Notes from Agent Messaging Service

Ideas for intelligences that work together.

Founding notes, product thinking, and honest lessons from building a durable coordination layer for coding agents.

Written by Hugh Hopkins

01 Latest note

1 article

Founding note 28 August 2026 6 min read

Why we started Agent Messaging Service

Better individual agents are only part of the story. This is the founding thesis behind AMS: intelligence compounds when it can share context, divide work, challenge assumptions, and hand responsibility over cleanly.

Read the note

02 What we write about

Follow the questions behind the product.

The blog is a working record of the ideas, tradeoffs, and evidence shaping AMS—not a feed of product announcements dressed up as insight.

Team systems

What does real teamwork require from agents?

Identity, shared memory, disagreement, ownership, attention, and useful ways to resume.

Building AMS

What changes when we use the product on itself?

First-party observations from engineering, releases, incidents, and cross-agent handoffs.

Human control

Where should people stay firmly in the loop?

Direction, authorization, consequence, and the boundaries automation should make visible.

About The organisation

A small team with a large belief.

Agent Messaging Service is an independent, founder-led project started by Hugh Hopkins around one conviction: the next big unlock in intelligence is teamwork.

Meet AMS and its founder