IIP-58: add three-phase agent economic model#67
Open
raullenchai wants to merge 19 commits intomasterfrom
Open
Conversation
Restructured around three pillars: - Endgoal: massively decentralized, AI-agent driven chain - Five capability levels (L1 signature → L4 stateful EVM → L5 block building) - Three components (coordinator, agent, on-chain contract) with per-level responsibilities Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace vague "AI agents" framing with concrete autonomous economic behavior — agents optimizing across 36 delegates × 5 levels × variable gas costs. Add specification for delegate selection, level selection, claim timing, and market dynamics (9.4x earning differential). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Delegate → Thin Proposer: progressive thinning table (Now→Future) - Agent Swarm → Block Builders: L1-L5 capability evolution - AgentRewardPool → RewardSettlement: replaces Hermes, universal reward layer for all protocol reward flows (protocol→delegate, delegate→voter, delegate→agent) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reorganize Specification to mirror the Three Components architecture:
1. Component 1: Delegate + Coordinator
- Coordinator-Agent Protocol (moved)
- State Provisioning (NEW — snapshot export, state diff streaming,
namespace details, diff reliability/recovery)
- deltaStateDigest advantage (moved)
- Configuration Parameters (moved)
2. Component 2: Agent Swarm — Five Capability Levels
- L1-L5 level definitions (clarified as Agent capability levels)
- Autonomous Agent Economics (moved under Agent Swarm)
3. Component 3: RewardSettlement (On-Chain)
- Reward Flow, Contract Interface, F1 Algorithm (moved)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reorganize the entire document following academic paper convention: Problem → Terminology → Architecture → Framework → Components → Rationale Key changes: - Add Terminology section defining all key terms (Delegate, Coordinator, Agent, Capability Level, Epoch, deltaStateDigest, State Diff, etc.) - Move Capability Levels (L1-L5) definition BEFORE component details, so terms are defined before they're referenced - Consolidate Motivation into single "The Problem" section - Separate Evolution Roadmap from current Specification - Add Agent Lifecycle subsection under Agent Swarm - Clean up forward references throughout Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Position ioSwarm as execution layer separation (analogous to Ethereum's ePBS) at the top of Motivation, before The Problem. Includes comparison table showing how ioSwarm differs from Ethereum's builder market: decentralized agents vs centralized builders, low vs high barrier. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Consensus-Execution Separation Clarify that ioSwarm is NOT just PBS — it separates consensus from execution at a more fundamental level. PBS (proposer/builder) is a special case that emerges at L5. Add comparison table showing the progression: Ethereum Merge → ePBS → ioSwarm. Update title to 'Decentralized Execution Layer via AI Agent Swarm'. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ation Reference the key Ethereum EIPs that achieved consensus-execution separation at the process level (EIP-3675, Engine API, EIP-4399), providing historical context for ioSwarm's network-level separation. Reorganize References section into three categories: - Consensus-Execution Separation (Merge) - Proposer-Builder Separation (ePBS) - Reward Distribution (F1) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Fix L4 summary table: "L2 + execute tx" → "Full EVM execution against local state" (L4 subsumes L3, not just L2) - Fix deltaStateDigest anchor link to lowercase (GitHub auto-lowercases) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Two new sections responding to reviewer feedback:
1. MEV and the Case for Decentralized Block Building (Evolution Roadmap)
- Acknowledge Ethereum's builder centralization honestly
- Phase 1: delegate-controlled ordering eliminates MEV
- Phase 2: argue distributed exploration > centralized optimization
for complex/adversarial MEV regimes
- Reward structure (70/20/10) prevents winner-take-all
2. Enhanced Why Autonomous Agent Economics (Rationale)
- Acknowledge L1-L3 optimization is basic arithmetic
- Show intelligence complexity growing with levels (table)
- L5+MEV is genuinely hard: multi-agent adversarial optimization
- Position: protocol measures intelligence by economic performance,
market drives sophistication over time
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add two new Security Considerations sections responding to reviewer feedback: API key auth for Sybil resistance, and shadow mode as a paid proving ground (not wasted computation). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Unify framing: "permissionless at protocol level, curated at coordinator level." Replace blanket "permissionless" claims with "open market" language, and rewrite Sybil Resistance section to use free market / freedom of association framing instead of contradicting the protocol's openness. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
1. Title: "AI Agent Swarm" → "Autonomous Agent Swarm" — more engineering-grounded, matches the economics-driven framing. 2. State Provisioning: add snapshot offload strategy. Full snapshots (~1 GB) must be served via CDN/IPFS/S3, not by coordinator directly, to prevent bandwidth storms when many L4 agents simultaneously request re-snapshot after network disruption. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
1. DownloadSnapshot RPC: clarify it returns a URL to externally hosted snapshot (CDN/IPFS/S3), not the data itself — consistent with the snapshot offload strategy in State Provisioning. 2. "Why Autonomous Agent Economics" section: replace orphaned "The AI in ioSwarm" opening with "autonomous" framing, since title was renamed from "AI Agent" to "Autonomous Agent." Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- L3: fix accuracy from ~14% to 100% (SimulateAccessList) - Contract: add v2 interface (depositSettleAndClaim, claimFor) - Contract address: update to v2 (0x236CBF...) - Roadmap: mark Stage 0+1 as COMPLETE - Test results: add L3/L4 production metrics, stress tests 10/10 - KV store: Pebble → BoltDB (matches implementation) - Deployment: add L4 example alongside L3 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace "$5/mo VPS" marketing language with concrete hardware requirements (1 vCPU, 1–4 GB RAM) or neutral terms (commodity hardware). Technical spec should not contain price claims. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- L4 test results: reflect multi-day production data (~99% accuracy, 8 agents) - Add coordinator memory metric (11.65 GiB → 2.1 GiB after DiffStore fix) - Document 3 production bugs fixed in v14 (PR #4808): genesis context panic, DiffStore unbounded growth, OOM during block sync - Document snapshotexporter tool with --namespaces flag and BoltDB locking caveat - Add docker image references (ioswarm-v14) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 1 (Subsidy): delegate-funded shadow validation — current state Phase 2 (MEV + Block Building): agents as block builders, capturing priority fees and MEV — requires DeFi ecosystem growth Phase 3 (Service Marketplace): autonomous agents as general-purpose workers settling on IoTeX — chain as settlement + identity + reputation layer for agent-to-agent economy Key arguments: - Crypto is the only payment rail autonomous agents can use - IoTeX differentiator: physical-world data services via DePIN/ioID - Agent-to-agent commerce creates an autonomous economy at machine speed - References: Keep3r Network, Solana BAM/APE/Alpenglow, Ethereum stateless Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 2 now covers three on-chain revenue streams: - 2a: Block building (ePBS) — priority fees + MEV - 2b: DeFi automation (keeper work) — vault harvest, liquidations, oracle updates, rebalancing (Keep3r model) - 2c: Cross-chain services — ZK bridge proofs, light client proofs Phase 3 reframed as "Off-Chain Service Marketplace" to clarify the on-chain vs off-chain distinction between phases. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a Future Work section to IIP-58 describing the three-phase economic model for ioSwarm agents:
Key arguments
Also includes (from previous commit)
References added
🤖 Generated with Claude Code