I agree with you, @Daniel Krasner, that making people rethink their conception of computing is a laudable achievement.
Yet when you look back on the analogy with Socrates and Plato, Socrates certainly was "the originator" of many ideas that inspired some of his contemporaries. Yet without Plato and Xenophon, who actually wrote the ideas down and shared them with the world, Socrates would be forgotten, possibly with his ideas. Who knows what the world would be like in that case...
The point is, the goal should be not only to generate ideas, but to generate knowledge. Ideas may eventually vanish, but knowledge is (usually) preserved and accumulates. You might recall a recurring theme in Alan Kay's talk about Da Vinci, Newton and Ford, and his point about knowledge and context being superior to smartness and/or great inspiration.
I believe Dynamicland is trying to generate "knowledge" (through the working system), and not only ideas. However, if the group falls apart (and according to the article summarising the sociologist's findings, there's at the moment just one person apart from Victor working at Dynamicland) without sharing any of its findings, then the knowledge generated by the ideas around Dynamicland would remain in the form of PaperPrograms and other such prototypes. These may not reflect the full vision of Bret Victor, but at least they're openly available.
I think it's a trivial observation that out of knowledge that's lacking but openly available, and knowledge that's complete but esoteric, the suboptimal one proliferates. I don't want for the knowledge around Dynamicland to be lost. I believe (hope?) that the group is still functioning, but naturally I'm worried that exactly that might happen. They don't want to be sharing their progress openly (at least not online), which will result in a loss of the whole system if they're undertaking too big a problem.
This is partially what happened to the Frank system developed at VPRI. The working system is available, but not discoverable (I have a link to it somewhere, but there's no other way to get to it, I believe) nor working out of the box (and naturally has no documentation). However, VPRI at least published many writings that documented aspects of the system; which is something that Dynamicland lacks.