It keeps being brought up. I suspect as a sort of character assassination technique. Sort of the disaster analysis version of bike-shedding. "well I don't know much about the physics of carbon fiber in high pressure environments, but I know a cheap gamepad when I see one." However, I expect that such a commercial off the shelf gamepad to be probably the most reliable well tested piece of equipment on the sub. it is highly unlikely to go bad and is super easy to carry a spare.

Also no analysis I have seen actually knows how the sub was controlled or bothers to explain how you would do it better than a gamepad. If I had to guess, the motors and ballast controls went into an array of controller boards in a rack somewhere. the control circuity(probably something like i2c or canbus) tied into a central microcontroller. and then a laptop or tablet is used for the human IO(gamepads and screens). There probably was not much redundancy in the system, but I am not sure that is a bad thing, redundancy massively complicates the design and introduces new error conditions. (look into the causes of the US Navy destroyer John S. McCain collision, it was a highly redundant controller system. and the helmsman failed to realize he had no control, it was really a training issue) often it is better to just have the ability for manual control. that is, go back to the rack of motor controllers and manually start twisting knobs and jumping wires. but none of this matters, because it is not what failed.

It cheapens the narrative to go after stupid superficial things like the gamepad, that did not cause or contribute to the fault rather than what actually caused the engineering failure.