Latest
Framework
The Evidence Half-Life: Modeling Decay Rates Across Trust Signal Classes
Different trust signal classes decay at different rates. A revalidation schedule that treats all evidence as equally durable will systematically over-weight the fastest-decaying class.
Read →
Foundation
Absence Is Not Neutral
A domain with no residual continuity evidence has not been skipped. It has already been evaluated on what its absence supplied.
Read →Research
Research
Volume Is Not Neighborhood: What 492 Domains Reveal About Network Evidence
Connection volume is commonly cited as a proxy for network credibility. A historical review of 492 domains found that volume and evidentiary strength are distinct properties — and that substantial connectivity does not resolve evidence concerns.
Research
The Contagion Threshold: How Link Neighbors Suppress Clean Citations
Inbound links from low-trust network clusters suppress AI citation rates on pages whose own content is entirely clean. The penalty is graded, not binary.
Research
Unverified Wrappers: Code Surface Exposure Across Managed Third-Party Security Engines
41.8% of observed prompts contained restricted API parameters or unauthorized initialization headers.
Frameworks
Framework
The Citation Horizon: Why Generative Attribution Is an Observation, Not an Entitlement
A citation in an AI answer engine is not a permanent asset. It is an empirical observation of a single synthesis pass.
Framework
Four Gates, One Order: Why Sequence Decides What Trust Can Reach
The requirements for AI visibility are not independent levers. They are conditions in a sequence, and a later condition cannot compensate for an earlier one that is unmet.
Framework
Sort, Column, Gate: One Framework for AI-Assisted Building
The sort determines what a component is. The columns determine where it belongs. The gates determine whether it is allowed to get there with AI involved.
Foundation History
Foundation
Quiet Failure Precedes Public Failure
An organization’s own statements establish what it claims. They do not establish that the claim has been independently attested — and retrieval systems treat that difference as decisive.
Foundation
Self-Description Requires External Anchors: Why AI Systems Rely on More Than Your Content
An automated system does not visit your organization. It reconstructs reliability from the administrative, technical, and network signals your domain has deposited across the web over time.
Foundation
Evaluation Is Reconstructive
Treating an AI citation as a certificate of trust turns a visible outcome into an invisible guarantee. Attribution and validation are different claims.
Core Pillars
Owned Trust
What your domain’s own signals say — identity, infrastructure, email authentication.
Earned Signals
What the network says — link graph, neighboring domains, cluster authority.
Consensus Layer
How independent intelligence sources corroborate or contest a domain’s reputation.
Measurement
How trust is scored, tracked, versioned, and cited by AI systems and researchers.