Comment by pmbauer

> I think there is some hazard in assuming that what agents fail at today they will continue to fail on in the future.

I think there is some hazard in assuming a seemingly exponential curve has no asymptotes, otherwise known as faith.