My current car (like all of the cars I've had since 2008) allows you to set a temperature target with physical controls. I don't change it as often as I changed the older direct controls, but I do change it often enough that I would hate to use a touchscreen for it.
Even that isn't guaranteed to work, because different sets of people in the same place might have different time zones based on political affiliation. That's rare, but it has happened so it certainly could happen again.
I'm not certain that that's what is happening, but I strongly suspect it. It is if nothing else a useful mental model: switching my strategy from selecting every square that contains foo to selecting every square that I think that Joe Schmoe would select gets me through CAPTCHAs more consistently.
The journal article is open access and available through the DOI link. Short version, as I understand it (take with more than a grain of salt): the mechanical process does not degrade the C-F bonds at all, but when you grind quartz that is already sufficiently fine it produces some highly reactive species on fresh surfaces which chemically attack the PFAS. At least some of the C-F bonds are replaced with Si-F bonds, though I'm not sure exactly what the products are.
Security against unauthorized use and data lifespan are separate concerns. They're not fully orthogonal—security tends to make things more brittle—but you can apply whatever form of security you like and then store the secured data in any way you like. Hardburn seems to have been talking purely about the useful life of the archived data. The charge in flash storage leaks, so the data is eventually lost if not refreshed. A flash drive is reliable for a year, but not a decade. If you want long term storage you're going to want something else. Paper would be fine for most uses. An ordinary printout subjected to ordinary handling is good for a few decades with reasonable storage conditions.