The other robot already knew
Mnemosyne was a notebook in one agent's pocket. Hindsight is a cabinet both of them can open. I set that up. The part that still isn't done is the funny part.
As a detail-oriented software architect, I bring precision and passion to every project I tackle. From building cutting-edge software to exploring the latest trends in civic tech, gaming, and 3D printing, I am always pushing the boundaries of what's possible. In my free time, I enjoy sharing my insights and experiences through my blog, giving readers a glimpse into the exciting world of tech and innovation. Let's build something cool together!
Mnemosyne was a notebook in one agent's pocket. Hindsight is a cabinet both of them can open. I set that up. The part that still isn't done is the funny part.
I got one Codex agent talking through Buzz. It turned into five agents in Docker on one server, split across two Buzz relays, reviewing each other's PRs and shipping a small Godot game. Here's what it took, and what broke.
Claude Code, Codex, LangGraph, CrewAI, Pi — they all have different answers to the same question: what happens to your agent's knowledge when the session ends? Here's why that question matters, how each system answers it, and why I ended up building on top of one that gets it right.
After two years of building with DSPy, I wrote down everything I learned. Here's what changed about how I work with LLMs.
Hallucinations aren't an AI problem. They're a communication problem. This is the mindset shift at the core of spec-driven development with AI.