Here's the thing: If you built something like Quicklaunch instead (https://en.wikipedia.org/wiki/Quicklaunch), your g forces would already be an order of magnitude lower. So the question is, why not do that instead?
They could have lowered the g forces by having a longer arm, they essentially picked the g forces as acceptable and then designed the system around that constraint.
As to why not use a gun, barrel damage is probably a deal killer for Quicklaunch and other such designs. Quickly spinning something in a vacuum and letting go should allow for launches roughly as fast as you can pull vacuum. Supporting something like a Starlink constellation seems possible with their design which could be incredibly profitable.
Starlinks are pizzabox-shaped. That's a terrible shape to launch unless you have a stack of them, and this won't be able to launch a stack of them.
As for barrel wear...it's possible that that would be higher than for a normal gun, but I'm pretty sure that the rather famous author of that design took that into consideration. (It doesn't seem like SpinLaunch's design is immune to firing wear either, considering the sudden rush of air into a space with hypersonic machinery in motion.)
You can have a door close behind the projectile to limit the amount of air that gets into the main chamber. I suspect spin launch to include that simply to reduce the amount of vacuum pumping needed. The idea showed up on one of the early space gun designs for similar reasons.
Starlink’s shape is optimized for it’s launch system, while I doubt spaceX is going to use a competitor anytime soon they would seriously consider sufficient cost savings.
You can reach somewhat higher accelerations by making the center of the rotor thicker than the periphery, but you start getting into ridiculous ratios pretty quickly.
I don't think Quicklaunch would have to suffer significant barrel damage, maybe some fouling that can be cleaned off. The pressures and temperatures required are pretty low, and the required chemistries are pretty tame; it's just the velocities that are high.
The g forces are from rotation not how fast you spin up. To take it to the extreme an arm that’s slightly longer than the radius of the earth for sometime moving at Mach 22ish hits ~1g, because that’s orbit. Hitting Mach 22 or even Mach 10 on an arm that’s 500m means vastly higher g-forces.
And how long is the arm supposed to be in the intended orbital system? Because if you want a 500m arm, you literally have a millions-of-cubic-meters-sized vacuum chamber -- and not only that, you have a vacuum chamber with a highly disadvantageous shape (a flat disc). That sounds like a mildly insane design to me.
Submerging the gun barrel in the ocean presents huge engineering and maintenance challenges. It's not impossible but building anything underwater is extremely expensive. And then you have to deal with currents, corrosion, storms, leaks. People who aren't experienced in marine engineering tend to drastically underestimate the costs.
One thing I was thinking about some time ago was that if you wanted such infrastructure while still being able to adjust azimuth and elevation, perhaps an artificial lake would work? At the very least you avoid salt water this way. Of course at that point you can't pick and choose your launch site after the fact, but that's not a regression compared to stationary launch pads anyway.
