The probability of getting a UUID collision is different from an intentional attempt to find a SHA-1 collision. One is guessing a truly random ~128 bit number (really - it’s effectively the output of a CSPRNG). The other is exploiting a weakness in the construction of a 160-bit value that is intended to be a secure mapping of input -> output. It’s not even remotely the same problem and saying evidence of a sha-1 collision indicates anything about a UUID collision is plain wrong (and that part isn’t on a humor forum).