"except when you spend 2 hours in a conference explaining why people should use free software. At that point, as a service user, RMS says something about someone else's computer."
As to "spend[ing] 2 hours in a conference explaining why people should use free software": Be careful about the trap of concision as explained by Herman and Chomsky in "Manufacturing Consent". It takes time to lay out new ideas, ideas which mainstream media never discuss. Software freedom is still (even 30+ years after the fact) regarded as a new idea.
Regarding saying something about someone else's computer: I don't clearly understand your point. Are you sure this isn't giving someone convincing reasons to favor free software? Can you point to an example of where RMS is a service user and makes demands of that service beyond recommending what they ought to do to get the software freedom they deserve?
"I agree with the logic of your argument, but it sounds like a lawyer's argument : it's ok by the law eventhough i'ts not super "ethical"."
What's not ethical about not having control over someone else's computer?
I doubt he explained anything proprietary as being "closed". He spends considerable time correcting people when they summarize his views and use that term because (as he also explains) that's a reference to open source, a group founded over a decade after the free software movement which ostensibly promotes a development methodology purpose-built to remove the ethical underpinning of free software. See https://www.gnu.org/philosophy/words-to-avoid.html#Closed for more on this. In practice, open source philosophy is disposed of in the face of powerful, reliable proprietary software ("How can I get a copy [of that proprietary program]?") and that's why proprietary software businesses like open source and don't care to discuss anything in terms of software freedom. But software freedom is far better for computer users in cases where the two philosophies conflict or diverge. See the section named "Different Values Can Lead to Similar Conclusions…but Not Always" of https://www.gnu.org/philosophy/open-source-misses-the-point.... for how this plays out in practice.
As to using someone else's tracker (tracking is what those devices do most of the time, so why not acknowledge that's what they chiefly are by calling them by their rightful name?): this is right in line with understanding that software freedom concerns control of one's own computer. Remote services (such as /., Hacker News, Google mail, and more) might run on nonfree software. That's a bad choice for the computer owner (all computer owners deserve software freedom) but it doesn't say much about use of the service because service users don't have a say about someone else's computer. By the same token, those that choose to own and keep trackers on their person have already divulged the location of the tracker (and thus themselves), and that choice was never Stallman's to make. So the choice for Stallman (or someone using someone else's tracker) comes down to making the call or not.
I'm not convinced that's an example of your point -- "[RMS] cares about his copyleft philosophy and GPL dogma more than the ability of free software to use free software". And some parts of the blog you pointed us to are simply wrong, like saying "[GPLv3] doesn't work for end-user software" is also wrong in that that's not true (there are many GPLv3, GPLv3+ -- GPLv3 or any later version -- and even GPLv2+ programs intended for users to run) and mischaracterizes the underlying issue. The heart of the issue has to do with understanding why GPLv2 licensors won't switch to GPLv3 (which would be fine but risk revisiting this problem again later) or GPLv3+. Furthermore, this is hardly an intractable problem, particularly among copyright holders who want to look out for their users' software freedom.
As http://gplv3.fsf.org/rms-why.html rightly points out, "both GPLv2 and GPLv3 are copyleft licenses: each of them says, “If you include code under this license in a larger program, the larger program must be under this license too.” There is no way to make them compatible.". So LibreCAD could relicense to be either GPLv2+ or GPLv3 which would allow making a program with code from LibreCAD and code from a GPLv3 program. Asking a GPLv3 or GPLv3+ project to relicense under a lower version GPL would be heading in the wrong direction for preserving user's software freedom. Asking the FSF to do a worse job of defending software freedom is a silly thing to ask.
GPLv3 has substantial practical improvements over GPLv2 which all GPL licensors ought to consider. A few examples: nicer termination terms than GPLv2 -- under GPLv2, any infringement causes a infringer to lose their rights under the license unless the copyright holder(s) restore those rights. GPLv3, on the other hand, grants an infringer some time to come into compliance to account for accidental infringement. GPLv3 has anti-TiVOization, anti-DRM, and patent licensing language not explicitly present in GPLv2. GPLv3 has a more modern source distribution requirement than GPLv2 (which was written before Bittorrent). Again, as http://gplv3.fsf.org/rms-why.html rightly points out, "Further advantages of GPLv3 include better internationalization, gentler termination, support for BitTorrent, and compatibility with the Apache license.".
http://gplv3.fsf.org/ has complete details on GPLv3's advantages and http://gplv3.fsf.org/rms-why.html concisely states multiple reasons to switch to GPLv3. There are plenty of good reasons to license under GPLv3 or "GPLv3 or later" to avoid this problem in the future.
As to "spend[ing] 2 hours in a conference explaining why people should use free software": Be careful about the trap of concision as explained by Herman and Chomsky in "Manufacturing Consent". It takes time to lay out new ideas, ideas which mainstream media never discuss. Software freedom is still (even 30+ years after the fact) regarded as a new idea.
Regarding saying something about someone else's computer: I don't clearly understand your point. Are you sure this isn't giving someone convincing reasons to favor free software? Can you point to an example of where RMS is a service user and makes demands of that service beyond recommending what they ought to do to get the software freedom they deserve?
"I agree with the logic of your argument, but it sounds like a lawyer's argument : it's ok by the law eventhough i'ts not super "ethical"."
What's not ethical about not having control over someone else's computer?