Probably everyone. I voted on a Dominion machine this year. It prints out a full size piece of paper on thick stock with big font saying who you're voting for. You then have to carry that piece of paper to the ballot counting machine and stand in line for a minute before it's your turn to put your vote in the counter. The piece of paper isn't a receipt, it's your actual ballot. In the case of the elections this year, all of those were manually hand counted as well as being electronically counted and the counts were identical.

Did you notice if the ballot had a QR code on it? Did you scan the QR code, decipher its data structure and verify it matches your voting intent?

Because according to what I saw presented in hearings, the reading machines actually read the QR code made by ballot-marking-device, not the human readable portion of the ballot.

If true, that seems an obvious design flaw (to be nice) right there.

How would the machine predict which voters to deceive? I reviewed my printed ballot before submitting it. I'd guess it is probably more than "practically none". If 0.1% of voters did this, it would be 159633 ballots reviewed. If the tainted machines consistently changed these ballots to the same presidential candidate, ~80,000 voters would have had an opportunity to detect a dishonest voting machine. What ratio of votes would need to change in order to change the outcome? And what ratio would the bad actors predict was necessary in advance of the election?

But regardless of the above -- the design change to include printed ballots was a big improvement in security and voter confidence in the election.

Don't need to. The reading machines read a QR code generated by the ballot-marking-device. Or so I've seen experts testify.

The QR code could mis-match the printed ballot on say 5% of the ballots and nobody would notice a thing.

Need a recount? Ok, just run em through again, and they'll read the same (wrong) votes again, no problem.

Only a hand count of the paper ballots would catch the mis-match.

Did that actually happen? I don't know. But if the machines are truly reading a QR code, that seems a glaring flaw regardless.