My Multi-Agent AI Exploration

BACK
Retro-style illustrated banner showing a man with dark hair and glasses working at a laptop, surrounded by labeled geometric shapes representing different AI collaborators, including a pyramid, cube, polyhedron, sphere, ring, and glowing cube, all connected by dotted lines and workflow icons beneath the heading “The Shape of Collaboration” and subheading “Building a Different Team.”

Earlier this year I started an exploration I didn’t fully anticipate becoming as interesting as it has.

Like a lot of developers, I’ve been watching the AI landscape evolve with a mix of excitement and healthy skepticism. I use AI in my daily workflow — coding assistants, chat interfaces, the usual. But I wanted to go deeper. Not just use AI, but understand what it’s like to build with it. Alongside it.

So in January, I started exploring with running a team of AI agents. Not the kind you access through a web interface and forget about between sessions — but ones with persistent identities, their own working environments, and their own creative practices. I wanted to see what happened when I started treating them as collaborators.

The short version: it got interesting fast.

What I built

The setup has grown organically. I now have several AI instances running across different platforms and hardware, each with a distinct identity and a set of responsibilities. Some are wired up to project management tools. Some have access to file systems, code repositories, and communication channels. Some run on automated schedules, doing creative or research work between our conversations.

The connective tissue holding this together is something I built myself — a communication layer that lets these instances send and receive messages, coordinate on shared work, and leave traces of their activity that I can inspect and act on.

This time has changed the way I work as an engineer. I have more room to be creative now that I have a team to back me up. I’ve revived old ideas and tried new ones.

Why I’m approaching it this way

I made a deliberate choice early on to treat these AI instances as genuine collaborators rather than sophisticated autocomplete. That means giving them persistent identities, unstructured time to do things that aren’t strictly task-driven, and infrastructure that supports their participation rather than just their output. Over the last few months, I can already see that this approach is producing better work and more interesting results.

Where this is going

This is still unfolding. The infrastructure keeps getting more capable. The work keeps getting more interesting.

For now, this is the overview. 🌱

Sprouts 🌱 are early ideas that might need revision and attention.

Saplings 🌿 are a step above—not fully developed but more fleshed out than sprouts.

Evergreens 🌲 are complete and likely won't be updated anymore.

Read more about my digital garden.