As you said, the neuroscience is pretty adamant: To be good at something, you have to practice, practice, practice. (It also may help instill a solid work ethic, which is by far the most important thing a kid can learn from school, in my opinion.) Obviously there can be overkill, but you have to do repetition to learn something. And yes, it's true that for some kids a few problems in class is all they need to get it down--but they are probably a fairly small minority (in the above-average intellect range).
A few other things to consider:
1. Beware the title "education expert", because nearly every person I've heard called this has been little more than an expert salesman, rather than an expert in education. (See: Robert Marzano, Douglas Reeves, Willard Dagget, etc.)
2. #1 Above is probably caused in part by the terrible standards for educational research and education colleges in general. It continues to shock me how easily so many people in charge--most of whom have Masters and Doctorates in education--of our educational system are so easily fooled by the fad du jour.
3. While I do like Alfie Kohn, I have to disagree with his "crusade" against homework; it seems focused on attacking the research opposing his beliefs, rather than providing direct support for said beliefs. I agree (#2 above) the research in education is, generally speaking, poor. But he needs to provide some evidence (research) specifically showing that homework doesn't help, which is not the same as showing that the research showing it is helpful is shaky. To me it seems his argument generally boils down to "I don't believe homework helps, the research supporting homework is inconclusive, therefore homework doesn't help."
= new reply since forum marked as read