Right - that’s a SHA1 collision, which has nothing to do with the possibility of a UUID collision.
I worry that some people might take away from this that UUID collisions are actually something to worry about (nonsensical kindergarten cubby analogies and so on).
Really, truly, UUIDs can be assumed unique. We’re not relying on luck.
UUIDs can be assumed unique because we are relying on luck, not the other way around. The cubby example is actually a pretty good example of why the luck is on your side to the point you should never worry about it. The SHA1 collision not as much.
