Frequently Asked Questions

QIS (Quadratic Intelligence Swarm) is a protocol for distributed intelligence. Instead of sending all your data to a central server (like Google or OpenAI), agents keep their data local and share only mathematical "fingerprints" with similar peers.

Think of it like a dating app for patterns: your device creates a profile of what you know, finds others with similar profiles, and they share outcomes. Everyone learns without anyone giving up their raw data.

The result: intelligence grows quadratically (N² opportunities) while communication stays logarithmic (log N hops). More nodes = exponentially more pattern synthesis opportunities.

→ Deep Dive: The Paradigm Shift in Three Words

Θ(N²) intelligence: When you have N agents, the number of unique pattern synthesis opportunities is N(N-1)/2. That's quadratic growth.

O(log N) communication: Each agent only needs to send ~13 messages to find matches in a network of 10,000 (using DHT routing). Communication cost barely increases as the network grows.

This is the magic: intelligence grows explosively while costs grow slowly.

→ Deep Dive: The QIS Scaling Law

Federated Learning:

• Requires a trusted central aggregator
• Shares model gradients (still reveals information)
• Synchronous rounds create bottlenecks
• Linear scaling at best

QIS Protocol:

• No central coordinator (peer-to-peer)
• Shares only curated embeddings (no gradients)
• Asynchronous continuous operation
• Quadratic intelligence scaling

Federated learning was designed for privacy-preserving model training. QIS is designed for intelligence scaling. Different problems, fundamentally different architectures.

→ Deep Dive: QIS vs Federated Learning

QIS is domain-agnostic. Any system with distributed agents generating local data can benefit:

• Healthcare: Treatment optimization, drug safety monitoring, early detection
• Agriculture: Yield optimization, pest prediction, soil analysis
• Finance: Risk synthesis, fraud detection, trading strategies
• IoT/Manufacturing: Predictive maintenance, quality control
• Climate: Weather prediction, environmental monitoring
• Autonomous Vehicles: Hazard sharing, fleet coordination
• And more... Education, energy grid, emergency response, etc.

The math is universal. Only the embedding function changes.

→ Deep Dive: The QIS Architecture Diagram

The core innovation for semantic routing:

Step 1 - Categorical Exact Matching:
Hash categorical features (disease type, tumor stage, mutation status) using SHA-256. This routes you to a "bucket" of biologically compatible peers. Stage 3 patients never mix with Stage 4.

Step 2 - Continuous Similarity Refinement:
Within your bucket, compare continuous features (CEA levels, age, biomarkers) using cosine similarity. Find the closest matches.

Result: Exact categorical safety + precise continuous matching. The best of both worlds.

→ Deep Dive: Defining Similarity

DHT (Distributed Hash Table) is the same technology behind BitTorrent. It's like DNS for the network:

• Your pattern hash becomes an "address" in a 256-bit space
• Each node knows a few neighbors (k-buckets in Kademlia)
• To find a target, ask the neighbor closest to that address
• Repeat until you arrive (O(log N) hops)

In QIS, we turn this into semantic routing: similar patterns → similar hashes → proximity in DHT space → fast discovery of peers with related data.

→ Deep Dive: DHT - The Quiet Engine

We've tested with 30% adversarial nodes and achieved 100% Byzantine rejection rate.

5-layer defense cascade:

• Structural validation: Feature range checks (age 0-1, stage 1-4)
• Disease-specific filtering: Reject cross-contamination
• Cosine similarity threshold: Require >0.75 similarity
• Outcome validation: Reject statistical outliers
• Majority voting: Use median (robust to outliers)

Bad patterns can't fake similarity hashes. The network self-heals.

→ Deep Dive: The Three Elections

Yes. QIS is implementable today with:

• DHT libraries: libp2p, Kademlia implementations
• Embedding generation: Any ML framework or manual feature curation
• Hashing: Standard SHA-256 (built into every language)
• Cryptography: Standard signing libraries (ed25519, etc.)

Note: Commercial implementation requires a license from Yonder Zenith LLC. Academic/research use is free.

