Fantastic and exciting stuff! I wonder how much of this meteoric progress in actually creating novel mathematics is because the training data is of a much high

Fantastic and exciting stuff! I wonder how much of this meteoric progress in actually creating novel mathematics is because the training data is of a much higher standard than code, for example.