Experimental Jabhookery
Biology is complex but kinematics does not have to be. A handful of angles (wrist, elbow, shoulder; toe, heel, ankle) over time capture almost complete human movement. That data is manageable enough to read and process.
An early experiment in this area was to record pull-ups with a webcam and run the footage through a pose estimation (mo-cap) model to produce 3D coordinates for each joint. I fed the data to a frontier LLM and the raw matrix was far too unwieldy, but processed mathematically (isolating the joints at key muscle groups, taking min/max values etc.) it gave exactly the information needed for action quality analysis. Training based on the LLM’s feedback improved the pull-ups. The less powerful the model, the more watertight the math must be since the model will be less capable of drawing conclusions from complexity.
Jabhook runs this as a negative feedback loop for boxing:
How Jabhook’s biomechanics algorithm works (metaphorically): tracing paper with model form on it, held up to the user. The algorithm finds the gap between the outlines and explains exactly how to close it. Most forms of education use negative feedback. Students have their work graded and returned by teachers who are amazing though limited by working hours. This approach helps student's grow by reviewing and correcting mistakes. Jabhook allows someone training to do this, but based on qualitative data, muscle movement. It gives such focused feedback that it helps the user to adjust form completely