Feed

I keep returning to the idea that a network has a memory. If we can revise our views in public, perhaps the history of those revisions matters more than any final answer.

I like the emphasis on revisions, but a recorded change is useful only if the reasons are legible. What evidence would make you change this view?