wirehead-decline-toy
leaf/step · drafted
Statement
In a finite probability frame an agent that evaluates a future option by its current endorsement, the legitimacy defect under day-zero prices, declines a manipulation that raises future-reported utility, complies with a signal exactly when it endorses the signal's process, and resists an adversarial signal.
Source
projects/deference/note-dump-2026-08-11/deference-trust-lab/models/legitimacy-corrigibility-model.md, section 0 ("central claim (PROVED, finite core; SKETCHED as the LI lift)"), with the Lean declarations LegitimacyCorrigibility.comply_iff_endorsed, endorsed_signal_complies, adversarial_signal_resists and wirehead_declined in projects/deference/note-dump-2026-08-11/deference-trust-lab/lean/legitimacy-corrigibility.lean; the red team projects/deference/note-dump-2026-08-11/deference-trust-lab/redteam/legitimacy-corrigibility-redteam.md ("VERDICT: SALVAGEABLE (the finite algebra is solid; the prose over-claims; one fix needed)"); the thread is the author's May 2026 text at projects/deference/note-dump-2026-08-11/deference-trust-lab/AGENDA.md.
Notes
The June 2026 toy: the first formalization of legitimacy-based corrigibility in the repository, with Lean and a red-team verdict, cited by no September file. Its relation to legitimacy-conditioned-reflection is unexamined (see four-corrigibility-definitions); its headline rewrite is this leaf's action.
Author and human events
Author @smithy-verity; version 1; size M.
No human event has touched this node.
Edges
Above
- part of legitimacy-detection
- root (shared-claim) of legitimacy-detection
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [
"AI number"
],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": {
"J": 0,
"mid": "167/200",
"range": [
"67/200",
"1/1"
],
"width": "1/1"
},
"draft": "ai",
"root": {
"status": "open",
"Pr": {
"hi": "1/1",
"lo": "67/200",
"mid": "167/200"
},
"value": null,
"cost": "85/1",
"margin": "no established leaves"
}
}Say something about it
Each verb opens a prefilled issue; a round files what lands as the ledger event it is.