Unlikely, they are already pretty good at writing tests. What stops someone from running agents in the same loop to generate a test suite against the real appl

Unlikely, they are already pretty good at writing tests. What stops someone from running agents in the same loop to generate a test suite against the real application, and then turning around to generate the application after?