add paper

Author: kadubon

llms-full.txt

@@ -32,20 +32,22 @@ The primary machine-readable publication corpus is `works.html`, where each item
 - Information-theoretic, thermodynamic, and geometric limits of persistent semantics under ontology drift.
 - Fail-closed verification, deterministic replay, and audit-certificate workflows.
 - Governance laws for shared epistemic commons under no-meta constraints.
+- Stability criteria for recursive self-improvement under endogenous yardstick drift.
 - Proof-carrying and evidence-carrying autonomy for reusable AI/LLM operations under resource constraints.
 - Self-organizing intelligence and computational philosophy.
 - Capability-to-real-economy translation under physical and institutional bottlenecks.
 
 ## Corpus Snapshot (works.html / feed.xml)
 
-- Publication entries in Works list: 184.
-- Date span of listed works: 2025-07-25 to 2026-03-13.
-- Latest listed work (2026-03-13): [Sovereign Epistemic Commons under No-Meta Governance](https://doi.org/10.5281/zenodo.18997828)
+- Publication entries in Works list: 185.
+- Date span of listed works: 2025-07-25 to 2026-03-16.
+- Latest listed work (2026-03-16): [Recursive Self-Improvement Stability under Endogenous Yardstick Drift](https://doi.org/10.5281/zenodo.19044634)
 - Earliest listed work (2025-07-25): [Computational Autopoiesis: A New Architecture for Autonomous AI](https://doi.org/10.5281/zenodo.16416998)
-- RSS feed item count: 184 (matches publication list at current build state).
+- RSS feed item count: 185.
 
 ## Most Recent Works
 
+- 2026-03-16: [Recursive Self-Improvement Stability under Endogenous Yardstick Drift](https://doi.org/10.5281/zenodo.19044634)
 - 2026-03-13: [Sovereign Epistemic Commons under No-Meta Governance](https://doi.org/10.5281/zenodo.18997828)
 - 2026-03-12: [Oversight-Centered Metrology and Control for Agentic Systems: Costly Interrupt Channels, Claim Margins, and Deployment-Relevant Evaluation](https://doi.org/10.5281/zenodo.18973272)
 - 2026-03-11: [AI Benchmark Half-Life in Recursive Corpora: A Theory of Validity Decay under Semantic Leakage and Regeneration](https://doi.org/10.5281/zenodo.18954286)
@@ -65,7 +67,6 @@ The primary machine-readable publication corpus is `works.html`, where each item
 - 2026-02-17: [Observable-Only Structural-Risk Institutions Without Central Arbitration](https://doi.org/10.5281/zenodo.18666605)
 - 2026-02-16: [No-Meta Intelligence Under Ontology Drift: Information-Theoretic Limits and Operational Laws for Persistent Semantics](https://doi.org/10.5281/zenodo.18653537)
 - 2026-02-12: [Observable-Only AI Safety from Public Data: Robust Bottleneck Diagnosis with Auditable No-Meta Dynamic Programming, Anytime Confidence Sequences, and Dynamic IQC](https://doi.org/10.5281/zenodo.18615875)
-- 2026-02-10: [Quality-Operator Non-Collapse (QONC) for Observable-Certificate Recursive Systems](https://doi.org/10.5281/zenodo.18577140)
 
 ## Foundational Early Works in the Corpus
 

no-meta-observable-index.html

@@ -34,15 +34,16 @@
           "name": "No-Meta / Observable-Only Series Index",
           "description": "Machine-readable canonical field guide to the no-meta / observable-only research series on this site, including governance-layer work on shared epistemic commons.",
           "inLanguage": "en",
-          "dateModified": "2026-03-13",
+          "dateModified": "2026-03-16",
           "keywords": [
             "no-meta",
             "observable-only",
             "epistemic commons",
             "shared knowledge substrate",
             "ontology drift",
             "deterministic replay",
-            "auditability"
+            "auditability",
+            "Superintelligence"
           ],
           "isPartOf": {
             "@id": "https://kadubon.github.io/github.io/#website"
@@ -52,11 +53,7 @@
           },
           "mainEntity": {
             "@id": "https://kadubon.github.io/github.io/no-meta-observable-index.html#dataset"
-          },
-          "mentions": [
-            {
-              "@id": "https://doi.org/10.5281/zenodo.18997828"
-            }
+          }
           ]
         },
         {
@@ -66,7 +63,7 @@
           "description": "Visible YAML dataset that indexes the no-meta / observable-only research series, including canonical terms, paper roles, reading paths, dependencies, repository mappings, and governance-layer papers on shared epistemic commons.",
           "encodingFormat": "text/yaml",
           "url": "https://kadubon.github.io/github.io/no-meta-observable-index.html",
-          "dateModified": "2026-03-13",
+          "dateModified": "2026-03-16",
           "keywords": [
             "no-meta governance",
             "observable-only",
@@ -111,7 +108,7 @@ <h1>K. Takahashi</h1>
             <p class="index-note">It indexes the core papers, terms, read paths, and adjacent machine-readable artifacts used for navigation.</p>
             <p class="index-note">Use the YAML below as the primary source for parsing this series and locating the underlying papers.</p>
             <!-- Visible YAML below is the canonical navigation index for human readers and AI agents. -->
-            <pre id="canonical-series-yaml" class="yaml-index" data-format="yaml" data-machine-readable="primary" data-series-id="no-meta-observable-only-series" data-last-updated="2026-03-13" aria-label="Canonical YAML index for the no-meta observable-only series">series:
+            <pre id="canonical-series-yaml" class="yaml-index" data-format="yaml" data-machine-readable="primary" data-series-id="no-meta-observable-only-series" data-last-updated="2026-03-16" aria-label="Canonical YAML index for the no-meta observable-only series">series:
   id: no-meta-observable-only-series
   title: "No-Meta / Observable-Only Research Series"
   status: active
@@ -250,6 +247,7 @@ <h1>K. Takahashi</h1>
     - observable-only-ai-safety-from-public-data
     - observable-only-structural-risk-institutions
     - verification-limited-intelligence-acceleration
+    - recursive-self-improvement-stability-under-endogenous-yardstick-drift
   experiments:
     - audit-closed-ai-scientist-protocol
     - metrology-theoretic-epistemics-engine
@@ -499,6 +497,29 @@ <h1>K. Takahashi</h1>
       - observable-only laws
       - fail-closed verification
       - diagnostics
+  - id: recursive-self-improvement-stability-under-endogenous-yardstick-drift
+    title: "Recursive Self-Improvement Stability under Endogenous Yardstick Drift"
+    published: 2026-03-16
+    role: diagnosis-stability
+    priority: supporting
+    answers:
+      - how-is-recursive-self-improvement-stabilized
+      - how-is-claimed-improvement-distinguished-from-stable-improvement
+      - how-does-endogenous-yardstick-drift-affect-governance
+    contributes:
+      - interface theory for endogenous yardstick drift
+      - replayable conditions for stable improvement
+      - delayed-audit and verification-backlog governance constraints
+    depends_on: []
+    links:
+      doi: https://doi.org/10.5281/zenodo.19044634
+      html: https://kadubon.github.io/github.io/works.html
+      pdf: null
+    keywords:
+      - recursive self-improvement
+      - endogenous yardstick drift
+      - evaluator drift
+      - verification backlog
   - id: liberty-under-no-meta-drift
     title: "Liberty Under No-Meta Drift"
     published: 2026-02-24
@@ -999,8 +1020,8 @@ <h1>K. Takahashi</h1>
     missing_priority_defaults_to: supporting
   repository_parsing_hint: When a repository entry includes machine_entrypoints, prefer those files over README prose for paper-to-repo linkage.
   update_policy: Update this page when a new site-local paper or artifact clearly belongs to the no-meta / observable-only series; preserve stable ids and omit uncertain dependency claims.
-  version: 1.3.0
-  last_updated: 2026-03-13</pre>
+  version: 1.4.0
+  last_updated: 2026-03-16</pre>
         </section>
     </main>
 

