What does NeoAmorfic™ do?
NeoAmorfic™ builds a physics-first intelligence layer that converts time-series and trace telemetry into auditable structural signals. The core is QEIv15™, a modular entropy-physics kernel. On top of QEIv15™, we ship domain engines (EEG, CO₂, AELIE™, Protein, QaaFS™, F1) and a governance product for AI/ML inference called X-40™.
What is QEIv15™?
QEIv15™ is a deterministic physics kernel that computes structural invariants and stability indicators (including entropy families, Δ-dynamics, curvature/stiffness, φ-field structure, and λ-style stability measures) from real input signals. QEIv15™ is not a statistical model and does not require training data. It is a modular engine designed to generalize across domains because the governing quantities are physics-based, not dataset-based.
Is QEIv15™ “AI”?
QEIv15™ is best understood as measurement + structure inference. It produces physically grounded indicators that can govern decisions and pipelines. If you define intelligence as “the ability to produce reliable action under uncertainty,” then deterministic governance is a powerful form of intelligence—especially in mission-critical systems where “close enough” is catastrophic.
What is X-40™?
X-40™ is a runtime governance layer for LLMs and ML systems. It sits above inference outputs and returns an actionable decision such as ACCEPT or REQUIRE_VERIFICATION, plus reason codes and audit fields. X-40™ is designed to reduce the operational risk of deploying probabilistic systems in regulated or high-stakes environments.
What is the core differentiator?
X-40™ is built around dual evidence:
- Channel A: Behavioral traces (uncertainty dynamics, confidence margins, repetition/degeneracy signals).
- Channel B: Structural evidence from QEIv15™ anchors (Φ, κ, ΔS families) computed independently via ResearchCore.
One channel can fail silently; two independent channels produce a stronger governance basis. In addition, X-40™ includes deterministic envelopes (math verification, unknowns enforcement, forbidden-output redaction) that make failure modes auditable and controllable.
Do you store customer content?
X-40™ is designed to operate in privacy-first modes. Customers can choose to store no user content and retain only minimal audit fields (indices, reason codes, hashes). In regulated contexts (legal, finance, health), this matters as much as raw performance.
Who is this for?
NeoAmorfic™ serves two buyer archetypes:
- Engine buyers (domain teams who want physics structure signals in EEG, climate, markets, etc.).
- Governance buyers (teams deploying LLM/ML in production who need safety, auditability, and controlled automation).
How do we work with customers?
For X-40™, we onboard by defining workload profiles (facts/math/unknowns/attack or ML-specific profiles), capturing baselines per prompt class, and configuring policy thresholds. The deliverable is a hosted API (or on-prem container) plus license keys, profiles, and a reproducible benchmark protocol.