I was just crowing the other day with my parents about the filesystem resilience of the simple Amiga Workbench system. I reminded Dad how you can't just pull the plug on Windows since 3.0. You can't pull the plug on DOS if you're in an app, for that matter. But Amiga, you could just hit the power button whenever you pleased, and the filesystem was none the worse for wear, because it was designed well.
Back in the heyday of Raspberry Pi 3, I was tinkering with a kit, but I moped into a Linux User Group meeting, complaining how often I'd ruined an sdcard filesystem after being forced by circumstance to pull the plug. And a nerd, an AWS Engineer by profession, ignored or misread the situation of the humble SBC, and he told me 'just install ZFS' and also 'stop pulling the plug'.
He did not understand the Raspberry Pi, and I let him know exactly how I felt about his idiotic, misplaced advice. Yes, sure, ZFS for the big systems with gobs of memory and UPS backups. Personally I kinda liked btrfs.
Also the author of today's article cannot spell "gigabytes" and for a storage specialist, that's a real bad look, kids. Learn how to abbreviate, and you communicate better. Spell your units wrong, and you appear uneducated and inexperienced.
But any filesystem in 2026 that simply can't survive a power-cut does not deserve to be in an OS that is High Availability and Redundant and promising >4 nines of uptime. I mean seriously. If this guy's hacks involve mounting rofs everywhere, then ufs (or whatever OpenBSD still uses) has failed again, in 2026, after 56 years' worth of trying to get it right.