Side projects we build to test current AI models on problems we know well. They aren't products and we're not selling them. Most of our client work is under NDA, so these are the things we can show.
An AI driver lapping Averton Park, in a car driven by a motorsport-class physics model. Everything on screen apart from the logo and the font is generated in code, the car included.
Runs in the browser. Best on a desktop, with the sound on.
A brush-model tyre with a thermally coupled tread and a pressure-dependent carcass. Double-wishbone kinematics with chassis compliance. A 566 PS flat-six through a slipping clutch and a Salisbury differential. Every slide, kerb strike and lock-up comes out of the simulation.
The engine note is synthesised from individual combustion cycles through a modelled exhaust, and the tyre squeal follows published squeal acoustics.
120 channels across eight pages, with any two laps overlaid and a running time delta, for anyone who wants to check the physics model's working.
Most of the sim was built with Fable 5.0. The second version's graphics, AI driver and telemetry tools were done with OpenAI's Astra.