[PartiallyTyped]

I do not have a lot to add to the conversation, except some examples from ML, in particular, the Support Vector Machine was conceived much earlier than it was published but Vapnik's work was repeatedly rejected, yet SVMs are one of ML's brightest achievements.

Similarly, the Neural Network was repeatedly dismissed, with Minsky being a prominent example.

Not all ideas amount the same, but perhaps, at least in ML, we need to be open to exploring as many ideas as possible instead of performing grad student descent on the current consensus valley.