Also, any serious discussion of election security has to grapple with two simultaneous realities:

- there's no evidence that any US election outcome has ever been altered by hacking

- there are real, exploitable vulnerabilities in many parts of our election infrastructure

I've written a bit on what these vulnerabilities are and how to fix them, See, e.g., this brief article:
https://georgetownlawtechreview.org/wp-content/uploads/2020/07/4.2-p505-522-Blaze.pdf

@mattblaze I think people focus too hard on theoretical impossibilities and not enough on actual threat models.

Paper ballots in Brazil had a habit of disappearing, teleporting, being damaged on route... In the US voter suppression and gerrymandering are so bad that frankly you don't need to hack anything to control what votes are counted.

I would guess that sometimes the security gains of a simpler, faster, more accessible process can be more important than using paper slips in an insecure process that is effectively impossible to be audited in practice.