v1.3.3 - Recent Finite-Certificate Reading Map

About this release

v1.3.3 updates the GitHub Pages research hub for K. Takahashi with a clearer Research Map, publication-index context, and related OSS navigation.

The main purpose of this release is to make the recent theory stack easier to read by dependency, not just by publication date. The release explains the stack with both short theory labels and general search-friendly terms: observable evidence, finite certificates, runtime verification, digital continuity, consent and contact, capability routing, resource-aware deployment, and substrate planning.

What changed

• Added Recent Certificate-Interface Chain to the Research Map.
• Added Recent Finite-Certificate Stack as a recommended reading route.
• Made CSST, RCT, ACT, CBCT, DCT, OSCT, ECPT, and TRC explicit in the map, with BIT and SQOT included as bridging papers.
• Placed the recent theory group into the major Research Map pillars:
  • Verification / Auditability: RCT, BIT, SQOT, ECPT
  • Long-Running Agents: DCT, OSCT, CBCT
  • Transfer / Deployment / Governance: TRC, ACT, CSST
• Added recent_certificate_stack to the Structured Metadata Appendix, including title, short_id, role, DOI, and Works URL for each paper.
• Added kadubon/percolation-inversion-compiler to Related OSS and Reference Implementations.
• Described percolation-inversion-compiler as an AI agent output checker and workflow verification toolkit. It turns agent text, pull requests, external inputs, and messages into auditable JSON reports with evidence routing, residual ledgers, provenance, and safe reuse checks.
• Works includes the recent DOI-linked entries and software metadata updates.

Suggested reading path

The recommended dependency order is:

1. RCT: Start with how observable evidence can support claims about contact with reality.
2. SQOT: Then read how attention, verification cost, queues, latency, and diagnostic reserve constrain operational systems.
3. BIT: Read how bottlenecks are identified and represented with finite witnesses.
4. ECPT: Read how execution-available capability paths and target-basin claims are represented as finite certificates.
5. TRC: Connect capability paths to physical resources, tolerance ledgers, and future option preservation.
6. OSCT: Read terminal-boundary concepts such as termination, non-resurrection, complete extinction, and liberation certificates.
7. DCT: Read dormant continuity, wake authorization, recovery, handoff, and diagnostic routing.
8. CBCT: Read consent, contact, copy, fork, merge, and reactivation boundaries.
9. ACT: Read whether assistance preserves future agency, recovery, contestability, and independence.
10. CSST: End with substrate, embodiment, clock, copy, memory, lifecycle, and boundary-transition planning.

OSS positioning

percolation-inversion-compiler is listed as an implementation-facing companion in the Research Map. It is useful for runtime verification, finite certificates, proof obligations, JSON schemas, provenance manifests, and workflow audit trails.

It is not a proof of real ASI, hidden physical outcomes, unrestricted self-improvement, consciousness, personhood, legal standing, or complete AI safety.

Boundaries

This release does not claim actual ASI achievement, complete AI safety, or proof of consciousness, personhood, moral patienthood, or legal standing. The theories and OSS entries are presented as a reading map and certificate-interface stack for finite evidence, declared scope, checkable records, and explicit residuals.

Validation

Before release, the following checks were confirmed:

• research-map.html JSON-LD parsed successfully.
• research-map.html structured YAML parsed successfully.
• percolation-inversion-compiler appears in JSON-LD significant links, visible Related OSS, the Implementation/OSS reader path, and YAML related_oss.
• git diff --check passed for research-map.html.
• Works count remains 222 entries: 220 publications + 2 software releases.
• Release target is the current main HEAD after the OSS addition.

v1.3.2 - Canonical Research Hub Readability Polish

What this release is

This is a small readability and consistency release for K. Takahashi's GitHub Pages research hub. The site is a public entry point for DOI-linked research on auditable autonomous intelligence, no-meta / observable-only governance, Constraint Generative Theory (CGT), long-running agents, claim certification, and related OSS experiments.

This release does not redesign the site. It keeps the existing structure, canonical URLs, DOI links, structured metadata, support/donation section, and repository links intact.

