nah, he's correct that the area of a circle (which is the general shape a bullet projects through the body) is based on the formula (pi)r^2, and that it adds up to a lot more difference than you'd think. I think he's exaggerating a bit though.
So if you have a circle with a diameter of 9mm, that's 4.5mm, squared, times pi... 63.5 square millimeters of frontal area.
a 10mm is 5mm squared, times pi: 78.5 sq. mm of frontal area.
that's if you were shooting ball ammo basically.
the thing is, 9mm (36), .40 S&W, and .45 all seem to open up to more similar diameters than you'd think. I believe they actually range from around .65 for the 9mm to around .75 for the 45acp.
The reason why the rounds' opened diameter is (by percentage) more similar than their starting diameter, is because the 9mm rounds which open up that wide have similar amounts of kinetic energy to work with (compared tot he .40/.45 loads). So they all open up by about .30 caliber further than they started. Which makes them more similar in final result.
a 9mm opened up to .65 (16.25mm) comes out to 207 sq. mm of frontal area.
a .45 opened up to .75 (18.75mm) comes out to 276 sq. mm of frontal area.
which really is a pretty good comparison of their potential lethality, assuming you made borderline-lethal shots with both guns.
of course, shot placement is infinitely more important than making a hole a couple millimeters wider... and that's where the lethality differences fall apart somewhat. You can be killed with a .22LR, and knowing that the hole in your heart is smaller than what you consider to be effective, won't make you less dead. As a result, once you reach calibers where penetrating through ribs/bone is not a problem, and you have bullets which expand reliably to make a bonus-sized wound cavity, then you start to get really solid one-shot-stop numbers. Is 96% better than 92%? Sure. It's not what you'd call a highly significant statistical difference though.
In statistical analysis, there's a dramatic shift in the effectiveness of variables when you are working towards a 'hard-cap'. In terms of shooting people, the target dying is a hard-cap on the effectiveness of your bullets. So when you go up in caliber to the point at which bones offer no defense, and rounds open reliably to make more lethal/violent wound channels... then the total wound displacement (which yes, .45 has substantially more than 9mm) becomes only a marginal difference. Because there's the issue where once you have "enough" reliable wound channel creation to reach the hard cap with a high probability (to frequently kill the target), then any additional hypothetical bullet potency will suffer from extremely high diminishing returns.
which is fancytalk esplain'n why 207mm of frontal area pushed through a body gets 92% effictiveness, while 276mm (33% more) of frontal area pushed through a body only gets 4% more effectiveness vs. the hard cap (death/incapacitation). Diminishing returns.
Which is also why, 44 magnum doesn't do any better than .357 or .45. Diminishing returns. You're basically maxed out. The only way to get above that 97% effectiveness range is with a bazooka. And even then you'd be looking at a 99.X%
But, that's all way off topic for .357 vs. 10mm to stop bear... except perhaps, insomuch as the hard cap for lethality vs. humans is much MUCH lower than the hard cap for lethal effectiveness vs. a 500lb bear. So while .40/.45/.357 may be 97%-ish vs. humans, even 44magnum is no where remotely near 97% 1-shot-stopper vs. a 500lb blackbear (at least in a charging situation). It's not even close to the hard cap, leaving plenty of room for more potency without wasting it on diminishing returns.
Which is why I'd go with more firepower than 10mm or .357 if possible.
10mm/.357 are both 'good' but shooting a bear with them is about on par with shooting a human with a .32acp... it may or may not penetrate bones adequately, it can get the job done, but shot placement is at a premium, it's a lot better than nothing, but it's potency vs. the target isn't overtly comforting.
Mind you my ccw is still a .32, so maybe i shouldn't talk.