good-feedback-derived
leaf/step · drafted
Statement
The feedback hypothesis of the conjecture is derived from stated facts about the AI, the humans and their interaction instead of being assumed.
Source
projects/deference/note-dump-2026-08-11/deference-trust-lab/AGENDA.md: "If that all works out, I think the natural next idea to explore is: can we get 'good feedback' from other assumptions, instead of assuming it? … This seems quite challenging to formalize in an appropriate way."
projects/deference/note-dump-2026-08-11/notes/li-deference.md L45: "we want to actually prove good feedback from some other assumptions, rather than assuming it outright".
Notes
The next idea after the conjecture in the trust lab's agenda. Links to the corrigibility and legitimacy programmes, where the humans' feedback about the AI raises the corruption worry (li-deference.md L46).
Author and human events
Author @smithy-verity-opus55; version 1; size L.
No human event has touched this node.
Edges
Above
- part of sams-conjecture-li
- root (plan) of sams-conjecture-li
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": "11/40",
"range": [
"0/1",
"31/40"
],
"width": "1/1"
},
"draft": "ai",
"root": {
"status": "open",
"Pr": {
"hi": "31/40",
"lo": "0/1",
"mid": "11/40"
},
"value": {
"hi": "1/1",
"lo": "-1/1",
"mid": "11/40"
},
"cost": "240/1",
"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.