>
Sometimes it helps to contemplate the nature of remote possibilities. Most times, it does not.That's very true. And for a lot of people, and I'd guess HN has a very dense concentration of them, that's very hard to not do, because we've been trained to do exactly that, for professional work.
Cybersecurity is exactly that. What are all the remote possibilities you can come up with, and how do you defend them? Then reality hits, and it's so much worse than you ever feared, because you didn't think that Intel would release chips that couldn't keep secrets if you asked in the proper way. "I speculate, with my little gadget, the contents of all of physical memory..."
What's SRE, ops, any of those fields, but trying to imagine the remote possibilities of how things go down, and how you can mitigate it, or at least fail cleanly? And then, of course, reality hits, and you end up with someone shutting down a datacenter edge router with a dial up modem, because they'd been sufficiently paranoid about things to insist that there would still be a POTS route in, for when everything else went wrong and in-band signaling no longer worked. Which, of course, happened.
The high reliability hardware sorts, I expect, deal with something similar. "But what if a super high energy cosmic ray flipped this bit?"
I don't know how to turn this off when it's not helpful. I've done it too long. :/