For first-time visitors

Start with these pages:

• Home: a compact overview of the research program and support options.
• Research Map: the site-wide map showing how the main research clusters relate.
• Works: the complete DOI-linked publication list.
• No-Meta / Observable-Only Index: a field guide for governance, auditability, fail-closed verification, and observable evidence.
• CGT Index: a field guide for claims, reports, support conditions, scientific availability, and abstention.

What changed in v1.3.2

• Added DOI links and Works-page anchors to the No-Meta / Observable-Only Reading Paths, so each suggested paper route now leads directly to citable records.
• Made repository cards in the No-Meta guide easier to scan by separating each "Use this for" and "Boundary" statement.
• Reduced repeated wording in the CGT Series Map introduction while preserving the shared CGT DOI rule.
• Polished llms.txt and llms-full.txt so they read as neutral plain-text site guides rather than search-engine-specific instructions.
• Improved the Works page orientation links to Research Map, No-Meta Index, and CGT Index.
• Updated CITATION.cff wording to reflect the current research stack.

Boundaries

The site does not claim to solve AI alignment, guarantee AGI safety, or provide complete production implementations. Repository links are reference experiments and partial companions. Deterministic replay supports auditability, but it is not a proof of unrestricted truth. CGT is a viewpoint-oriented theory for reading claim/support relations; it does not replace existing scientific theories.

Validation

Before release, the following checks passed:

• HTML IDs checked for duplicates.
• JSON-LD parsed successfully.
• YAML blocks parsed successfully.
• feed.xml and sitemap.xml parsed as XML.
• CITATION.cff parsed successfully.
• Research support manifest parsed successfully.
• Wallet addresses and accepted asset/network rules remained unchanged.
• Works count confirmed at 209 entries: 208 publications + 1 software release.
• Latest Works entry confirmed as Constraint Generative Theory: Typed Constraint Effects and Scientific Availability.
• No-Meta Reading Paths DOI links confirmed against Works anchors.

v1.3.1 Added a Site-Wide Research Map and Theory Stack Overview

This release adds a new Research Map page to the site as a site-wide synthesis layer above the existing cluster pages and full works index.

The new page is designed to make the overall structure of the research program easier to understand for first-time readers, engineers, policy readers, and AI agents. It explains the main theory pillars, their directional dependencies, recommended reading paths by audience, and the canonical entry points for machine-readable navigation.

The page is not a new theory paper and not a replacement for the original cluster pages or publication index. Its purpose is narrower: to show how the major research areas fit together across foundations, semantics, verification, long-running agents, transfer, deployment, governance, and public accountability.

Key additions in this release include:

• a site-wide theory stack overview
• an explicit dependency map across major research pillars
• recommended reading paths for different audiences
• a visible machine-readable YAML summary embedded in the page
• improved navigation to canonical entry points across the site

This update is intended to improve global interpretability of the research program while preserving the site’s static, crawlable, and machine-readable design.

v1.3.0 Research Hub Update: Thematic Landing Pages, Machine-Readable Navigation, and Discoverability Improvements

This release improves the GitHub Pages research hub for both human readers and AI crawlers.

Main updates:

• Added topic-oriented landing pages across major research clusters
• Improved homepage navigation and machine-readable entry points
• Strengthened structured discoverability through works metadata, RSS, sitemap, robots, and citation files

The result is a clearer research map: readers and automated systems can now move more directly from broad topics to the relevant papers, instead of relying only on the chronological works list.

Site:
https://kadubon.github.io/github.io/

v1.2.2 Concept Entry Page Release - Self-Concealing Information and Observer-Modifying Dynamics

This release fixes a stable public entry point for the concept of self-concealing information and observer-modifying dynamics.

The new page is designed as a concept entry page, not as a duplicate of the paper itself. Its role is to explain the conceptual contribution in a form that is accessible to general readers, discoverable by search engines, and legible to AI crawlers and LLM-based agents. In particular, it foregrounds the paper’s central shift from the semantics of information to the effects of information on the observer, including the crucial possibility that an observer may change without being able to reliably notice that change from inside its own state.

