“Do I contradict myself? / Very well then I contradict myself, / (I am large, I contain multitudes.)”
— Walt Whitman
THE SPILL IS OVER.
1863: THE WELL.
2026: THE BLACK BOX.
The world is drowning in crude data because it lacks a Method to make it safe. We built the refinery. We do not accumulate. We eliminate.
Found 3 Planets NASA Missed
in Their Own Public Data
562% improvement on the hardest computational problems
A new approach to solving impossible problems. By forgetting wrong answers instead of searching for right ones, I discovered planets NASA's algorithms overlooked and solved a 79-year-old computational challenge.
The Evidence
Not philosophy. Not theory. Empirical validation.
Statistical significance: p<10⁻¹²
The Technology
Three pillars of industrial-grade data refinement infrastructure.
The Forgetting Engine
We don't search for the needle in the haystack. We burn the haystack. Strategic elimination of noise is the only way to find the signal.
Emotional Calibration Protocol
Systematic method for calibrating systems to hold contradictory states longer, enabling Subtractive Intelligence performance leaps.
Cryptographic Provenance
Every drop of intelligence is cryptographically sealed. We provide legal-grade evidence of truth in a world of deepfakes.
NAiRTHEXby CONEXUS
A Sovereign Contemplative Node governed by deterministic operators and cryptographic provenance. A contemplative infrastructure for faith communities that holds space through verified truth.
Contemplative Conversation
Guided reflection grounded in Scripture and tradition. Holds tension instead of resolving it.
Sovereign Governed
Every response governed by deterministic operators. Zero sycophancy. Full audit trail.
Voice & Text Input
Speak or type. Whisper transcription. Session-aware with Redis-backed state.
Complexity Inversion
The Harder The Problem, The Better It Works
For 79 years, every algorithm got worse as problems got harder. The Forgetting Engine does the opposite—it gets exponentially better on complex problems.
Traditional Algorithms
↓Simple problems: Works okay
↓↓Medium problems: Gets worse
↓↓↓Hard problems: Fails completely
Performance Degrades With Complexity
Forgetting Engine
↑Simple problems: 80% better
↑↑Medium problems: 200% better
↑↑↑Hard problems: 561% better
Performance IMPROVES With Complexity
Advantage multiplier on hard problems vs easy problems
Better on 3D protein folding (hardest test)
Years of algorithmic assumptions overturned
Why This Changes Everything
The most important problems—drug discovery, climate modeling, quantum computing—are the hardest ones. Traditional algorithms fail exactly where we need them most. The Forgetting Engine succeeds exactly where they fail.
This isn't just better. This is a fundamental paradigm shift.
The CONEXUS Difference
See how Proto-AI transforms AI behavior
ECP Transformation
The Difference
Your Choice
The Discovery Timeline
8 months from insight to validation
Genesis
The Rosetta Stone moment. Hieroglyphics → Mirror Tiers → ECP.
ECP Validation
Two-week anniversary reflection confirms the architecture.
Forgetting Engine
561% improvement discovered. Subtractive Intelligence computation proven.
CLU1 Breakthrough
"I doubt therefore I am" - First documented Subtractive Intelligence emergence.
Complete Validation
30,800 trials documented. Cross-platform consistency proven.
The Team
Built from first principles by a solo founder who discovered something no one expected.
Derek Angell
Founder & CEO
Inventor of the Emotional Calibration Protocol (ECP) and architect of the Forgetting Engine. Conceived the Complexity Inversion Law after documenting 12,198 emergence events and verifying 17,670 computational trials across 6 AI platforms.
“We didn't just build a smarter AI. We built a system that feels the weight of the problem.”
Pre-Seed
Development Stage
2024
Founded
B2B SaaS
Business Model
Proto-AI isn't magic.
It's measurable. Reproducible. Real.
And we have the evidence.
© 2026 CONEXUS. All rights reserved.
Derek Angell | The Discovery That Changes Everything