Mario Zechner explains why he built a minimal coding agent harness and discusses how AI agents are impacting open-source software.
All I wanted was a shitty coding agent that is truly mine. And I’d have loved to just tell you why and how I built pi. But then Peter decided to make it the agentic core of OpenClaw. And now pi is collateral. So yes, this is a talk about pi. But it is also a talk about how agents are destroying OSS, how I deal with that, and a plea to slow the fuck down.
x.com/badlogicgamesgithub.com/badlogiclinkedin.com/in/mariozechner/
Timestamps
– Intro and motivation for building pi
– Act 1: Building pi and the frustration with existing agent harnesses
– Why current context management in tools like Cloud Code and Open Code fails
– The importance of minimal harnesses and the "Terminal" benchmark
– Introducing pi: A self-modifying, extensible agent core
– The "YOLO" security philosophy and extensibility through TypeScript
– Examples of pi extensions (chat rooms, NES, Doom)
– Act 2: OSS in the age of "clankers" and how to fight them
– Act 3: A plea to slow down and stop the "slop" in software development
– How agents create "enterprise-grade complexity" and why humans are still the bottleneck
– Practical advice: How to effectively integrate agents into your workflow
---
Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page:
ai.engineer/talks/RjfbvDXpFls-pi-coding-agent
Join our upcoming AI Engineer conferences (see ai.engineer landing page):
- AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers
- AIE CODE in SF, November 10–12 — the top AI Coding conference returns
- AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney
---
Why it's here
4 people recommend this
What people say. Recommended for its critique of bloated AI agent harnesses and its insights on building minimal, self-modifying agent cores.
s
sbysbon Hacker News • Recommends
On “Harness Engineering for Self-Improvement”
There is a really good video by the author of pi.dev (which I have used to build some of my own harnesses): link to this video The basis of the argument is that the labs are constantly pushing updates to their system prompts that are used in claude code or codex, which are exceptionally bloated and change the sand beneath your feet with every update. By rolling your own harness, as long as you keep up with the latest advances and changes in the ecosystem, you capture a lot of the 'control' that LLM-based development feels like it strips from you. Obvious disclaimer that I use pi.dev when I am aiming for consistency, not absolute quality. Custom harnesses on pi are what I ship, claude code is still my f…
