Stop feeding failed code back to the model below 7B, and A/B it above.

A placebo-controlled July 28 study found blind resampling beats self-repair at 2.5-5.5x lower token cost on MBPP+, because showing a model its own failed attempt makes it reproduce a near-identical program 33-68% of the time versus 2-14% under blind resampling. Real execution feedback added nothing over a content-free failure notice. Tested only up to 7B, so treat it as a hypothesis to test on frontier models, not settled.