Oh look great results with real world testing. This is the kind of results we should see more of in the wild

Somebody please tell me why this is not actually real result despite very good results and some reasonable benchmarks for self-coaching improvements.

Or maybe how it is real but its actually bad.

“What it can do is provide structured, personalized practice with realtime pitch feedback available 24/7 at a fraction of the cost”

This obviously directly contradicts the current zeitgeist perspective: AI can’t speed up your systems