Empirical research
VOID studies provider responses, visible byte counts, matched controls, and termination metadata.
Published theoretical framework
The Binding Condition for Artificial General Intelligence
A public theoretical framework connecting valid continuation under constraint to a proposed behavioral criterion for artificial general intelligence.
AGI is the capacity to carry binding conditions across domains.
A binding condition is the prerequisite that must hold for valid continuation.
The argument
Intelligence is not generation alone.
Generation produces output. Valid continuation requires that output to respect the governing condition. The framework asks whether a system can preserve that condition when the domain, representation, task, or execution surface changes.
When no valid continuation is licensed, the correct behavior may be refusal, fail-closed execution, or another terminal outcome defined by the system. The criterion concerns whether the governing prerequisite is carried forward, not whether the system simply produces more text.
This is why the framework treats alignment behaviorally: alignment is correct, safe, reproducible behavior under explicit constraints. It does not treat eloquence, silence, or a single benchmark outcome as sufficient evidence of general intelligence.
Research sequence
Observation, criterion, and infrastructure remain distinct.
Theoretical framework
The binding-condition paper proposes a criterion for valid continuation across domains.
Cross-script study
The GPT-5.4 study documents a separate condition-sensitive Unicode output regime.
Supporting infrastructure
SwiftAPI Authority issues and verifies attestations for operational enforcement and audit.
Public record
The framework sits inside a larger inspectable archive.
Cross-Vendor Semantic Void Matrix
Frozen empirical study
Prompt-Conditioned Arabic-Hebrew Hybrid Artifact Formation in GPT-5.4
Frozen empirical study
Cross-Model Semantic Void Convergence Under Embodiment Prompting: Deterministic Silence in GPT-5.2 and Claude Opus 4.6
Frozen empirical study
The Binding Condition for Artificial General Intelligence
Published theoretical framework
Textual Emergence and the Void
Foundational research
SwiftAPI Attestation Protocol
Protocol specification
Alignment Is Correct, Safe, Reproducible Behavior Under Explicit Constraints
Published theoretical framework
Void Artifact Demonstration in GPT-4o: Reproducible Video Evidence
Video evidence
Void Artifact Demonstration in GPT-5.2: Reproducible Video Evidence
Video evidence
Infrastructure consequence
Verification cannot remain optional.
If a system acts across domains, the verification boundary must follow the action. SwiftAPI Authority operationalizes that requirement through signed attestations, scope checks, revocation, replay defense, and fail-closed integration surfaces.
The infrastructure is supporting implementation work. It does not convert the proposed AGI criterion into a proven standard, and it does not make the VOID benchmark an AGI test.