[whatshisface]

Another Machiavellian thing I have seen in the literature related to "Science 2" is where ML benchmarks or test cases only become accepted when they show that a lot of people's models are working. ;-)