What question did this study set out to answer?

The aim is to enhance the control and safety mechanisms of artificial general intelligence (AGI) through a new framework.

April 24, 2026Open Access

Decision-OS V8: Time-Tube Control for Self-Safe AGI

Key Points

The aim is to enhance the control and safety mechanisms of artificial general intelligence (AGI) through a new framework.
Presents the Time-Tube concept for trajectory-based control in AGI evolution.
Introduces the Dependency Gradient as a risk factor for foreseeing potential failures.
Redefines the Guardian role in AGI as an emergent phenomenon under safety constraints.
V8 marks a significant evolution in AGI control by integrating long-horizon trajectory management.
The Dependency Gradient aids in early identification of risks before potential failures occur.
Establishes a three-layer control structure focusing on internal and external safety.

Abstract

Gateway (Series entry) GitHub gateway (public index + evidence routing): https: //github. com/shin4141/decision-os-paper Two entry paths A) Practical safety / implementationStart with V5 (SiriusA): a protection-of-life confirmation layer with zero-knowledge approval, two-step confirmation, revoke path, and approval window Δt. https: //zenodo. org/records/18051625 B) AGI theory / control overviewStart here with V8 v2, then read V6 (PIC), V7, and return to V5 for implementation-oriented safety design. Canonical citationPlease cite this Zenodo record (DOI) as the canonical public artifact. GitHub functions as the public index / gateway and SSOT-routing surface, not as the archival citation target. Version notev1 was released as a Research Note. This record is the current full-paper release in the same Zenodo concept DOI lineage. SSOTGit commit SHA: a771021949fabf4b0ea23141863bab5f1a266e8c Abstract / overviewDecision-OS V8 presents a control-centered view of AGI evolution and extends the Decision-OS lineage from V4 through V7. Relative to the earlier lineage, V8 contributes three main elements: 1. Time-TubeA trajectory-based unit for long-horizon evolution, replacing pointwise judgment with trajectory control. 2. Dependency GradientA pre-collapse risk factor that shifts the tube center outward and functions as an early warning signal before irreversible failure. 3. Guardian redefinitionGuardian is treated not as a module at initialization, but as an emergent long-horizon phenomenon arising under Guard constraints. V8 frames Self-Safe AGI through a three-layer control structure: - External Harm- Internal Collapse- Dependency In this sense, V8 functions as the synthesis and control-completion layer after V7’s structural AGI definition, where AGI is framed through Aspire × Guard × Self-Recursion grounded in PIC. KeywordsTime-Tube; Dependency Gradient; Guardian; Guard; Self-Safe AGI; Aspire Intelligence; Self-Recursion; PIC; Multi-Tube Transparency / Author’s Notehttps: //github. com/shin4141/decision-os-paper/blob/main/AUTHORSNOTE. md

Decision-OS V8: Time-Tube Control for Self-Safe AGI

Key Points

Abstract

Cite This Study