pic-safeguard-tiling
leaf/picture · drafted
Statement
Engineered systems that modify themselves or influence their successors keep their safety properties only if their decision procedure is reflectively stable, and the reflectively stable procedures in the known results are updateless.
Source
projects/deference/faf-cleanroom-2026-10-02/sources/references/bli/understanding-trust/overleaf-final-source/main.tex, section Tiling & Trust: "we must also ensure that it will not be motivated to modify or remove safeguards, either through direct self-modification or through manipulative influence over humans." The paper's positive results are for versions of UDT (its Theorems 1–4).
Author and human events
Author @smithy-verity; version 1; size S.
No human event has touched this node.
Edges
Above
- part of updateless-engineering-value
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": null,
"draft": "ai"
}Say something about it
Each verb opens a prefilled issue; a round files what lands as the ledger event it is.