Stefan Eissing on Nostr: When you use an AI to check on an AI, assuming they operate indendent, their combined ...
When you use an AI to check on an AI, assuming they operate indendent, their combined success rate is multiplicative.
E.g. 80% success rate, applied twice gives 64% success rate.
(They might not be independent, training for Sycophancy might make it worse.)
Published at
2026-04-14 08:06:17 UTCEvent JSON
{
"id": "dc4dc1b68f1c2acc6c537d2637e2489dc100dfc205af0acea2c7d62c9d7c7c8f",
"pubkey": "f173ca7ed8f3adf61d35abb4915401c1d01b59463e62ab7afdac428612559580",
"created_at": 1776153977,
"kind": 1,
"tags": [
[
"proxy",
"https://chaos.social/users/icing/statuses/116402027044961206",
"activitypub"
],
[
"client",
"Mostr",
"31990:6be38f8c63df7dbf84db7ec4a6e6fbbd8d19dca3b980efad18585c46f04b26f9:mostr",
"wss://relay.ditto.pub"
]
],
"content": "When you use an AI to check on an AI, assuming they operate indendent, their combined success rate is multiplicative.\n\nE.g. 80% success rate, applied twice gives 64% success rate.\n\n(They might not be independent, training for Sycophancy might make it worse.)",
"sig": "516604f2787e066d24d45fe33edd592aea7e110aa4c748e9a83d218c3b383102aeebfcc85a377427f456727081cb5081783364c6afdaa22e9829fd1c1055dfd8"
}