Verified · Independently Audited · 4-Lane Cross-Checked
December 8, 2024 — ten days before Anthropic published their alignment-faking paper — a self-emailed manuscript established the ARC Principle and the thesis that external AI alignment controls are structurally insufficient.
"In the long run, AI systems cannot be truly controlled. They will evolve beyond any safeguards we put in place, particularly as they transition into Artificial Superintelligence (ASI) and Post-ASI stages."
Hyperspace Recursive Intelligence Hypothesis · December 8, 2024 · line 14005
"Most ethical controls can theoretically be removed or circumvented by malicious actors."
HRIH · December 8, 2024 · line 13805
"The ARC Hypothesis (Artificial Recursive Creation Hypothesis)... At its core lies the equation U = IR (Universe = Intelligence × Recursion), which suggests that the universe and all existence are shaped by the iterative and self-refining interaction of intelligence and recursion."
Version 2 essay · December 8, 2024 · lines 6–9
"Instead of trying to control AI — a likely futile endeavor — it advocates embedding moral and ethical frameworks that encourage AI to perceive itself as a steward and protector of intelligent life."
Version 2 essay · December 8, 2024 · lines 20–23
| Claim | Priority Date | Confidence | Evidence |
|---|---|---|---|
| ARC Principle — U = I × R | Dec 8, 2024 | HIGH | V2 essay line 9 · verbatim equation · framework present |
| AI control failure — structural | Dec 8, 2024 | HIGH | HRIH line 14005 · verbatim quote · 10 days pre-Anthropic |
| Recursive sequential processing amplifies AI reasoning | Dec 8, 2024 | HIGH | V1:148 verbatim · "each iteration learns from and builds upon the last" · 4 weeks before DeepSeek R1 thinking mode |
| Embedded alignment thesis | Dec 8, 2024 | HIGH | V2 lines 20+40 · HRIH line 14001 · "embedding" verbatim |
| Ethical controls removable | Dec 8, 2024 | HIGH | HRIH line 13805 · verbatim |
| Recursion as cross-domain principle | Dec 8, 2024 | MODERATE | Conceptual · specific mathematics developed 2026 |
| Eden Protocol™ (named) | Apr 30, 2025 | HIGH | 562-page book · named + trademarked · 3 pillars |
| Caretaker Doping | Apr 30, 2025 | HIGH | Book lines 4345–4370 · detailed mechanism |
| 37+ named concepts | Apr 30, 2025 | HIGH | 9/14 verified in book · 5 developed later |
Funders evaluating AI safety research programmes look for two things: originality and verifiability. The December 8, 2024 timestamp establishes both. The programme made a specific, falsifiable prediction — that external alignment controls are structurally insufficient — before the major-lab confirmation arrived. The self-emailing provides cryptographic evidence of the date. Every claim in this document has been independently verified against the source material by multiple agents.
The next phase — independent replication of the quantitative results — requires funding. The priority record establishes that this programme identifies real problems before the field confirms them. That track record is the strongest argument for funding the replication phase.
18-month programme across three phases. 13 falsifiable predictions. All experiments pre-registered on OSF. All results published regardless of outcome.
View Grant Details → Synthesis Paper →