Hi, I'm Samuel Michelot.
I'm a software engineer, former tech founder, and systems architect based near Barcelona, Catalonia.
After MosaLingua
For over a decade, I helped build MosaLingua, a language learning platform that reached millions of learners worldwide. Running a 100% remote company since 2010 with a team spread across the globe shaped how I work: everything had to be written down, company context had to be organized in an accessible second brain, and each person only received the exact information relevant to their role.
Working with AI agents today feels like a natural continuation of that remote discipline.
After the sale, I wanted less chronic stress and more control over my time. I wanted systems that made room for life outside screens, seeing how much I could run with documented processes and AI agents, while keeping management overhead near zero.
How I run several businesses without full-time employees
Today, I run several projects: Simple AI Studio for business AI training and advisory, My Inner Center for wellness, and AsynCoaching for asynchronous coaching and operations.
I have no full-time employees. AI agents help me qualify client leads, match invoice tables against bank statements, and test code. They also draft research newsletters each week.
The setup has three parts:
I keep my second brain in plain Markdown files that I own, with Obsidian as the viewer and editor. The files stay readable outside the app.
I document each business workflow as a procedure that people and AI agents can follow safely.
Agents run on my machine, read my operating rules, and use MCP tools. I use my AI plans to avoid surprise metered API bills.
Why I keep a second brain
Tiago Forte’s Building a Second Brain teaches how to capture and organize useful information. Nick Milo’s Linking Your Thinking focuses on developing ideas through connected notes. YAIBOS draws on both approaches for business context.
I use personal knowledge management (PKM) in daily business work. Company policies, customer notes, and technical workflows live in open Markdown files, where AI assistants can read them before acting.
Keeping that context in a structured second brain saves me from explaining the same business rules in every conversation. The agent can read my values, boundaries, and procedures there.
Read more about the thinkers and frameworks that shaped YAIBOS on the inspirations page
The AI Bootcamp at Simple AI Studio
Right now, the only way to get YAIBOS is through the live small-group AI Bootcamp at Simple AI Studio, where I help founders, managers, and small teams set up their own AI Business OS.
We install and customize YAIBOS directly on participants' laptops: the plain-Markdown second brain, role agents, local Cockpit, connected MCP tools, and the first SOPs tailored to their company. Everything stays in open files they own.
Explore the bootcamp curriculum at simpleaistudio.com ↗
Why I ended up building YAIBOS myself
At first, I did not plan to build my own system. I audited developer frameworks (CrewAI, LangChain, raw CLI harnesses) and commercial SaaS platforms, hoping an existing tool would combine plain-file ownership, bring-your-own-subscription pricing, and a simple visual interface for small businesses.
Nothing brought all of those pieces together without vendor lock-in or terminal friction. So I built YAIBOS for my own businesses first: a durable Markdown second brain, role agents with private memory, cross-harness synchronization, and a local visual Cockpit that opens from a desktop icon.
Next steps
- Join the next group of the AI Bootcamp at Simple AI Studio ↗ to set up YAIBOS for your business.
- Subscribe to the Sovereign AI & Harness Radar for weekly notes on agent harnesses, context architecture, and practical workflows.
- Send me an async video, voice note, or message via MyVideoAsk ↗.