Pi was built when there were already agent harnesses around. Here’s why Mario Zechner(@badlogicgames), found them suboptimal and built Pi, a minimalist self-modifying agent:
#1 - Mario initially was a believer in Claude Code:
"I was a believer in Claude code because they were the first that packaged agentic search up in a really compelling package. And at the time that fit my workflow really well. Everything around the LLM was kind of nice and tidy and easy to understand.
I was super happy. I was proselytising Claude code."
#2 - Reverse engineering Claude Code highlighted the degradation that Mario felt as a user:
"I personally like simple tools that are stable and that I can rely on. Even if they have non-deterministic parts, all the deterministic parts should be as stable as possible.
That was just not the experience with Claude Code around summer 2025. They would take away your control of the context. They would inject stuff behind your back, which is bad. Then, your workflows stopped working because there's now a system reminder that you don't even see in the UI that would modify the behaviour of the model. They would also do this to the system prompt.
I built a little service where I can track the progression or evolution of the system, prompt and tool definitions and, with every release, it was messing with stuff.
That just messed with my workflows and I don't appreciate that."
#3 - PI was built with an appreciation for simple and reliable tools:
"If I commit to a development tool, I want it to be a stable, reliable thing like a hammer. I don't want my hammer to break a different spot every day. That's terrible.
We need somebody who goes the full velocity kind of way. But I don't want to work with a tool like that."
AI agents handle the boring stuff—so I never have to install a 300MB PDF editor or reader again (google chrome reader); self-hosted BentoPDF for the impatient pdf edit glory👉 https://t.co/b0oVTW3RK4
AI agents are great for the boring things. I always hated to install the bloated Adobe Reader 300+MB. @BentoPDF to the rescue with a local selfhosted install on my static site https://t.co/b0oVTW3RK4 . To be honest not much AI really used or needed.
Part II the Swimlane Complexity Grows I’m not chasing AI hype — I’m stress-testing what real hybrid development can actually deliver one commit at a time.
https://t.co/WV1TaIjuvk
Its enlightening to see a slow but steady trend towards more countries "doing" something for the future with AI. Go Europe, Go Japan go go go
https://t.co/mbcT4LtOWR #ai#future#europe https://t.co/wzitgQv02Z
Your time is something everyone needs to respect, and respecting the time of others should be on everyone's mind. This post just reminded me of another cool way to do just that "respect the time of others" by being B.R.I.E.F nice. https://t.co/kgSAjjFTbb
I have the privilege to speak at the #IRF International Road Federation, on behalf of #Strucinspect explaining the role of #AI in digital inspections. Tomorrow 14.07 18:00 (CEST) Vienna, 9:00am (PDT) Vancouver, 12:00 (EDT) New York. Register FREE https://t.co/ImIAtcznLf
Learn how to stream real-time simulation and training experiences from @Azure (powered by #NVIDIARTX) to HoloLens 2, using #UnrealEngine with @CesiumJS high-fidelity 3D terrain at #GTC21 on April 13.
Registration is free:
https://t.co/j1HrzREVrU