The idea that people will better learn from spending several weeks doing something useless (i.e. unnecessary design) instead of having that headed off at the pass is silly, when the specific thing they should be learning is to _not spend several weeks working on something without feedback_.
What if more could be learned / gained from several days or weeks of mistakes in the grander scheme?
I can't tell you how many times I've had to throw out code or designs from longer explorations, but the things I learned in getting there, for the sake of technical purposes, or heck, new ideas I could apply elsewhere in the future, were entirely worth the trouble. This is a pretty fundamental part of increasing skill — to be forced to struggle on one's own, especially if that person's personality is suited to the solitude needed to solve a problem.
Obviously it's smart to avoid larger mistakes or going in the wrong direction for the sake of the product and delivery timelines, but I think it's more realistic to expect these longer adventures to happen, and that an element of trust is necessary to allow that person to not only solve the problem, but grow in that journey.
I don't think that's what GP was suggesting at all.. everyone in the thread agrees that feedback is good.
However, the opposite extreme, spending time in meetings where lot of good ideas get proposed but most of them get killed, because somebody doesn't like a detail, or nobody champions them strongly enough, is not a good learning experience either.
