Here's some good info to get you started.
Farad
MOSFET
On cogging, it happens whenever the ESC cannot properly sense the position of the rotor. It is far more prevalent on startup because the rotor is not moving quickly enough for the ESC to sense its position and commutate the correct sequence of current pulses for smooth rotation. You are right, at low speed the motor does not draw much power. On startup, however, there is a short high current pulse when starting the motor, as it's basically stalled for that short time. Capacitors help with cogging because they can provide that extra burst of current that the motor needs on startup, and so the ESC better senses the position of the rotor and there is less/no stuttering.
ESC's heat up because, like all other electronic things, they are not perfect conductors. There is a certain amount of voltage lost across the junction of the MOSFETs in the ESC. It isn't much, but losses, in the form of heat, are related to the current passing though the ESC. Say you run a 4s pack (14.8V) and there is a voltage loss of 0.5V after passing through the MOSFETs in the ESC. Ths would mean only 14.3V is getting to the motor. So where does the rest go? Well if your motor at a given time is drawing 30 amps of current from your battery, then by Ohm's law:
0.5V * 30A = 15W
Which means that the ESC is dissipating that 15W as pure heat. So you can see why there is such a fuss over creating superconducting materials. No voltage loss would mean no heat!