I don't think this is true given the likely relative participation rates. Unless a large stakeholder sets the Refund standard particularly high, it's very likely that any faction able to drum up support on a one-time basis will be able to fund a proposal, simply because most witness voters aren't here all the time. (And if the Refund standard is particularly high it likely precludes all non-blockchain-development projects entirely. Perhaps it would be good to have separate pools.)
Consider 50% curation as an example. Given the support it has been shown, it's almost certain this could gather enough stake to vote into the development queue. Yet we thankfully aren't particularly close to a supermajority of witnesses supporting it. But if code development were funded and finished, it would essentially lock that faction into place, never allowing them to move on and leading to permanent conflict over the issue.
History has shown in the case of bitshares that the refund standard was set very high, and in such a way that it easily weeded out questionable proposals.
In general, a lot of your arguments make assumptions that I find very questionable. For example, I think you're flat out wrong on the support level of witnesses for 50% curation (along with some of the other proposed changes to curation). In my opinion, I think a strong majority were in favor of the idea, and the primary blocking issue was just finding someone willing to do the work for free when Steemit devs didn't show a lot of interest in taking on the task themselves.