In science and math, too often only the final result gets credited or published.

Maybe we should start crediting and publishing our intermediate results and attempts.

Humans would gain credit for providing a part of the solution to tough problems, LLMs would profit from the resulting data flywheel.