4 ms·
I generally agree with your statement about trust systems - but my point is that we have to end up trusting a lot of people anyway, even if it is a matter of tr
by ergothus 8y ago
I generally agree with your statement about trust systems - but my point is that we have to end up trusting a lot of people anyway, even if it is a matter of trusting the people that design the systems. I can't verify everything about my life and still have it. So we need to design not only based on mistrust, but also include post-facto verification systems so if questions are raised we can go back and apply the distrust we were too busy to do originally.
I was going to say that I can't sign up as a poll-worker because I'm in a mail-in-ballot state, but it appears I'm wrong. I'll look into that - it's definitely an accessible (for me) form of civic service. Thanks for the suggestion!
- 3pt14159 8y agoThe core difference is that if you trust, say, 5 people. You can quietly send them to 5 random polling stations and get them to relay back what happened. Basically even the extremely conspiratorial can trust the election if there are observers that can be there from start to finish. Saying "just trust our government's employees" is not a real solution.