He's got decades of this under his belt. Ask him about whether red sauce or green sauce is better for drive-through tacos, he'll talk about freedom and oppression the same confounding ways.
As for myself, I'm much more in the Torvalds camp. Substantive freedom is a practical freedom as well as a prophylactic one; it's the freedom not just *from* things but *to do and participate in* things.
Sure, I want the freedom to protect my data or change my software. But I also want the freedom to buy and use a device that I think is great, or to participate in the mobile ecosystem (sorry, RMS) because I find it to be useful.
RMS can't distinguish between the two, or between the different kinds of restrictiveness at issue—the commercial software restrictiveness that is certainly annoying and terrible for our world, but also the FSF-styled restrictiveness that shoots itself in the foot and ends up being exactly the same.
In both cases, the end result is that I can't do what I want with my software/hardware. The commercial interests are at least open about it: we don't want you to do that because it would hurt our profits. The FSF is less open about it: we're not responsible for this, it's their fault—you're free to do your own thing.
Yes, maybe in theory I could rewrite the entire GNU codebase from scratch or get a world of developers together myself to do my own thing, but substantively speaking, in terms of actual opportunity structures available to me right now, today, or next week, or indeed for most people *as themselves, during their regular lives*, there is about the same amount of substantive freedom and restriction involved.
If I want to do X with my tech, and company X won't allow it with their toolchain, and the open computing world won't support it for ideological reasons, the net result is still that *I* am *practically* unable to do X with my tech.
Part of the FSF problem is that they often delegitimize X. RMS's answers about, say, "cloud" computing or the mobile ecosystem are instructive here, and mirror common answers in free software developments from techs. "That is not a real thing, it's just marketingspeak, you are a victim of ideology, and no, we won't help you."
Operating under the Thomas theorem and using the well-respected argument made by Rawls, I'd say that RMS fails to distinguish between summary rules and rules of practice. For RMS, there are only summary rules—things that we decide or don't decide to do, and espouse for utilitarian reasons. All of his arguments are utilitarian in nature (though often convolutedly so). Even when they involve other people or "society," his arguments boil down to rational self-interest calculated according to a very narrow range of values and goods, discounting the rest.
He ignores the dimension of rules and practices that are oriented toward social life—toward behaving in ways that others understand and that enable one to substantively participate in public and group life by virtue of conceding them as ordering principles for "how the world works right now."
The FSF vision of computing is, ironically, radically individualist and lonely in this regard—it is all about "what I can accomplish on my own." The only "we" that it acknowledges is one that is made up entirely of people that have precisely the same ideological outlook, goals, desires, and summary rules as the self. All other forms of "we" are reimagined as secretly selfish people that *claim* to be a public, but are in fact actually seeking to dominate one another. For RMS, "we" hasn't happened yet and he is trying to bring it about through summary means—as a rational self-interest calculation.
But a world of identical "free-people" in which the "we" finally comes about by virtue of the universal embrace of FSF values simply doesn't and won't exist—people are different, desires are different, and that which is in one person's self-interest is never necessarily in everyone's self-interest.
To believe in the existence of the group and participate with the group on group terms as an end in and of itself—to acknowledge and concede that different people have different forms of self-interest, but that we can all establish norms of practice and collaborate nonetheless in intelligible and useful ways, while *at the same time* also taking care, in a different way, to nurture our own self-interests within this framework—is beyond RMS.
I don't think he's just willfully antisocial, I think he actually doesn't get it. Everyone, to him, is purely a rational utility maximizer in a very simplistic way—it's about money and control, not values, not social life, not interests, not identity, and not long-term planning or life balance.
As for me, I'm happy to "allow" the phone networks to surveil me so that I can receive calls from my daughter more often, and when I do this, I see myself as participating in a very real, normal thing—communication, as it is normally done today, with understood concessions that (if and when the time is right) may have to be adjusted. This simple understanding is totally beyond RMS as far as I can tell.