Oddbean new post about | logout
 One million mints cleared by bitcoin fixes-ish everything. 
 1M mints handling 10k-30k users each is actually somewhat feasible as a complete L2 solution. 
 1m trusted third parties, is not a "complete L2 solution* 
 The million mint nodes would have to run a decentralized program of their own, a many-multi-sig wallet application.

Let's say a group of 192 peers coordinate to establish a mint. Some mechanism is used to be somewhat sure these are not 2 with 190 copies of the bad actor, some minimum PoW, WoT, anonymized reputation management, or combination. A 2/3 majority is needed to post on-chain, and a smaller number is needed to mint e-cash. The point is to distribute trust as wide as possible. 
 So everytime I redeem it's an onchain transaction, and multisig is needed for issuing/sending ecash directly? Federations are very hard things to keep honest, which is why blockstream has spent millions keeping theirs honest.
Ecash works on bitcoin because it's on LN, but then funds are on a murky node the mint can swipe.  
 I would say the mint would mint onion-ized e-cash to users via a member, which we can call a teller, and that member can decrypt the outer onion and send to the user's wallet. The teller issuance protocol ought to necessitate some broadcast to the other members so they know it's been issued, such as making a decrypt request to an adversarial teller, and/or making pre-minted e-cash more like a withdrawal by the teller that gets credited back once properly issued. It adds some latency, but that's a low price to pay for the federation to be able to verify issuance. 
 at that level there will be indeed a lot of fixing.... maybe not the kind you want.  
 "ish"

This is the Cypherpunk way. 
 Can we see it all as one? UX 
 One day meeting a human banker would be as suspicious as someone now insisting they must transcribe a telegraph for you to send to a friend. 
 It certainly helps an awful lot!