Halseth is making an *actual* proposal for zk channel announcements in LN gossip! :
https://delvingbitcoin.org/t/zk-gossip-for-lightning-channel-announcements/1407/
This topic has fascinated me for the last couple years. LN has a kind of unique position to blaze the trail of creating pure cost based, fully private, anti Sybil/DOS of the type that every privacy network needs.
(Disclaimer: I don't believe PoW works, long term, for anything outside of bitcoin. Probably. )
What’s wrong with PoW?
Specialization, mainly. You want the cost per unit to be fixed for all users.
If the PoW scheme requires the user to commit to some user id in the hash, would that satisfy specialization?
Very interesting, and there is PoC already! Thanks for sharing :)
As you can see here:
https://github.com/halseth/output-zero/issues/10#issuecomment-2633912441
proving time without gpu acceleration is no good as of yet (15 mins on my kinda normal laptop). Let's see if optimization brings it down a ton.
Aut-ct proving times are 1-2 seconds. I think curve trees are a better solution for this, but it is extremely debatable.
