Actually, consider (say) Fulton county. Of 120,000 votes, 100,000 had to be "adjudicated". You really do have to ask what the /real/ vote was, if they felt the need to "interpret" the votes of 5/6ths of the voters. One just /might/ think this is wildly racist disenfranchisement.
Nobody really knows how folks in these cities actually voted, and the Democratic Party would like to keep it that way.
Section 230 is actually VERY important for small sites. Removing it will only keep the powerful in power.
What needs to be done is to actually enforce it. In particular, if you're 'fact checking' someone, you're producing your own content, and that's publishing.
We want neutral platforms. The trouble is these folks are using the platform shield while publishing. I think SCOTUS had a recent ruling demonstrating areas folks can certainly sue under Section 230.. This will get better. Really, people need to choose to get off these bad actor platforms. The government is not the appropriate solution to this. Seeing how quickly Fox has declined, if the people /know/, the move to Gab, Parler, etc. en masse can happen very quickly too.
Going lower level, I have thought that if Trump or friends would produce a phone without an uncensored app store, you could really market it as "this phone doesn't spy on you or sell your data or censor you". Everyone wants this, but many aren't aware the current Apple/Google do in fact do some of these things, or that they have/could have a choice.
Actually, if you sum up the state by county, with the official registered voter percentages, and use the portion who are registered Republican+Libertarian, that's a partition, and Democrat+all parties including independent, that's a partition. You actually come up to within 2400 votes of the exact state total. I did the code for this a week or two ago.. As if all independents, 99.9%, went to Biden. It could be a total coincidence.
Added votes don't necessarily mean a hacked machine.
For example, if contrast/gamma/whatever on the scanner can be adjusted (or any other way the ballots would be unable to be determined by the computer -- extra markings, bleed through, etc?), the ballots are considered uncertain. In this case, the ballot images go into a separate directory, and an election official can then choose to manually assign them to a candidate.
The employee scanning the ballots in would be none the wiser, and it looks like they're just doing their job.
I believe Rudy discussed this in the Pennsylvania hearing, and it seems a very practical way to cheat.
The question I have here is, what is the raw data being reported?
If it reports total votes as well as percent, and TV and BV are derived fields, of course they have decimals in them (take TV/0.534 and BV/0.448, equals 3606490). Now, there's no reason to DO it this way, except to be able to play shenanigans, but 'machine is capable of shenanigans' doesn't prove they are actively doing them.
If the machine is directly reporting TV and BV, that's pretty solid.
So the strength of this depends entirely on which of these fields is raw data.
They had the original count, and then the re-tally which they requested not be reported. The difference gives them the number of ballots that must be destroyed to match in a real recount. Since the first two 'reported the same', give or take, the third one will too, because they've used the second to determine the evidence destruction needed. The shredder trucks would square with this theory...
There is a real question why they did their first recount but requested the new tally not be reported. Could it be that they now have a 'target number' for destruction etc. to match their original numbers? This would square with the existence of shredder trucks. Let's not be naive and assume the ballots themselves are unmodified, all present, third time around... Otherwise, why do it this way?
Risk Limiting Audit... As opposed to what they previously did, an actual partial recount.
They take a negligible number of votes, "randomly" chosen. Look at the article, "80% of the vote, they take 27 votes to verify". It's a sham.
I was thinking. Maybe the Sharpies do work and do scan... individually. But if you mark something else it spoils.
For example, suppose on one side of the ballot you have D/L/R. And on the exact opposite side, for a different race, you have D/R/L.
Then, if you mark D on both, the bleed through doesn't matter. But if you mark R or L, the bleed through would make races double-marked.
If you live in an area where they did this, look at the sample ballots online and see if they played ordering games (when you print it out). I'd be real curious.
It might be worthwhile to randomly sample very active users and offer them the option instead. If you do it the other way around, just on statistics, you may end up with certain ...interested... parties moderating.