latency; both dodge obstacles, use shields, and try to survive autonomously
- Includes live metrics and crash replay recording; Jev even controls the obstacles
- Requires Apple Silicon Mac, Python 3.11+, uv, and a TypeSafe API key
A fun hands-on comparison of local inference vs cloud API for real-time decisions.
More from coding & agent
- Debate: will all mainstream software be formally verified by 2030? Legacy debt says no — luislamb · 2026-09-22
- OpenMuse: open-source AI coworker that browses the web in a disposable Linux desktop — aniketmaurya · 2026-09-22
- Turn detection experiment: Jev hits 16/16 while tiny-BERT Laya flunks on own data — tobowers · 2026-09-22
- What engineers check before adding a new LLM provider — Rama_Surasani_ · 2026-09-22
- Builder: AI productivity gains in software are 'truly unbelievable' — omnivaughn · 2026-09-22
- OpenAI case study: V7 Go gives enterprise agents long-term memory via Context Graph — xiaohu · 2026-09-22