X @Elon Musk
Elon Musk·2025-08-10 15:53

Training Improvement - The model should feed the results back into training to improve the one-shot probability [1]