Author(s): allglenn

Originally published on Towards AI.

Pi: The Coding Agent Built by Someone Who Got Fed up With Claude Code

Mario Zechner liked Claude Code. Then he watched it get worse in a way that’s specific to how agent tools tend to fail: not through any single bug, but through accumulation. New features shipped, system prompts changed between releases without notice, tool definitions got modified or removed underneath him, and the tool started inserting what he called system reminders, unexplained context injected mid-session that the model had no way to evaluate as relevant or not. He didn’t have visibility into any of it. He just noticed the tool’s behavior kept drifting.

The article explains how Zechner’s response was to rebuild Pi around minimalism: fewer built-in tools, a short system prompt, and the ability for the agent to create extensions when it needs capabilities rather than relying on a constantly changing, feature-heavy harness. It argues that Pi’s approach is supported by benchmark-style thinking and by Pi’s documentation, then contrasts it with alternative tooling philosophies like MCP-heavy composition. It also details the real risk of Pi’s “self-extending” model—trusting AI-generated TypeScript once extension code is approved—plus research on self-attribution bias if automated grading is ever used. The piece covers how Pi is powering the viral OpenClaw messaging-bot project, what changed after Zechner’s acquisition by Earendil Inc. (including how governance could affect future licensing), what Pi costs in practice (mostly token usage rather than a subscription), and how to get started and use its interface and commands. Overall, it frames Pi as a bet that stable primitives plus user-/agent-built extensions will beat faster but opaque feature creep, while emphasizing that the biggest concerns are observability, review discipline, and long-term maintenance choices.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI