I can see how I gave the impression that those writings were some substantial, coherent blog post when in fact they are several comments strewn across several previous rationality diaries.

I can summarize. One can easily Google up an impressive array of org-mode setups and blog posts describing GTD workflows which seem extremely compelling. The problem with all of these turns out to be that using somebody else's GTD setup in org-mode is like wearing somebody else's shoes. Whereas in a program like Nozbe every choice has been made in the most general possible way, setups that you'll find online will tend to be ultra-specific to the person from whom you obtain them, often in ways you don't even realize.

Then you are forced to start fiddling, and this is already doom, because you are supposed to be getting stuff done, not fiddling. But at some point you become seduced by this vision of having a hyperoptimized, automated system in elisp and plain text that practically runs your life for you and frees you from having to make any decisions, frees you to think about important things. Except, the opposite happens. You're spending more and more time on the damn org-mode setup, trying to figure out why something didn't get moved where it was supposed to, trying to figure out why a calendar item didn't remind you when you needed it to, searching for a note fruitlessly, botching a package install. The GTD literature harps on the importance of having a "trusted system" and finally you admit that you actually trust org-mode less than just writing things down on loose paper at this point.

At least, this was my experience. I never became proficient with elisp but I am a programmer. Maybe if I had started out knowing lisp things would have been different but I frankly doubt it.