ROBOTS ATE MY HOMEWORK is an unusually curious room for thoughtful, creative people who want more from AI than faster output. We follow strange questions through culture, research, ordinary life, and real experiments, then see what they change about the way we understand, notice, judge, and make.

I’m Mia, a writer and AI Narrative Architect. I’m interested in how ideas, language, references, standards, and systems start to form a recognisable world around a body of work, and what AI needs to understand before it can be useful inside it.

That’s also why I’m interested in what happens when AI gets more interesting jobs than writing things for us: when it investigates, challenges, remembers, compares, cultivates, appraises, and occasionally knows when to stay out of the room.

RAMH is where I follow those questions in public. Some discoveries become essays. Some become strange experiments. And the methods worth using again may eventually become robots inside RobotsOS.

Every week, I follow a question far enough to find out whether it wants to become an essay, an experiment, a method, a system, or occasionally nothing useful at all (also allowed).

• Free posts (every Wednesday): complete essays, cultural investigations, arguments, experiments, and occasional practical methods designed to change what you notice, understand, question, or want to try.
• Exclusive posts (alternating Fridays): more thorough investigations, original methods, and builds that give curiosity something more substantial to do. Paid members also get the complete RobotsOS library.

If you need a map, three loose routes run through the archive:

Not really. RAMH isn’t a beginner AI classroom, but you also don’t need to understand APIs, build agents from scratch, or spend your evenings comparing model benchmarks.

If you’ve opened ChatGPT, Claude, Gemini, or another AI and had a conversation, you’re ready to read RAMH. To use RobotsOS, you need to be comfortable uploading a file, answering questions, and bringing a real piece of work into the chat. The robots handle the methodology. You keep the judgment.

Completely new? Open whichever AI you already have, ask it to help with one real problem, and talk back when it gets something wrong. That’s enough to begin.

A tiny note on the terminology: RobotsOS contains skills, agents, and playbooks. A skill guides you through a method. An agent takes on more of the job. A playbook gives you a complete setup, sequence, or working environment. Around here, I usually call all of them robots, because life is short and product taxonomy isn’t why you came.

RobotsOS is the applied library inside paid RAMH: a living collection of unusual AI collaborators, each with a specific responsibility, method, boundaries, and way of seeing.

There are robots that cultivate half-formed ideas, research past the obvious, pressure-test your thinking, appraise your archive, catch where the work has gone soft, and help you decide what AI should touch in the first place.

Others help AI understand more of your context, references, standards, and ways of working, so it can become more useful without taking over the whole room.

You can download them, adapt them, argue with them, or call compatible ones directly inside the AI you already use through Robots MCP. The library keeps growing as new questions and experiments survive long enough to deserve a robot of their own.

Inside the living library

Workflows, collaborators, playbooks, context systems, experiments, and working environments built from real RAMH methods. Each gives AI a specific responsibility, useful context, and clearer boundaries than “help me with this,” and each is designed to change something about how you think, notice, judge, research, or make.

And inside Paid RAMH

Original paid editions that stand on their own: deeper investigations, participatory experiments, methods, and builds that help you practise or install an idea in your own work.

Paid RAMH, RobotsOS, and Robots MCP all belong to the same membership. The paid posts give ideas more room. RobotsOS gives the methods worth keeping a durable form. MCP simply lets compatible robots come to you inside the AI you already use.

Yes. You don’t need to be technical.

You need to be curious, willing to bring real material, and comfortable having an actual conversation with the machine rather than accepting its first answer like a royal decree. Most RobotsOS systems ask questions, explain what they’re doing, and guide you through the process. If you can upload a file, answer honestly, and say “no, that isn’t what I meant,” you’re ready. Start with the quiz or ask the RobotsOS Advisor. They’ll find you a sensible first room.

You just want to try something. Right now. Here's what to do.

Already comfortable working inside your AI? You can also connect Robots MCP and ask: “How can I use Robots, and what is available to me?”

That’s your way in. You don’t need to tour every room on the first day.

You don’t need to arrive with a perfectly organised brain, a five-year plan, or a beautiful AI workflow. Bring the strange question, the half-built thing, the draft that has gone suspiciously beige, or the archive you know contains more than you can currently see.

There is probably a robot for it.

And when there isn’t, that is usually how a new one begins.