This is all kinda blind speculation and there will be a formal report eventually, but as a general outline:

-When carbon fiber fails, it tends to fail spectacularly: completely and suddenly. So you can think of it not as "crushing a tin can" but more "smashing a glass lightbulb, but from all sides at once".

-If we randomly assume they were halfway down (no idea on where they actually were but as a blind guess 50% is a good starting point) that's about 200 atm of pressure. 1atm = ~15 psi, so thats about 3,000 psi. For comparison, a typical firehose is roughly 100 psi. And that can do serious damage to people: if a badly threaded cover pops off a charged hydrant, there is enough force behind that to break bones. If you were sitting next to the hydrant it'd hit you faster than you could react - you'd only know it after you'd been hit. The water outside the sub is at 30x that pressure.

-Lets assume just as an arbitrary approximation that in the first instant of the carbon fiber failing catastrophically, an area roughly equivalent to a 3ft diameter circle fails (it probably actually fails by buckling in a line then milliseconds later splitting and shattering, but we're just approximating). This means that the water that flows through is pushed by 30x as much pressure as a firehose, and that pressure is coming in across 200 times as much area as a firehose (which are typically 2.5in diameter), so there are basically 200 of those 30x-power-firehoses coming through at once.

-A 2.5in firehose will do ~300 gpm. 6000 firehoses would be 1.8 million gpm. The internal volume of a 2m diameter/4m long cylinder is about 2,500 gal. That would be completely full of water in 0.001 seconds. Of course in reality water doesn't hit full speed instantly, fluid flow is far more complex than just multiplying through like this, etc. But this just drives home that we're talking very very small fractions of a second.

-Yes, compression = heating and when its super fast there isn't much time for heat transfer so its adiabatic: wikipedia has an example under "adiabatic compression" for 10:1 compression going to about 500dec C (in an engine) and this is more like 200:1. But remember that air has low specific heat capacity and also doesn't weigh much. The specific heat capacity of water (i.e. humans, plus those 6,000 firehoses worth of water) is ~4x that of air, and the density is ~1000x as much. So if you have equal volumes of air and person, and you heat the air by 4,000 deg C, that contains roughly enough energy to heat the person by 1 deg C. And also refer back to "there isn't much time for heat transfer". So chances that this actually matters beyond detailed physics calculations are slim.

Bottom line: completely obliterated by the force of so much water under so much pressure. By the time any water entered the sub it should have been over faster than a human could perceive. No explosions or incineration though, just force.

Also, common misconception: pressure alone doesn't hurt you. You would not be directly hurt by spending time anywhere from the complete vacuum of space (0atm) to the challenger deep (1,000 atm). Obviously there are other little complications like you can't breath in 0atm and that'll kill you quickly, but the pressure itself won't. Conversely at high pressures oxygen becomes toxic which isn't great for staying alive, but the pressure itself isn't the issue. Very rapid and therefore very violent pressure CHANGE, however, can and will kill you in many horrible ways.