Even allowing for tolerances, a breaker or fuse can carry 100% load indefinitely in an open air ambient temperature of 40°C. The three hours is a time limit from the NEC not the protective device. The problem is devices are mounted in enclosures often with other heat producing items.
Which is why, when you want a “100% rated” circuit breaker, it usually has to be either a Main, where it stands alone by itself connected to bus bars, a separately enclosed breaker, or all by itself in a switchgear / MCC cubicle.
The only exception I have seen to this is Sq.D I-Line panels and I think that’s because their direct bus connection is designed to help dissipate heat away from the breakers.
When a breaker is listed to be APPLIED at 100% of its rating continuously (aka “100% rated”), testing establishes that the correct operation of the circuit breaker can carry the full rated current without tripping
and that it does not exceed temperature limits that could render it unsafe. That test is required to be done in an enclosure with a
specific minimum air volume, which is why it is really difficult to get a breaker to pass when mounted in a panel with other breakers. It’s also required that the conductors connecting to a 100% rated breaker are 90C rated, but sized based on the 75C column.
So back to the issue of breakers being “80% rated”, it’s indirect. If, for a basic continuous load or one that is mixed with continuous and non-continuous, you size the conductors for 125% of that load as you are required, then you size the breaker to protect
those conductors, you will END UP with the breaker only carrying 80% of its rating continuously. Because the panel manufacturers know that this is how the breakers will be APPLIED, the heat rise testing they do for them MOUNTED IN THE PANEL is based on that 80% application.