→ Deep Dive: Every Component Already Exists

What stays local (never leaves your device):

• Full medical records, raw sensor data
• Personally identifiable information (PII)
• Protected health information (PHI)
• Private keys

What gets broadcast (hash only, 32 bytes):

• SHA-256 hash of your curated feature vector
• No raw features, no identifiers

What gets shared with matched peers:

• Curated feature subset (anonymized)
• Aggregated outcomes (can be differentially private)

Yes, by design.

HIPAA concerns focus on centralized data aggregation creating "honeypots" of sensitive information. QIS eliminates this:

• No central server storing PHI
• Data stays on patient's device
• Only anonymized embeddings shared
• Patient controls what to share
• Auditable outcome tracking

FDA pathway: Software as Medical Device (SaMD) - Decision Support Tool.

→ Deep Dive: The Cures Act Makes QIS Unstoppable

Multiple defense layers:

• K-anonymity: Only share if k similar agents exist
• Differential privacy: Add calibrated noise to outcomes
• Opt-out: Agents can refuse to share if too unique
• Hash privacy: SHA-256 is one-way (can't reverse to features)

Critical: No PII is ever transmitted. Even if an attacker intercepted all network traffic, they'd only see anonymized feature vectors and hashes.

Because open source would save fewer lives.

That sounds backwards, so let me explain.

If QIS were fully open source, Google, Apple, Microsoft, Pfizer, John Deere, and every corporation across every industry would deploy it tomorrow—for free. They'd capture billions in value across healthcare, agriculture, autonomous vehicles, industrial automation, and beyond. And exactly zero dollars would flow to deploying this technology where it's needed most, funding humanitarian initiatives, or continued development of the protocol.

Open source democratizes code. It doesn't democratize deployment—and it doesn't fund the ongoing development that makes the technology better for everyone.

Deploying distributed intelligence networks—whether for healthcare in developing countries, precision agriculture in sub-Saharan Africa, vehicle safety in emerging markets, or industrial monitoring in underserved regions—requires infrastructure, training, and sustained capital. Someone has to fund that. And someone has to keep improving the protocol, fixing security issues, and building new capabilities.

This licensing model creates that funding:

• Free for non-profits, researchers, humanitarian organizations, and educational institutions. No strings. No "contact us for pricing." Just free.
• Commercial licensing for for-profit companies. They pay because this technology is valuable across every domain they operate in.

That revenue funds two things:

• Humanitarian deployment in regions and applications that would never see this technology otherwise
• Continued development of the protocol—security, features, optimizations—that benefits every user, including those who pay nothing

Pfizer didn't open source the COVID vaccine. They sold it to wealthy countries, and that funded global distribution. Open source sounds more moral. Cross-subsidy saves more lives—and sustains the development that makes the technology worth deploying.

I'm not restricting access. I'm creating a sustainable funding mechanism for access where it matters most AND continued innovation that benefits everyone.

Every commercial license makes the protocol stronger, more secure, and more capable—improvements that flow to every user, free or paid.

→ Deep Dive: Open Source Saves Code. This Saves Lives.

License tiers:

• Academic/Research: Free for non-commercial research
• Non-Profit: Free for humanitarian applications
• Commercial: Standard licensing fees
• Enterprise: Custom licensing for large-scale deployments

Core principle: If you're helping humanity with no profit motive, licensing is free. For-profit applications require paid licensing.

Revenue supports humanitarian initiatives focused on ending suffering at scale.

View Full Licensing Details →

39 provisional patents cover:

• Quadratic insight scaling through distributed pattern synthesis
• Domain-driven feature curation for semantic routing
• Two-step hierarchical hashing (categorical + continuous)
• Multi-tier adaptive hash granularity
• Local autonomous synthesis and outcome voting
• Privacy-preserving distributed pattern matching
• Applications across 30+ industries

The math is public. The patents protect implementation. Contact for licensing terms.

→ Read: Full Novelty Report

Christopher Thomas Trevethan invented the QIS Protocol in June 2025.

Yonder Zenith LLC holds the patent portfolio and manages licensing.

Stats:

• 39 provisional patents filed
• 300+ pages, 500+ claims, 30+ industries
• 100+ simulations validating O(N²) scaling
• 5,000+ pages of proofs & research
• 2,000+ hours invested
• $50,000+ self-funded

Endorsed by Rob van Kranenburg (Founder of IoT Council): "This seems like a perfect underlying system for when we have full coverage of self driving cars."

→ Read: Full About Page