I understand why people would like discussing GPL with RMS, but RMS is not a legal scholar nor has a legal background or education. If we look into the history of GPL we see that RMS himself did not arrive at the finer legal details by himself, but rather through the help by his friend Eben Moglen who is a professor of law and legal history. RMS tend to focus on the philosophy and the strategic use of GPL towards free software goals.
The concept of "linking" in GPL also came from Eben Moglen. The legal theory was fairly simple, ie, would a non-technical judge see a linked program as one large single copyrighted work or two independent separate works (originally in US legal context only)? RMS initially viewed it as two works, but Eben Moglen had a different view about it, which RMS then changed his opinion to.
Technically, dynamic linking looks like use, and not the creation of a combined or derived work.
Using an interactive dynamic language with FFI, I can connect to a shared library at the REPL and interact with its functions, just as if I were calling programs in /usr/bin.
While I have presented a series of arguments, which in my opinion could not be refuted logically, it was not possible to change RMS's mind about anything, so the discussion has ended with everybody keeping their original opinions, unchanged.
Nevertheless, I am immensely grateful to RMS for his past work. I am a very good programmer and i could have been able to write all the programs that I use daily, including an operating system, device drivers, compilers or a text editor.
Nevertheless that is a huge amount of work, which would have taken many years of my life. Due to the open-source movement, I was able to benefit of the work shared by countless others, thus saving an enormous amount of time.
I usually take advantage of the open-source nature of the programs to make customizations that would have been impossible when using proprietary programs. However, making various modifications to an open-source program is orders of magnitude less work than writing the program from scratch.
While the great quantity of open-source programs existing now is the work of a very large number of developers, RMS has played a critical role in catalyzing this evolution, which must not be forgotten.
I also agree with the author of the article that in general all the targets set by RMS about free software and hardware remain as valuable and important as ever, but the strategy and tactics advocated by him are no longer effective and by now it cannot be expected that he will ever change in the future.