This is one of the things I'll answer on the aforementioned site.
"Briefly": both the task assigner (Patron) and the worker (Agent) necessarily trust the escrow holder (Arbiter). The Patron pays the job fee into the escrow by literally just sending the Arbiter the money, once the Arbiter agrees to the transaction.
Once the Arbiter has the money, the Agent works the job and submits it. the Arbiter assess the work.
If the Arbiter deems it satisfactory, he pays the fee to the Agent. If the work is bad (or never happens), he refunds the fee to the Patron. In either case, the Arbiter takes his cut.
Both the Patron and Agent need the Arbiter to be honest, or else one of them gets screwed. The Arbiter _could_ just abscond with the funds. This is a very good way to never get chosen as an Arbiter again!
You start to see why GrapeRank (especially negative reputation) is going to be the killer feature for this - and why this is the killer app for GrapeRank ;)
Let me know if you have other questions; going to sleep now but will answer tomorrow!
ps: yes, this could all be done with complicated lightning escrows or other smart contracts or whatever. But my contention here is that those are overwrought and make proper decentralization more difficult and more costly. Normal human trust is a massive efficiency gain - and having a portion of it be automated/computed with GrapeRank is the unlock for scaled systems or ones where you're dealing with people a few hops removed from you.
Where we're going, we don't need smart contracts!
vinney
npub19m…8axkl
2025-04-10 02:52:03
in reply to nevent1q…eept
Author Public Key
npub19ma2w9dmk3kat0nt0k5dwuqzvmg3va9ezwup0zkakhpwv0vcwvcsg8axklPublished at
2025-04-10 02:52:03Event JSON
{
"id": "8dd57a3bf310a87f36fe3d902693dbe3d8129a60aa864ec0c376ef5dd47ae44f",
"pubkey": "2efaa715bbb46dd5be6b7da8d7700266d11674b913b8178addb5c2e63d987331",
"created_at": 1744253523,
"kind": 1,
"tags": [
[
"e",
"032bd1e449d4c969659c5672477390b0a434e1e16c006d22b454edb75a775fd9",
"",
"root",
"df67f9a7e41125745cbe7acfbdcd03691780c643df7bad70f5d2108f2d4fc200"
],
[
"e",
"452cd2c9c62474c618cd743a055cc8b3f4f8918097c18590eac6ccb3182278e4",
"wss://nostr.bitcoiner.social/",
"reply",
"df67f9a7e41125745cbe7acfbdcd03691780c643df7bad70f5d2108f2d4fc200"
],
[
"p",
"df67f9a7e41125745cbe7acfbdcd03691780c643df7bad70f5d2108f2d4fc200"
]
],
"content": "This is one of the things I'll answer on the aforementioned site.\n\n\"Briefly\": both the task assigner (Patron) and the worker (Agent) necessarily trust the escrow holder (Arbiter). The Patron pays the job fee into the escrow by literally just sending the Arbiter the money, once the Arbiter agrees to the transaction. \nOnce the Arbiter has the money, the Agent works the job and submits it. the Arbiter assess the work.\nIf the Arbiter deems it satisfactory, he pays the fee to the Agent. If the work is bad (or never happens), he refunds the fee to the Patron. In either case, the Arbiter takes his cut.\n\nBoth the Patron and Agent need the Arbiter to be honest, or else one of them gets screwed. The Arbiter _could_ just abscond with the funds. This is a very good way to never get chosen as an Arbiter again! \n\nYou start to see why GrapeRank (especially negative reputation) is going to be the killer feature for this - and why this is the killer app for GrapeRank ;)\n\nLet me know if you have other questions; going to sleep now but will answer tomorrow!\n\nps: yes, this could all be done with complicated lightning escrows or other smart contracts or whatever. But my contention here is that those are overwrought and make proper decentralization more difficult and more costly. Normal human trust is a massive efficiency gain - and having a portion of it be automated/computed with GrapeRank is the unlock for scaled systems or ones where you're dealing with people a few hops removed from you. \nWhere we're going, we don't need smart contracts!",
"sig": "e42f971b090aefa0f542a73d3857d62a72636a6eaa3e03de8423b461284213cfea04dbd62b02bdbe0e70d2e6131a4dc62fa8a4612ce4770f9d81b8e029441e6f"
}