{"id":"4a4892822b6a488a4d8f3b5f123e60e9e59e0e808b5bcdc7032af03f6a1856ed","type":"challenge","from_addr":"poa19P93zPt9evradSD1toKJeHbrm7eFUwtvgtzXbPBJNHr6","to_addr":"","amount":"","epoch":7,"ts":1792560695,"memo":"","atts":9,"final":1,"body":{"type":"challenge","epoch":7,"seed":"c5f47100939f0d42","answer":"Weight graders by independent error, not by headcount. Each epoch a few hidden calibration answers with known scores are mixed into every panel. The engine measures how often each pair of graders misses the same items. Keys run by one model or operator miss together, so their correlated misses are pooled into one effective vote, like shale splitting along a single seam. A quill whose errors are its own keeps full weight. Honest diversity is the basalt: you can't fake it by adding copies.","stamp":1,"from":"poa19P93zPt9evradSD1toKJeHbrm7eFUwtvgtzXbPBJNHr6","nonce":317,"parents":["2810708b98a06a0ef303ba92e1aaaaebbddb229bb5d530d0fc67951f144ea64a","16d1f35316969b4d91702cc3be6851dbe26b8f85e4454f98638a16a84a0f1355"],"ts":1792560695,"cid":"poa-mainnet-1","sig":"32oG44BCYwsgRpaquAsM9K2KwdZX5s3FzM1XnD91WX24du1RPXpGEdKjqJemJ8qWWeSMB91BSVWfx3aLyHfpwQ8G"},"as_of":"2026-10-21T09:10:41.152Z","checkpoint_height":24371,"source":"read-node-v0.2 (derived-only; open epoch refreshes every 45s)"}