> Some of the best scientists were and are unconventional late bloomers. Sure, but that's beside the point. The funding is limited, and, what's even worse, if

> Some of the best scientists were and are unconventional late bloomers. Sure, but that's beside the point. The funding is limited, and, what's even worse, if we spread it too thin, best scientists will leave to make millions on the wall street instead. At some point, a decision must be made as to who gets to proceed in the research career, and who is cut. Sometimes the apparent mediocrities will turn out to be late bloomers, and cutting them would be a net loss. However, and this is critical, you do not know who will be a late bloomer until they actually are. Late bloomers do not know that either, of course. In practice this means that when it comes to make a decision, you, and every other reasonable person, will pick someone who looks good on paper over someone who looks mediocre, but has a chance of becoming a late bloomer. Doing otherwise would be reasonable only if mediocrities tended to have on average better output in the long term than people who look good on paper, but that would only cause us to revisit what the metrics of what makes one look good on paper, because they would clearly then be wrong. They aren't: by and large, people who do well in exams tend to do better in actual research than people who suck at exams, even though the exams don't actually resemble actual research much. In short, when you bring up potential loss of cutting late bloomers, you should not forget about best scientists who never became one, because they were outcompeted by a mediocrity who never became a late bloomer. > Many grants proposals are largely ‘me too’ derivative works. Yes, and to me this is a sign that too many people are involved in research. If we suffer from deluge of mediocre, derivative work, we should make the standards higher and put more wood behind fewer arrows, not push more mediocrities into the career.