fidelity bond is a stupid concept. proof of stake rebranding.
Who has more money to lock up: bad actor criminal chain analysis feds? or plebs who just want to mix for other plebs?
You have better idea how to fight sybill attacks (which is generic problem of any P2P network)? Without FBs bad rich actors can spin up thousands of makers and have big probability to be all "counterparties" in every single coinjoin.
reply
An entirely valid point to raise, but the answer isn't trivial at all.
It's rather analogous to the design of Bitcoin: who has more resources to apply to mining, the aggressors or the 'defenders' (which can include people just wanting to earn value from the system, not only idealists)?
One set is a lot more open than the other, and that set has a positive financial incentive. Also coinjoin has an additional quirk: more than one "agressor" can actually interfere with each other, if each one individually is trying to deanon all the coinjoins.
Anyway I think you can list several reasons why a fidelity bond approach to making Sybil attacks hard is flawed; what's a lot harder, is to come up with any viable alternative without requiring identities.
reply
Yep, this is how I came up with the idea of paying for sessions with AMP sends, and a string of ideas I'm not completely satisfied with relating to clients giving delayed feedback via selective sharing and once a user has found a few score good peers they can just stick with them and tippie toe in the new fields once in a while to try pick up a new one as inevitably some will go offline.
I think a protocol a bit like LN's scheme of PSBTs and such like things would make it possible to do them with lower counterparty risk than a kind of security bond system by leveraging game theory and cryptography.
Not something I'm gonna dwell on much right at the moment, I have plenty of puzzles to unravel before I can move on to something new, but that's my take on it anyhow.
reply
proof of stake has nothing to do with a fidelity bond
reply