I lurk. Sometimes I actually have an intelligent thought, so post..
How ya doin??
The reason for the twist is to decouple the pair from external influence by a time varying magnetic field. For cat5-e and up, the 4 twisted pairs were twisted with different pitches to prevent communication between the pairs. So while the twisting does indeed "choke" the rfi/emi, it does so by eliminating the reception of the nefarious stuff. The term "choke" is kinda like an element put in place to remove or block something that got in. Yes, it is a nitpicking distinction to your accurate statement, but I'm big on nitpicking.
Also like I stated before, if its peak current which is very short duration, I would suspect that the power supply caps are still charging during this state so whatever benefits this cable is trying to provide from a current carrying capability is actually reduced from the charging times required by the power supply caps.
I did not think it necessary to repeat what you stated, you were quite concise and accurate on those points.
Actually, I've used 24AWG copper to carry 400 amps. It was at a temperature of 1.9 kelvin, and it was a pulse with a time constant of 200 milliseconds. Given those conditions, the wire never exceeded 100 kelvin.
One must make a distinction between the ampacity of the conductor/insulation scheme, and the results of a transient test. If you examine the trip curve of a standard breaker, you will find two areas of interest.. The thermal curve, which is used to protect wires from slight overloads which can possibly cause fire, and the magnetic curve where massive fault currents will clear the breaker within a few line cycles.
Since the graphs are in the microsecond realm, clearly a 14 awg run will not go up in flames.
Cheers, John