I thought it was more scientific than entrepreneur focused... the competing goals of "science" are to think up a prediction generating falsifiable model, and to break someone else's model. Classic computer science analogy: Nothing get crypto CS more excited than proving some algorithm fails.

Some hard sciences are more open to failure than others. Good luck publishing a negative result chemistry paper, with the unusual exception if the purpose is to disprove someones previous published positive result paper.

Cranking it back to the article, although a "failure based curricula" would be brilliant for teaching history, especially history of sciences, its a complete waste of time for background stuff for average people. Its important for a historian of chemistry to know people used to be really "into" the four elements air water fire and earth, but probably not a good use of limited teaching time to test everyone on it.