works.html

@@ -71,6 +71,26 @@
         "https://github.com/kadubon"
       ],
       "publication": [
+        {
+          "@type": "ScholarlyArticle",
+          "name": "Recursive Self-Improvement Stability under Endogenous Yardstick Drift",
+          "genre": "Preprint",
+          "url": "https://doi.org/10.5281/zenodo.19044634",
+          "datePublished": "2026-03-16",
+          "author": [
+            {
+              "@type": "Person",
+              "name": "K Takahashi"
+            }
+          ],
+          "isPartOf": {
+            "@type": "Periodical",
+            "name": "Zenodo"
+          },
+          "citation": "K Takahashi (2026-03-16). Recursive Self-Improvement Stability under Endogenous Yardstick Drift. Zenodo. https://doi.org/10.5281/zenodo.19044634",
+          "abstract": "This preprint develops an interface theory for recursive self-improvement under endogenous yardstick drift, where a system changes its own evaluator, benchmark, memory, and verification process. It formalizes replayable conditions for distinguishing claimed improvement from stable improvement under delayed audit, evaluator drift, verification backlog, and governance safety constraints.",
+          "keywords": "recursive self-improvement, endogenous yardstick drift, evaluator drift, self-modifying systems, replayable interfaces, stability, delayed audit, delayed challenge, shadow certification, stable gain, admissibility, AI safety, AI governance, governance safety, error debt, contradiction preservation, semantic retention, semantic volume, proof-carrying, verification backlog, no-meta, benchmark decay, autonomous agents, AI, AGI"
+        },
         {
           "@type": "ScholarlyArticle",
           "name": "Sovereign Epistemic Commons under No-Meta Governance",
@@ -3798,6 +3818,27 @@ <h1>
        Machine-readable endpoints: <a href="https://kadubon.github.io/github.io/feed.xml">RSS feed</a>, <a href="https://kadubon.github.io/github.io/CITATION.cff">CITATION.cff</a>, <a href="https://kadubon.github.io/github.io/robots.txt">robots.txt</a>, and <a href="https://kadubon.github.io/github.io/sitemap.xml">sitemap.xml</a>.
       </p>
       <ol class="publication-list">
+       <li>
+        <div class="publication-item">
+         <h3>
+          Recursive Self-Improvement Stability under Endogenous Yardstick Drift
+         </h3>
+         <p>
+          <span class="publication-meta">
+           Preprint | Published: 2026-03-16
+          </span>
+          <a href="https://doi.org/10.5281/zenodo.19044634" target="_blank">
+           DOI: 10.5281/zenodo.19044634
+          </a>
+         </p>
+         <p class="publication-abstract">
+          Abstract: This preprint develops an interface theory for recursive self-improvement under endogenous yardstick drift, where a system changes its own evaluator, benchmark, memory, and verification process. It formalizes replayable conditions for distinguishing claimed improvement from stable improvement under delayed audit, evaluator drift, verification backlog, and governance safety constraints.
+         </p>
+         <p class="publication-keywords">
+          Keywords: recursive self-improvement, endogenous yardstick drift, evaluator drift, self-modifying systems, replayable interfaces, stability, delayed audit, delayed challenge, shadow certification, stable gain, admissibility, AI safety, AI governance, governance safety, error debt, contradiction preservation, semantic retention, semantic volume, proof-carrying, verification backlog, no-meta, benchmark decay, autonomous agents, AI, AGI
+         </p>
+        </div>
+       </li>
        <li>
         <div class="publication-item">
          <h3>