No, I used a reasonable approximation. Sqrt(2^128) is roughly in the ballpark of what is needed to have 50% of a collision, in fact it is slightly lower than th

No, I used a reasonable approximation. Sqrt(2^128) is roughly in the ballpark of what is needed to have 50% of a collision, in fact it is slightly lower than the actual number. Incidently sqrt(2^128) is 2^64, which is the number I used.