"I do not think there’s a contradiction between reproducibility and innovation. Contrary to common belief, we can improve both at once – by incentivizing failed

"I do not think there’s a contradiction between reproducibility and innovation. Contrary to common belief, we can improve both at once – by incentivizing failed results, and by funding “Red Teams” that would aim to refute existing dogma or would be entirely outside it." I don't know - the mechanism that creates the problem with reproducibility is that all teams are "Red Teams". What is published is what is unexpected, what deviates from a common understanding, but nothing confirming what we would expect. Most of what seems new and exciting comes out to be just random accidents - that is natural, but it is a problem only when we put too much weight to them.