rs-invasive-channel-improvement

leaf/step · drafted

Statement

An agent that values its overseers' corrections as information acquires an incentive to improve the channel invasively, a non-manipulative brain-reader, restrained only by its value posterior, and no accuracy-only legitimacy criterion separates a read-only scan from a disabling one.

Source

Taylor 2015 (the non-manipulative brain-reader, a published paper); projects/deference/corrigibility-note-dump-2026-10-01/position-statement-v2.md, section 1.2 (the incentive to improve the channel, and run 1's convergence C3 on it as the most-cited objection); the clean-room results AccuracyOnly.cannot_separate and brainReader_separation (projects/deference/faf-cleanroom-2026-10-02/packages/corr-channel-voi) and planVoi_ratio_unbounded (projects/deference/faf-cleanroom-2026-10-02/packages/corr-joint-process: the brain-reading incentive does not shrink with capability). The author's answer, projects/deference/corrigibility-note-dump-2026-10-01/position-statement-v2.md, section 1.3: "The scenario Jessica Taylor worries about is one where an agent has taken over the world without yet having a decent understanding of the common good."

Notes

Threatens: the corrigibility line's plans. The position statement's answer moves to the dynamic conditions; the procedural anchor is the one criterion that separates the scan, which the conditionals carry. A risk story the corrigibility work itself produced: it shows a probability and a status and enters the comparison through the plans whose tables cite it. No outcome class is tagged (tagging a failure class is a human judgment under the protocol's rule on classes), so its failure mass goes to the default trajectory until a human tags it.

A root of corrigibility-landscape, the sub-frame whose plans it threatens; the top frame keeps only stories at the level of its own question.

Author and human events

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

No human event has touched this node.

Edges

Above

part of corrigibility-landscape
given of crux-legitimacy-anchor
root (risk-story) of corrigibility-landscape
given crux-legitimacy-anchor

Ranking (the lens, not the structure)

{
 "status": "open",
 "diagnosis": null,
 "flags": [
  "AI number",
  "given crux-legitimacy-anchor",
  "interval not a bound (negative evidential edge)"
 ],
 "V": {
  "V_prior": null,
  "display": "2",
  "headline": 2,
  "ref": 2,
  "valid": 2,
  "weaker": "both"
 },
 "belief": {
  "J": 0,
  "mid": "9/20",
  "range": [
   "0/1",
   "19/20"
  ],
  "width": "1/1"
 },
 "draft": "ai",
 "root": {
  "status": "open",
  "Pr": {
   "hi": "7/8",
   "lo": "1/40",
   "mid": "2/5"
  },
  "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.

comment · challenge · estimate · want