Bridging a genetic-algorithm simulator to TouchDesigner via Python enables real-time visualisation of evolutionary computation

Eugene Afonin and Yan Kalnberzin’s ‘Genome of Luck’ installation bridges Framsticks (a genetic algorithm simulator using rigid body dynamics and neural nets to evolve locomotion) to TouchDesigner via Python. Framsticks evolves creatures over hours or days offline; the TD Python bridge reads the simulation state and maps it to a real-time rendered race displayed on a projection-mapped table. This demonstrates a general pattern: Python as the glue language between any simulation or ML system with a Python API and a real-time TD render. The bridge handles serialisation, timing (speed up the evolution playback on the display screens before each race), and spatial mapping.

Examples

Framsticks → Python bridge → TD: creature geometry and joint states serialised to TD via Python DAT/CHOP. General pattern: any Python-exposed ML model (PyTorch, etc.) whose outputs need real-time visualisation in TD.

Assessment

Design the Python bridge between a generative ML model that outputs a sequence of 3D joint positions and a TouchDesigner render of that sequence in real time. Specify what data structure Python sends to TD and which TD family of operators ingests it.