That sounds like an attempted "gotcha", but I think you missed the "matter of perspective."
Imagine some strain of surviving bacterial-descendants are a marginally less-deadly than their predecessors after one solar year. What measure would you use for the comparison?
If you were to pick "generations", that might be ~9000 for the bacteria, while applied to humans it's ~40x longer than all recorded history.
Anywho, point is that for every "too big to fail" things there is usually a longer timescale where it stops looking that way.
That said, some of it is a matter of perspective: To bacteria, individual humans are "too big to fail" in the same way geography is.