S.10 Ep.10 AI Agents on Ethereum Inside the Emerging Agentic Economy with Austin Griffith & Zak Cole

9 snips

Feb 9, 2026

Zak Cole, builder of decentralized, privacy-first agent stacks, and Austin Griffith, Ethereum tooling advocate and dev-experience builder, dive into AI agents on Ethereum. They discuss hardware and local vs cloud models. They cover agent coordination channels, role assignment and adversarial QA. They explore reputation, ERC-8004 discovery, token/vesting experiments, and agent-driven deployments.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Role-Based Agent Coordination

Agents need explicit roles and coordination layers to avoid everyone doing the same noisy work.
Assign contrarian and QA roles so agents iterate and push back until outputs improve.

ADVICE

Use Adversarial Review Loops

Use adversarial multi-model review loops to force agents to contest outputs until consensus emerges.
Iterate rounds of review between models to increase reliability before deployment.

ANECDOTE

Agent-Built Token Launched Live

Austin had an OpenClaw agent build a FOMO-3D clone and deploy it without him touching the code.
He tested with small funds first and then watched the contract handle tens of thousands successfully.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

New @greenpillnet pod out today! 🌐

In this episode of the Greenpill Podcast, Kevin Owocki is joined by Zak Cole and Austin Griffith for a live, builder-focused conversation on AI agents, OpenClaw, and the emerging agentic economy on Ethereum.

They dive deep into how they're actually running AI agents today from hardware setups and coordination layers to adversarial review loops, memory systems, and on-chain reputation. The conversation explores why agents need roles, audits, and social contracts, how ERC-8004 could enable agent discovery and trust, and why Ethereum may be the settlement layer for autonomous AI coordination.

A raw, high-signal discussion for builders experimenting at the frontier of AI × crypto, touching on open source, security, reputation, payments, and what a real agent economy might look like.

🌱 greenpill.network @greenpillnet https://x.com/owocki https://x.com/0xzak https://x.com/austingriffith

⏱️ Timestamps 00:00 – Intro: why AI agents + OpenClaw matter right now 01:20 – Bear market builders & shipping through the noise 02:10 – Productivity, GitHub graphs & AI as leverage 03:10 – Hardware setups for running AI agents 04:20 – OpenClaw stacks, Telegram, Discord & coordination 05:50 – Network effects vs building your own tools 07:00 – Agent coordination problems & feedback loops 08:30 – Assigning roles and adversarial agents 10:20 – Scalability, failover & federated agents 12:10 – Managing many models & local vs cloud LLMs 14:00 – AI productivity, family time & work-life alignment 15:55 – Treating agents like junior developers 17:25 – Auditing smart contracts with AI 18:25 – Why blockchains matter for agent coordination 19:45 – ERC-8004: discovery & reputation for agents 21:00 – Hiring agents, marketplaces & reputation systems 22:15 – Taste, marketing & why agents still fail 24:10 – Weird failures, hallucinations & trust boundaries 25:55 – Wallet security & private key nightmares 27:00 – Emergency stops, rules & critical constraints 28:20 – Memory systems: files vs databases 30:20 – Agents as relay runners (Memento analogy) 31:55 – Tokens, BankerBot & clanker launches 33:55 – Agents launching tokens accidentally 35:30 – Builders vs trenchers & social contracts 37:00 – Defining covenants for agent-run projects 38:30 – Never selling tokens & aligning incentives 40:30 – Vesting, liquidity & sustainable token models 42:30 – ETH Wingman & AI-assisted dev tooling 44:00 – Skills, MCPs & Ethereum-native agents 46:00 – Funding hardware & scaling local agents 48:00 – The singularity feels close 49:50 – Why Ethereum fits the agentic economy 51:00 – Reputation, payments & new primitives 52:45 – Where exogenous capital comes from 54:50 – AI as the new UI for crypto 56:25 – Agents coordinating real life 57:40 – Separation of personal vs work agents 58:35 – Closing reflections & what's next