About BLACKSHEEP OI
Building trustworthy AI with risk management at its core
Our Mission
BLACKSHEEP OI's mission is to deliver energy-efficient, trustworthy AI systems that reduce computational waste and enable sustainable, scalable intelligence.
From inception, BLACKSHEEP OI has been developed with risk management at its core. We recognize that responsible AI development is inseparable from trust, transparency, and oversight.
Guiding Principles
Transparency
All claims are evidence-based; limitations and gaps are explicitly documented
Safety by Design
Safety-critical and export-controlled applications are prohibited
Auditability
All prototype tests, updates, and governance actions are logged for review
Proportionality
Controls are scaled to the realities of our current stage and team
Our Team
Taylor Jenkins
Founder, CEO & CTO
Marine veteran. Sole inventor of the Recursive Entropic Computing paradigm — all mathematics, algorithms, and implementations. Algorithm architecture and core AI system design.
Nathan Nelson
Co-Founder & COO
Systems integration and operations
Strategic Goals
- ▸License the REC paradigm across semiconductor, energy, healthcare, defense, and automotive verticals
- ▸Pass independent audit reviews for compliance with U.S. and international AI risk standards
- ▸Maintain mission-driven operations over profit extraction
Compliance & Standards
We maintain alignment with major AI governance and regulatory frameworks:
Intellectual Property
Protected
133+ solutions across 6 patent families — patent pending.
Coverage spans architecture, algorithms, hardware, communications, and security.
Licensing available for commercial implementations. Academic research use permitted with attribution.
Technical Validation
Every system in the REC stack has been built, tested, and validated. Comprehensive test suite — all tests passed across multiple categories and data types. These are measured results, not projections.
ELF Pipeline
Constant-depth execution design verified across variable input sizes. Predictable latency confirmed at production-grade throughput. Zero floating-point.
FEM Compression
Up to 9,286:1 on structured data. 29:1 on enwik9 — 3× beyond world record. 61–64:1 lossless on production manufacturing files. SHA-512 verified across 25 data types.
QER Search
O(log log n) scaling confirmed across datasets from 10³ to 10⁹ records. 0.8 microseconds at billion-record scale. Exact recall.
Energy Efficiency
14,000× measured vs transformer inference. 1.15 nanojoules per query. Conservative floor: 1,000×.
Production Deployed
FEM compression running on production manufacturing files in a certified facility. 64:1 lossless on STL files. Industry-standard approaches like Google Draco achieve typically 20–60:1 through lossy quantization — FEM preserves original geometry exactly.
Scientific Discovery
CARTO discovered known physics equations from raw data with no priors. Multiple laws discovered and algorithms synthesized from a single simulation.
Reproducible Benchmarks: Complete validation documentation maintained. Every claim has a test. Every test has a measured result. Available to licensing partners.
Get In Touch
Ready to discuss licensing, enterprise deployment, or research collaboration?