This is a super interesting short paper!
Not for the formal results, which I confess I don't really understand. But for the meta-science.
It discusses how the same model can be implemented differently in code. So there's a need for independent coding, even when the researchers agree on the model.
Are scientists incentivized to share intermediate results? Some recent work has argued that credit-maximizing scientists would share intermediate results, whereas other work has argued that sharing to...