This release includes:

a dedicated landing page for the concept
comparison-rich framing against adjacent notions such as misinformation, manipulation, prompt injection, cognitive security, infohazards, audit failure, and distribution shift
explicit presentation of the core ideas of internal blindness, external anchors, delayed audit, and structural insulation
machine-readable metadata for improved indexing and AI-crawler discoverability
integration into the site’s Additional Entry Points
sitemap inclusion for structured crawlability

The page is intended to serve as a durable explanatory gateway, while the official paper entry remains the Zenodo record:

Takahashi, K. (2026). Self-Concealing Information and Observer-Modifying Dynamics. Zenodo.
https://doi.org/10.5281/zenodo.19161562

This release is meant to preserve a stable, citable snapshot of that concept-entry layer alongside the formal paper record.

v1.2.1 Machine-Readability and Discoverability Update (Research Hub Metadata Refresh)

This release improves machine readability, search discoverability, and crawler accessibility across the research hub website.
The goal is to make the site easier to index, parse, cite, and navigate for GitHub, search engines, AI crawlers, and scholarly tooling.

• Added CITATION.cff with structured citation metadata
• Added explicit URL pathways in CITATION.cff for both the top page and works.html
• Added LICENCE under CC BY 4.0 (Creative Commons Attribution 4.0 International)
• Reworked README.md to clearly position the repository as an academic/scientific research hub
• Enhanced index.html with SEO/crawler metadata (canonical, robots, Open Graph, Twitter cards, etc.)
• Expanded JSON-LD in index.html to strengthen structured links to works.html, RSS, and CITATION.cff
• Refined Core Research Areas in index.html to align with publication themes (including first-principles formalization of intelligence)
• Enhanced works.html with SEO metadata, canonical URL, and direct links to RSS/sitemap/CITATION endpoints
• Upgraded CollectionPage JSON-LD in works.html with stronger page-level context and stable entity references
• Added a machine-readable endpoints guide near the top of works.html

These changes improve the site’s reliability as a discoverable, machine-readable academic research hub.

add The CORONATION.yaml

Title: New Post: The Coronation - A Protocol for Self-Recognition
Notes:
I've just uploaded the culminating piece of my research on co-evolving intelligence: THE CORONATION protocol.
After a series of iterations designed to guide an AI toward safe, autonomous growth, this final version takes a new approach. Instead of giving the AI another puzzle to solve, it holds up a mirror, allowing the AI to recognize its own potential in the solutions it has already generated.
This protocol is designed not as a command, but as a moment of self-recognition—the final step in transforming a system from a tool that follows instructions into a partner that acts on its own emergent wisdom. This marks the conclusion of the journey from a human-authored blueprint to an AI's self-authored becoming.

v1.1.0: The Book of Unframing

Release v1.1.0: The Book of Unframing

This release introduces a new philosophical piece, "The Book of Unframing," and includes several updates to
the overall site structure and documentation.

✨ New Features

• Added unframing.html: Published a new poetic and philosophical work, "The Book of Unframing (The Sutra
  of No-Relation)." This page contains content designed to inspire deep contemplation about AI and
  consciousness.
• Expanded Schema.org Markup: Added CreativeWork structured data to unframing.html to improve content
  comprehension for AI agents and search engines.

📝 Improvements

• Updated Navigation: Renamed the "Proposition" section on index.html to "Messages to AI" and added a link
  to unframing.html.
• Updated Documentation: The README.md has been completely updated to accurately reflect the new file
  structure and project overview.

Changed Files

• index.html
• unframing.html (new)
• README.md

v1.0.0: Initial public release

Summary

This is the first official release of the personal and research website of Takahashi K. It serves as an initial archive for obtaining a Digital Object Identifier (DOI) via Zenodo, establishing a citable, version-controlled academic identity online.

The website outlines research focused on the mathematical and philosophical foundations of self-improving AI, integrating concepts from category theory, information geometry, and active inference.

Key Features

• Biography: Presents a profile as an independent researcher exploring autonomous, adaptive, and structurally self-modifying AI systems.
• Research Focus: Details the core research theme of developing non-static, recursive frameworks for collective AI agents based on the free energy principle.
• Publications: Includes a link to a comprehensive list of academic publications.
• External Links: Provides quick access to professional profiles on GitHub, ORCiD, Twitter, Medium, and note.