The main problem the FSF runs into, over and over, is that they're barking up the wrong tree if their goal is political change (which it's at least stated to be).
Drew doesn't really address this directly, although many of the things he points at I would say are related to this. RMS is a big reason why the FSF is the way it is ie. its why their site is a disorganized mess of philosophical ramblings, poorly organized (and often badly written, resulting in people believing things about their licenses that aren't accurate[0]) legal FAQs and kinda poorly written predictive fiction attempts from Stallman.
Your average user isn't going to give a shit about this. Your average programmer already knows and builds on top of FOSS - FOSS is the great portfolio establisher of the programming world at this point. You don't need to convince them. Your average user deals with far simpler (and frankly far more dystopian) problems like "my entire workflow broke because Google removed one of their apps and I was completely reliant on it", "Spotify removed all music of my favorite artist and now I don't have access to the music I paid for" and "Microsoft made it so that I can't use my computer without a Microsoft account anymore".
The FSF has an extremely myopic view here that if it's not invented in FOSS land, it's not worth talking about. The GNU Project has an attitude that actively works against bringing people over to FOSS tooling unless they go all-in. Most people don't... do that, they'd rather gradually replace things as they become an issue.
Your average user runs Windows and has an iPhone or an Android. Your average developer needs to target these platforms. What does the FSF provide through the GNU for developers and users? Preachy nonsense about how you should buy a Thinkpad and live like a digital hermit. Sure you can barely participate in modern society, but you're at least free!
FOSS successes come when a project does attempt to bridge the user problem. LineageOS is an extremely useful Apache licensed Custom ROM that's widely used as an Aftermarket OS for Android phones for example. LibreOffice (formerly OpenOffice, MPL licensed) only became a household-ish tool because it openly was meant as an alternative word processor that strives to at least be compatible with Office (and most of its layout is basically borrowed from the last version of Office before they introduced the ribbon - aka the most production efficient version of Office). People gradually replace, but the FSF demands a total switch in workflows if you want to play on their turf, which for most people is frankly impossible.
They can call Facebook and Google silly names all they want, but grandma needs them to schedule her bingo club on Friday. And unless you want to take up the effort of teaching 10 octogenarians how to use a FOSS WebDAV calendar app and Matrix, how to set their devices up for it and more importantly, how that works when their tablet is replaced 3 years later. It's basically unreasonable unless the workflow and experience is exactly identical to that of the tools it's replacing. And it basically never is.
The FSF needs to start trying to work for FOSS being a reasonable option for end-users if it wants to live - the developer work is done, it's been done for upwards of 20 years now. Their insistence to keep focusing on the dev side has just let Google et al. create the dystopia they warned about just by ignoring FOSS (ironically the very thing the FSFs proposition claims was easily avoidable if we just build the tools. - We build them. Then Apple and Google just steamrolled the entire relevance of FOSS through either having proper UX design or shoring it up.)
[0]: Anything to do with the AGPL.