Free software isn't supposed to be anything except free. If it's superior that's nice but that isn't really the goal.
The first one I think is pretty important, as the saying goes if the cure is worse than the disease its not worth it.
Security flaws is arguable, not enough people have an incentive to work on free software, not because it doesn't get used, but because companies themselves haven't and aren't built to contribute back. We could have solved this fairly easily, just not in the software world, but I digress.
Lastly, more features, by which I don't mean in the sense you are probably thinking i.e. the narrow mindset popularized by MVP thinking "can it do X". More in the sense, "can you do X in Y manner". If you have ever used software and thought "why does it do this stupid thing or take so long" etc., its what I mean. Corps often build MVPs and "working" software that is otherwise horrible, free open source software is often used and improved by end users. Technical users, often, but they wear a user hat as well. Users can complain and request things directly.
It can produce bloated software as well, but when you don't have a budget you can do things like refactor. It will take a while but it can happen, where I think often corp software is held back by layers of deadlines, product owners, and BAs - the users are so far at the end of the chain the software they learn to live with is an absolute nightmare. They use it for some business reason, not because the software is in the least bit good.
Superior, yes, but only for "some definition of superior."
It's not meant to be "better in every way" and could never be that anyway since "best" of anything, is always a compromised nuanced collection of competing interests. With the FSF there aren't competing interests, there is one overriding interest and all others are subservient to that.
Thus GNU software is always the "most free", but is also unsuitable for most users. Emacs may be free, but my mom uses MS Word. Her definition of "best" might be different to mine, not least because her goal is writing a letter, not analysing source code.
To the original author I suggest the following - let the FSF be the FSF and decide for yourself what tools best suit you. There are plenty of alternatives to GNU software for anything that matters. If gcc doesn't do done thing the use clang etc.
The "pure" FSF approach to computing is not meant for the masses, it is meant for a tiny few. You are not one of those few. (I mean that in the pragmatic, not pergoritive sense.)
Programmers see computers as "a thing" - 99% of the world see it as just another tool to get through the day. My mom cares about her computer about as much as I care where, or how, paper is made. She cares about her software about as much as she cares about her coffee machine.
In one sense RMS is a visionary leading us all to software nirvana. In another he's irrelevant. In many ways he can be both at the same time.
It doesn't matter whether emacs does this, or gcc does that. We don't really need a more pragmatic FSF - there are already lots of open source advocates filling that space.
If you want to be pure, the FSF is the home for you. If you want to be more pragmatic then find somewhere else.
Personally I am not one of the tiny group of perfectly-pure users with no reasonable phone or laptop. Personally I think the FSF stance is absurd in most cases and out of touch in most others.
But I completely respect their mission, and I completely enjoy many of the freedoms that mission has brought me indirectly. Open Source exists because of the FSF stance - a happy side effect of it maybe - and that is a very important thing.
The world needs extremists who model an alternative extreme. Homesteaders don't make me want to live off-the-grid, but they do inspire a model of more sustainable living, I'm encouraged to use alternative water for some uses, perhaps install some solar and so on.
Long may the FSF set a standard for impractical, unwavering dedication to an unrealistic computing standard. They set a boundary for the rest of us to see in the distance, and they allow a middle ground far closer to that boundary than we would otherwise aspire to.
The FSF is gloriously irrelevant, and long may it last.
Stallman's principles are important, the man saw really far into the future towards which we have been heading. However, in the GCC AST case these principles really got in the way of good work.
Free software is supposed to be the superior software. It seems obvious to me that superior software would be easy to interface with and build upon. GCC should have been no exception but he chose to make it hard on purpose because some proprietary software somewhere might interface with it some day and he didn't want to help those people. It's my understanding this is one of the many reasons why clang is so relevant today.
This unwillingness to help proprietary software shows up in other parts of the free software culture as well. For example, nobody seems to care much about binary interfaces, only source code that can be recompiled.