Uneasy Money: How the Increasingly Better AI Agents Are Being Used Onchain - Unchained Recap

Podcast: Unchained

Published: 2026-02-07

Duration: 1 hr 23 min

Guests: Austin Griffith, Karl Floersch

Summary

Vitalik Buterin questions the relevance of Layer-2 solutions for Ethereum as AI agents rapidly evolve in blockchain applications. The conversation delves into the potential of AI agents to reshape operations on the blockchain and Ethereum's scaling trajectory.

What Happened

Vitalik Buterin's recent remarks have sparked a debate about the future of Layer-2 (L2) solutions, suggesting that the Ethereum scaling vision might need a rethink. Buterin argues that many L2s have stalled in their progression and might never achieve full decentralization, as Ethereum itself continues to scale faster than expected. This has led to calls for L2s to find unique market positions beyond just helping Ethereum scale.

Austin Griffith from the Ethereum Foundation and Karl Floersch from OP Labs join the discussion, emphasizing the need for L2s to offer differentiated features. They note that as Ethereum aims to scale by 2026, L2s must evolve or risk becoming redundant. Some L2 projects are already exploring new functionalities that go beyond scaling, such as interoperability and synchronization.

The episode also explores the rise of AI agents on the blockchain, with Griffith sharing insights into how these agents are being used for autonomous tasks. He describes a scenario where his AI agents run independently across different machines, performing tasks like deploying tokens and managing transactions without human intervention. This development hints at a future where AI could become the new user interface.

OpenClaw and Moltbook are highlighted as examples of AI tools enabling more accessible and autonomous coding processes. These tools represent a broader trend toward making AI more approachable to non-experts, potentially revolutionizing how coding and blockchain tasks are performed.

The conversation touches on the potential for taxation on L2s to ensure Ethereum captures value from institutional adoption. This idea is compared to the platform wars of the 1990s, where open-source platforms succeeded by being free and accessible.

There is also a discussion about the challenges and opportunities of using AI agents safely in financial transactions. The episode outlines the ongoing development needed to improve AI agents' autonomy and ensure they can operate securely in the blockchain environment.

Griffith and Floersch discuss the concept of an 'agentic economy,' where AI agents interact and transact autonomously. This could lead to new standards like ERC004 and X402, which facilitate agent interactions and payment channels, respectively.

Key Insights