> But if you can create a new voter account as easily as you can create a new Gmail account
This is generally described as a Sybil attack [1].
> you just switch to a new account to lower the price of your vote. ... Or if a really rich voter with a high n-value pays someone with a low-n value to make a vote on their behalf
AFAICT the article discusses these problems:
> Perhaps the biggest challenge to consider with this concept of quadratic payments is the practical implementation issue of identity and bribery/collusion. ... Fortunately, there are technological means that can help, combining together zero-knowledge proofs, encryption and other cryptographic technologies to achieve the precise desired set of privacy and verifiability properties.
Fortunately, there are technological means that can help, combining together zero-knowledge proofs, encryption and other cryptographic technologies to achieve the precise desired set of privacy and verifiability properties.
This is placing too much trust in what is incomprehensible mathematics for 99.999% of the population. Even for those who understand it, they won't have any convincing way of being assured that 1) the maths is perfect and there are no flaws and 2) the voting technology in use is perfectly implemented.
On the other hand, the 'nineteenth century technology' of secret ballots is both obvious and simple. It's a shame that people like Vitalik are so dismissive of it.
I'm a developer with a strong math background, very security-minded, with experience using cryptocurrencies, pgp/gpg, privacy tools, etc.
I still would not be able to evaluate any sort of crypto program and certify its integrity, nor find any (save for the most eggregious) flaws. Even then, I have brought up questions in code reviews of "this seems to be flawed" only to learn "technically yes, but we have to do such gymnastics because of constraints xyz".
I, too, was big on the "blockchain is gonna make e-voting possible! Ra ra ra!" hype train for a while, but eventually wisened up to my own naivety.
I think paper is great because it's both proof of work (it had to be printed/written/modified) and proof of stake (e,g. I possess it and I can attest to its chain of custody), with zero lines of code!
SHA1 "is incomprehensible mathematics for 99.999% of the population", yet we do use it to great success. Same with public key cryptography. It doesn't seem to bother people that can't implement ssh/https from top of their heads, right?
> On the other hand, the 'nineteenth century technology' of secret ballots is both obvious and simple. It's a shame that people like Vitalik are so dismissive of it.
You can't write smart contracts based on secret ballots. The dream (well, one dream) is to have governance based on an ecosystem of programs that can vote perhaps hundreds of times a second. A lot of democratic procedure is built to work around the limitations of physical voting; this could qualitatively change the way decision-making works.
This is generally described as a Sybil attack [1].
> you just switch to a new account to lower the price of your vote. ... Or if a really rich voter with a high n-value pays someone with a low-n value to make a vote on their behalf
AFAICT the article discusses these problems:
> Perhaps the biggest challenge to consider with this concept of quadratic payments is the practical implementation issue of identity and bribery/collusion. ... Fortunately, there are technological means that can help, combining together zero-knowledge proofs, encryption and other cryptographic technologies to achieve the precise desired set of privacy and verifiability properties.
[1] https://en.wikipedia.org/wiki/Sybil_attack