I see this shit time and time again. People looking at an RPG system and attempting to dismiss awful game design as a concern with 'The GM can fix it' or 'The system doesn't matter'.
It's fucking infuriating. Good game design matters precisely because it means a GM doesn't have to fix stuff. A good, well designed system gives the GMs the tools they need to make running the game easier and to support its premise. Not having to waste time wrangling with a shitty ruleset means a GM is more free to focus on running the game.
It's the stupid idea that a bad rule isn't bad if it can be fixed or worked around. The very fact that it needs to be fixed or worked around is what makes it a bad rule. If a system is making the GM work even harder to run the game, forcing them to jump through hoops or fix things on the fly, then it's a bad system.
This still applies even if you've gotten used to it. A good GM can learn to run a bad system well, but all the investment it takes doesn't stop it being a bad system.
This all coexists with the caveat that a system being bad doesn't mean it's impossible to have fun in it. Roleplaying games are primarily social experiences, the people you're playing with are going to provide you with a lot of fun anyway, but it's all about whether the system is a useful tool or a cumbersome burden on the GM and the group as a whole.
Almost all systems have elements of both, useful and shitty elements, but it's the balance of them which is key for saying whether a system is good or bad, alongside other things like mechanical tone, degree of options, character distinction and player choice.