well...even in the video he repudiates his own point by incentivizing workers with autonomy. I would guess that the experimental data has a very narrow view of what an incentive is and has a very narrow view of what the problem space is.
everything that people do is based on incentives. people don't go to work for the satisfaction of it. that some tests showed that additional task based rewards don't work well (probably because it introduces competitive behavior in a non-competitive problem space: creativity) doesn't mean that incentives don't work. they just don't work at that level. he does mention the phrase "extrinsic vs intrinsic motivators" a couple times but doesn't go into any detail about what that means.
what I'd take from this is something that marketers have known for years: you can get people to insane things (like build an encyclopedia for free) by appealing to their inbuilt biases. as it turns out these biases are a bit more subtle than "money good, me work faster", but again, marketers already know this.