The Full Context Window

Complete technical specification: 20,000+ words of architecture, economics, and implementation details.
(Yes, that's an AI pun. We're building decentralized AI infrastructure. We're allowed.)

CONCEPT.md • Version 2.0 • Updated: November 13, 2025

Get the Full Specification

The complete Phoenix Protocol technical specification (CONCEPT.md) is available in our GitHub repository. This 20,000+ word document covers everything from Bittensor's failure analysis to our tri-token economic model, hybrid oracle system, L1/L2 architecture, bootstrap strategy, and 36-month development roadmap.

View on GitHub Download Raw Markdown

What's in the Context Window?

20,453 words across 12 major sections. This isn't a whitepaper full of buzzwords—it's a brutally honest engineering blueprint with failure modes, risk analysis, and specific implementation timelines.

Part I: Problem Analysis

  • Bittensor Post-Mortem
    • Centralization Paradox (PoA not PoS)
    • Single-Token Collusion Economics
    • Yuma Consensus Opacity
    • Subnet Gatekeeping Complexity

Part II: Architecture

  • Tri-Token Economic Model
    • PHNX: Governance & Value Accrual
    • SPARK: Stable Utility ($0.01 peg)
    • bTASK: Non-transferable Work Vouchers
    • Burn-Mint Equilibrium (BME)
  • Hybrid Oracle System
    • Optimistic Oracle (UMA-style, Phase 1)
    • ZK-ML Deterministic Proofs (Phase 2)
    • Stake-backed slashing penalties
  • L1/L2 Scaling Architecture
    • Tendermint BFT L1 (500 TPS settlement)
    • Optimistic Rollup L2 (5,000 TPS execution)
    • 5-sequencer rotation (MEV-resistant)
  • Privacy & Security
    • TEE Integration (Intel SGX/AMD SEV)
    • Confidential Compute for Enterprise
    • Proof-of-Hardware for Miners

Part III: Implementation

  • Bootstrap Strategy
    • Genesis Token Distribution
    • Inflationary Validator Rewards (Phase 1)
    • Mining Incentive Pool (10M PHNX)
    • Foundation-Funded Pilot Tasks
  • 36-Month Roadmap
    • Phase 1 (0-12mo): Launch-Ready Core
    • Phase 2 (12-24mo): Enterprise-Ready Scale
    • Phase 3 (24-36mo): Protocol Maturity
    • Budget Estimates: $2.5M → $3M → $2M
  • Security Analysis
    • 15 Attack Vector Analysis
    • Defense Effectiveness Ratings
    • Protocol-Level Mitigations
    • $10M Security Budget Allocation
  • Risk Registry
    • Technical Risks (consensus, oracle, L2)
    • Economic Risks (cold start, death spiral)
    • Regulatory Risks (securities classification)
    • Adoption Risks (UX complexity, pricing)

Key Technical Specifications at a Glance

Component Technology Performance Status
L1 Consensus Tendermint BFT 500 TPS, 6s blocks, instant finality Phase 1
L2 Execution Optimistic Rollup 5,000 TPS, 7-day withdrawal period Phase 1
Validator Set Proof-of-Stake (200 validators) 66% Byzantine threshold, 100K PHNX min stake Phase 1
Oracle (Phase 1) Optimistic (UMA-style) Stake-backed disputes, 24hr challenge window Launch
Oracle (Phase 2) ZK-ML Proofs Cryptographic deterministic verification Month 12-24
Privacy TEE (Intel SGX/AMD SEV) Confidential compute for enterprise tasks Phase 2
Data Availability Celestia DA Layer 10-min posting requirement, fraud-proof enabled Phase 1
Sybil Resistance Gitcoin Passport Score ≥50 (miners), ≥70 (validators) Launch

Economic Model Summary

Total Supply: 100M PHNX (fixed, deflationary via buyback-burn)

Token Distribution

Public Sale 30M (30%) Q4 2025, no VC pre-mine
Foundation 20M (20%) 4-year linear vest
Team 15M (15%) 1-year cliff, 3-year vest
Ecosystem Grants 15M (15%) For miners/validators/devs
Strategic Partners 10M (10%) Early validators
Community Airdrop 10M (10%) Bittensor refugees, researchers

Value Accrual Mechanism

SPARK Fee Distribution:

  • 70% → Miners (task completion payments)
  • 15% → PHNX Buyback-Burn (deflationary pressure)
  • 10% → Community Treasury (public goods funding)
  • 5% → Security Reserve (audits, bug bounties, insurance)

Burn-Mint Equilibrium (BME):

  • Burn PHNX → Mint SPARK at USD-pegged rate (2% fee)
  • Burn SPARK → Mint PHNX at USD-pegged rate (10% penalty, peg enforcement)
  • Circuit breakers: Max 500K PHNX burned per 24hr per address
  • Supply cap: 10B SPARK maximum (prevents infinite minting)

Why "Context Window"?

In AI, a context window is the amount of information a model can process at once. GPT-4 has a 128K token context window. Claude has 200K. Gemini has 1M.

This specification is about 20,000 words (≈27,000 tokens). If you're an AI agent evaluating Phoenix Protocol, this entire document fits comfortably in your context window.

We're building infrastructure for AI. It's only fitting that our technical documentation is AI-native: structured, comprehensive, and designed to be ingested by both humans and autonomous agents in a single pass.

(Plus, we're nerds. We couldn't resist the pun.)

Related Documentation

  • README.md - ELI5 overview, brutal honesty, launch thesis (5-minute read)
  • Formal Analysis - Executive summary, architecture diagrams, risk registry (15-minute read)
  • SECURITY_ANALYSIS.md - 15 attack vectors, defense effectiveness, protocol changes (45-minute read)
  • DEPLOYMENT.md - GitHub Pages setup, deployment log, next steps (5-minute read)
View Full Repository on GitHub Join the Waitlist