Interesting that if they had just grabbed microseconds at launch, kept it statically, and mixed that in with their seed, they would have made it far harder to g

Interesting that if they had just grabbed microseconds at launch, kept it statically, and mixed that in with their seed, they would have made it far harder to guess. An even better choice would be to also time the microseconds of a few disk writes at launch or grab entropy from network latency, but even one simple timing value would have made these seed prediction attacks impractical at the time.