"Only true peer-to-peer systems that need as little centralized elements as possible can survive."
I'm not sure how to interpret statements like this. Decentralization seems ideal. But some of the most successful services so far either begun or are centralized, not decentralized. If I'm not mistaken, the electric and water utilities, Google, and Facebook are all centralized for example. On the other hand, other services like email and the web (excluding DNS) are decentralized.
Is the intention behind this idea that control over those services shouldn't be centralized, while pooling resources could be ok, depending on the service involved?.
I'm not saying it's possible to remove control in a centralized service, but I am saying that there is a distinction. A blanket statement that "centralized services can't survive" would seem incorrect, and the existing statement of "as little centralized elements" implies that "as little" in the relative sense will be equal to little in the absolute sense, which might not be true. It could be that "little" practically equals centralization.
The reason such a misconception is dangerous is that one might put their efforts in designing a fully decentralized service that isn't adopted. Good design is often hard in that it challenges your assumptions in ways you hadn't originally anticipated.