This kind of stuff that Mr. Pink (someone should invite him out to dinner and find out if he doesn't tip :P) mentioned and guys like Alfie Kohn harp on about is intriguing, but there is one huge hole in the plot that I have never seen them address: The Problem of the Paycheck.

The studies that are being referenced lead naturally to the hypothesis "what would happen if I didn't give my employees any kind of artificial incentive at all?" (let's assume that they are all knowledge-work employees, say programmers for example).

The studies referenced in the talk and in Alfie Kohn's books suggest that you would see an increase in productivity, since that pesky, harmful incentive has moved out of the way.

Common sense (and experience, should anyone be foolish enough to actually try it) suggests that your workers will simply walk.

So while I think there is a lot of valuable insights to be gained from these experiments (e.g. people work like hell when something has meaning for them - see the works of Viktor Frankl for more on this), I think there's too much of a tendency to oversimplify and say "incentives are bad, mmkay?"