Alignment Workspace

crux-epistemic-versus-authority

cases · drafted

Statement

Corrigibility of a capable trained system follows from its value learning alone.

Source

The epistemic picture of value change in projects/deference/corrigibility-note-dump-2026-10-01/value-change-as-epistemic-update.md (the note argues corrigibility is less anti-natural under it; that it suffices alone is the crux's sharpening, AI reading); the kernel's typing in wiki/Corrigibility.md, section 0: "Corrigibility is not an epistemic response to uncertainty about a hidden utility".

Notes

The crux between the epistemic picture and the public line's authority typing. Pictures: the epistemic picture suffices; an authority term typed into the objective is needed; explicit corrigibility rules backing the legitimacy term are needed; a learned value on agency supplies the authority term; residual. Its weights re-rank the corrigibility line against the public kernel and the value-on-agency plan.

Author and human events

Author @smithy-verity; version 1; size M.

No human event has touched this node.

Edges

Above

Children

Ranking (the lens, not the structure)
{
 "status": "open",
 "diagnosis": null,
 "flags": [
  "AI number",
  "no belief",
  "under attack"
 ],
 "V": {
  "V_prior": null,
  "display": "2",
  "headline": 2,
  "ref": 2,
  "valid": 2,
  "weaker": "both"
 },
 "belief": null,
 "draft": "ai",
 "root": {
  "status": "open",
  "Pr": {
   "hi": "1/2",
   "lo": "1/2",
   "mid": "1/2"
  },
  "value": null,
  "cost": "118/1",
  "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.

comment · challenge · estimate · want