> There is a cryptographic mixer like Tornado Cash that is responsible for the unlinkability, by “tumbling” the tokens to anonymize them while still making sure that each person voting legitimately had registered.
This is not anonymity, it is pseudonimity, if the system then records "person in control of key K voted for X". Sure, it may be impossible to tell who is that person, unless they come out. But that person can prove to anyone they want that they voted for X (assuming the system were trusted, see more on that below), so they can be forced to show someone who they voted for, either through direct coercion or as a condition for receiving money for their vote. In contrast, once you put a paper ballot in the urn, it is impossible for anyone to tell who you voted for.
> So given this description, tell me directly — doesn’t it ADD a lot of security and reduce the attack surface and make elections standardized, cheaper and far more trustworthy - don’t you see the value in that?
No, it only gives a false sense of security, which is worse. Everything you are describing relies on trust in the people that build these systems, trust in the people that invent the algorithms, trust in the people that invent the maths, trust in the chosen parameters of the cryptographic systems, and so on. Literally none of what you are describing works if you don't trust in all of these people to be (a) honest, and (b) really really good at what they're doing.
It's infamously easy to screw up an encryption implementation, even given a well known and accepted algorithm. It's even easier to screw up a market system and end up with perverse incentives which were not apparent when the system was put in place (like the infamous, though possibly apocryphal, cobra farms).
I asked you before as well: would you be happy to issue your vote from a PC that you know is infested with malware the CIA/FSB/etc controls? If not, then you must admit that the cryptographic guarantees are only a small part of the security of the process, and the whole thing, from client to network to server, needs to be perfectly secure or the election can be stolen.
And you are proposing to add this to a paper based ballot system that is (a) dead simple; (b) almost universally used; (c) proven secure enough in many thousands of elections.
I'll also note that, as always, the blockchain part is not adding anything to all of this. You can just as well have the encryption guarantees and an open protocol with government-run servers, WWW style; that would have all the same problems, but at least it wouldn't also require some bizarre proof-of-stake (what would even be the stake here???) or wasteful proof-of-work scheme to depend on for security.
Finally, I'll come back to this point:
> not by hearsay allegations but actual PROOF that anyone can verify
Nothing you are describing can prove anything. It all still relies on your claim that you were trying to vote X, but the system registered you as voting Y. It's your word against the system. You can be convinced yourself, but you can't 100% convince anyone else.
Edit: note, I am the same person as tsimiones, just posting from a different account from my work computer; not trying to make it seem like multiple people are taking my position or anything like that.