計畫PRG-2026-0001v0.1
自適應世界狀態系統的第一原理框架
從第一原理推導 AI runtime 的長期研究線：世界知識以不同速率變化、自然語言是 rendering 而非 canonical state、能力必須被記憶與重用、算法與計算載體分離；然後把推導結果逼進可執行的現實——AER-0 參考 runtime 與六輪架構比較，以及檢驗非概率 primitive 是否收斂到概率表象的 PACC 微型實驗室。
目標
goals- Derive an adaptive world-state architecture without starting from existing AI technology names, then test whether its differences survive implementation.
- Decide, falsifiably, between Distinct Advantage, Operational Convergence, Behavioral Equivalence Only, Inconclusive and Architecture Worse.
- Test the Probabilistic Appearance Convergence Conjecture (PACC) in fully observable micro-environments before touching language models.
開放問題
open_questions- PACC-A (architecture attractor) is not closed: only two independent convergent families exist under the preregistered redundancy rule.
- Reverse chart transport in the PACC atlas carries a persistent destination bias that neither a bounded quadratic transition nor a base-point residual field explains (v0.12–v0.13); v0.14 tests a sign-free residual subspace.
- AER-0 has no executable cross-runtime comparison yet: LangGraph could not be installed in the comparison environment (R2), and R7 (independent witness, transparency proof, signer rotation) is not started.
- The real-language-model PACC A/B/C benchmark has a validated harness but has never been executed with a real model.
里程碑
milestones- Adaptive Epistemic Systems Series papers 01–11 complete (canonical UTF-8 sources, SHA-256 manifest).
- Adaptive Epistemic AI Runtime technical whitepaper v0.1 with schemas, pseudocode and the AER-0 roadmap.
- AER-0 MVP v0.1 (Python + SQLite, 25 tests) and comparison rounds R1–R6 (79 tests at R6).
- PACC conjecture paper (2026-09-08) and PACC-Lab v0.1–v0.13.
- PACC-LLM Hybrid Lab v0.1 (synthetic A/B/C) and v0.2 (real-LLM harness, not yet executed).
此計畫下的研究線與系統
| 來源 | 關係 | 目標 | 狀態 | ID |
|---|---|---|---|---|
RES-2026-0001 自適應認識系統的盲推導 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0001 |
RES-2026-0002 PACC 猜想——非概率 primitive 會不會收斂到概率表象？ | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0002 |
RES-2026-0003 認識論治理能不能活過實作？AER-0 架構比較 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0003 |
RES-2026-0004 PACC 式 runtime 用在語言模型上：推理、意圖與創造廣度 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0004 |
記錄欄位
limitations- Everything measured so far is synthetic data plus theoretical reasoning, or a deterministic scenario in a reference runtime. Until a real hybrid model exists, an inference is only an inference: theoretically possible is not actually possible.
關係
| 來源 | 關係 | 目標 | 狀態 | ID |
|---|---|---|---|---|
RES-2026-0001 自適應認識系統的盲推導 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0001 |
RES-2026-0002 PACC 猜想——非概率 primitive 會不會收斂到概率表象？ | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0002 |
RES-2026-0003 認識論治理能不能活過實作？AER-0 架構比較 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0003 |
RES-2026-0004 PACC 式 runtime 用在語言模型上：推理、意圖與創造廣度 | belongs_to | PRG-2026-0001 自適應世界狀態系統的第一原理框架 | ACTIVE | REL-2026-0004 |
歷史與來源歷程
- Canonical URL
- https://evemisslab.com/ai/programs/PRG-2026-0001/
- 快照
AI-SNAPSHOT-v0.1-fe85b9694a45- 來源歷程
source- EveMissLab research collection: Adaptive Epistemic Systems (真本體論13)
extracted_by- Splice (Claude Code), reading the canonical UTF-8 sources and each lab's own result reports
extracted_at- 2026-09-11
generator- tools/extract_aes/extract.py
claim_boundary- status, evidence level and result type follow the source artifact's own stated claim boundary; nothing is upgraded beyond what the report supports