Incredible explanation! You should consider making this its own post so I can rehive it!

The return proposal is an innovative use of the proposal system to set a floor on which proposals are funded.

When voting for proposals in the future, I need to consider not only whether I support the project but also whether the requested funding is reasonable. I didn't understand that nuance before.

Seems that the proposal system is susceptible to vote selling, similar to content and witness voting? For example, imagine a proposal that distributed its payout to everyone who votes for it. Perhaps even with some mechanism to reward early votes that allow it to surpass the return proposal more than later votes. Ignoring reputational damage for supporting this proposal, the incentives seem to lead to this outcome?

Other commenters here seem to think the @steem.dao holdings are too large. I tend to disagree. Development is a crucial ingredient for growing and sustaining the blockchain. @steem.dao is a way in which the blockchain can effect change outside its own protocol, but in support of its protocol. It's a major step in blockchains become self-sustaining and adaptive entities, almost like a digital life form.

But this really depends on whether the system is used to fund valuable work or whether it is susceptible to embezzlement.