Is Iterative Machine Learning Really a Crisis?

I was just talking to someone yesterday that said the opposite of what I'm about to say So maybe I'm saying something silly But I kind of feel like we treat Publications as like you can get a publication some publication in anywhere. If we could actually treat things Iteratively it would be better because then you can start on this process and then someone else can like pick it up And maybe someone else who like knows a little like which things to tweak would be the next one to look at it yeah.