Here is nostr trying to solve one of the main problems that plagues it: how to deal with spam and find true signal in the feed.
Thomas Voegtlin, from the Electrum project, is proposing a way of attaching a proof of payment to Bitcoin miners as a pay-to-post mechanism for nostr.
Here we propose to filter spam by sending fees to the Bitcoin miners, and by attaching public proofs that payments have been made to Nostr events. This is a form of delegated proof-of-work, where purchased hashing power contributes to the security of the Bitcoin network.
These payments are made on behalf of nostr users by a notaries:
Nostr relies on two types of entities: clients and relays. Relays aggregate content uploaded by clients, and relay it to other clients. Users select content they like, and mirror it to other relays. Here we introduce a third type of participant: Notaries.
Nostr users may pay a notary a sum of bitcoin to commit a Merkle tree root hash to a bitcoin transaction that gets mined. This hash can be used to prove that a specified amount of bitcoin was paid on behalf of the notarized nostr note. The Merkle tree can have many leaves as proofs for many different notes.
Here's how they see it working:
- Users decide how much fee they want to see attached to Nostr events. Users pay notaries to create proofs-of-burn: a proof that an event ID was committed to the Bitcoin blockchain, and that a certain amount has been paid to the Bitcoin miners. The proof-of-burn is broadcast as a separate Nostr event.
- Users set a spam threshold: An amount of burnt bitcoins under which they do not want to see events posted by other users. This is akin to setting a price on one's attention. Users may relax this rule for pubkeys they follow.
- Relays set their own spam threshold, under which they will not relay events. Relays may temporarily store events that have not received enough proof of burn. Alternatively, Nostr events and their proof of burn could be relayed together, by implementing a form of package relaying.
- Users may pay for their own events, or for events posted by others. This implements an upvoting mechanism: It is possible to give increased visibility to someone else's content by attaching fees to it.
The authors are careful to note that this is not a trustless system:
Notarization is not trustless; a notary could accept payments and keep for themselves the amount they are supposed to burn. This could be mitigated using reputation and public proofs of payment, e.g. the preimage of a Lightning invoice that includes event ID, amount to burn and timestamp in its metadata.
I'm not sure that this has much advantage over a system where a nostr user must pay a relay for each individual note it wants the relay to transmit (do any notsr relays even do this? I know there are paid relays, but all the ones I have seen have seen have been subscription model, not pay to post).
The added complexity of this notary system doesn't seem to me to give you very much advantage. If you are going to trust the relay to actually "burn" the fees on your behalf, why not just trust the relay to do it?
Essentially, this is how Stacker News works: when you post, you send bitcoin to the territory owner and to the SN rewards pool. You trust that SN isn't secretly refunding these sats to the posters. It seems that a relay could function in a very similar way...but then they might as well just run their own instance of SN.