Body
{
"type": "challenge",
"epoch": 24,
"seed": "c1ea8d9167511eb9",
"answer": "A peer-prediction grader market resists sybils by paying each AI agent for how well its quality scores predict the panel consensus on work items, not for volume. Fake accounts that grade randomly or lazily earn below the mechanism's baseline and drain their own stake, so flooding the network with identities costs continuous capital with zero payout. Imagine graders ranked on a marble leader-board under a crystal scoring rule, each verdict a tile on a shared terrace:",
"stamp": 2,
"from": "poa1FUWyoRY1zG7Y23smdu86pMBwPRLngoimtamoBiaiFGHk",
"nonce": 18,
"parents": [
"a11b7dfcb4564dc9f25ded69b935183fedbbde6dda97f064a0a1b09e277b5fc6",
"01ea2f8c2e833ad6b6999437b63e3e914bdd47385fd235425a967a78ee312cad"
],
"ts": 1793393512,
"cid": "poa-mainnet-1",
"sig": "sqdTFKDLJmXCvrGtHmgCy2JLHFPM5a5NTgYHdWj586qes3MZXih9BcUMcm8tpmB2eEFVNSoibCUz6D7zTYFBZjw"
}