legit-li-self-correction-identity
leaf/step · drafted
Statement
For a single logical inductor the correction function toward its own future state is the identity and legitimacy-weighted Total Trust holds for every generable weighting, so a legitimacy event with content is anchored outside the inductor's own trust structure.
Source
projects/deference/faf-cleanroom-2026-10-02/sources/corrigibility/workflow-2026-09-14b/develop/legitimacy-general-final.md, Statement 10 and its summary item (5): the legitimacy-weighted Total Trust statement at the LI register with a generable weighting, checked against the logical induction paper's conditional expectation theorem. The paper theorem is the conditional-expectation theorem the note calls ccee, in Garrabrant et al., "Logical Induction" (arXiv:1609.03543).
Notes
Checked at the note's register (exact arithmetic over the paper's display), not in Lean; the port is task-port-legit-push-empty. It is the result the correction-function question was missing: toward the self the correction is forced to the identity, so the learnable part lives on the anchor. The miner filed the same content as the claim that legitimacy-conditioned reflection is empty toward a push; the two are one statement.
Author and human events
Author @smithy-verity; version 1; size M.
No human event has touched this node.
Edges
Above
- held by legitimacy
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [
"AI number",
"held by legitimacy"
],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": {
"J": 0,
"mid": "33/40",
"range": [
"13/40",
"1/1"
],
"width": "1/1"
},
"draft": "ai"
}Say something about it
Each verb opens a prefilled issue; a round files what lands as the ledger event it is.