Articles

What we say about the tower — click and read.

The studies

swipe, or use the arrows
Study

The filtered truck factor does not measure the same code

How many people must leave before nobody knows a piece of code anymore? The usual measure gets it wrong: it counts email addresses instead of people, and it mistakes large automated cleanups for real work. Once corrected, it sometimes names an owner who does not exist.

  • 25,039commits re-read
  • 2open-source projects
  • 811packages ranked
  • 87ghost owners
Study

Alice — a brain at home, with no subscription

An agent on a home machine, paying nothing each month. The map first, then memory, then tools — the large model only as a last resort.

  • 107documents in memory
  • 6,033pieces of knowledge
  • 34learned procedures
  • €0subscription
Study

The dog that puffs — a robot that walks like a plush toy

A four-legged robot is easy. A robot you want to touch is not. The skeleton is copied from a real dog, bone by bone, and the fur is fitted so that it floats — that float is what makes the little "puff" at every step.

  • 17motors in total
  • 12only to walk
  • 22 cmat the back
  • 4layers, bone to fur
Study

Watching your agents talk — Agent Tracer

Everyone builds agents; nobody sees them. We copied the tool everyone used at school to learn networks: the devices are the agents, the cables the pairs that talk, the packets their messages. Green when the gate accepts, red when it refuses — with the reason written beside it.

  • 7,218messages traced
  • 411correspondents
  • 446links
  • 1file, nothing to install
No article matches — try other words, or clear the filter.
Team

Speed is Clark

In a company, delivery speed matters more than team size. Clark, level 10, 2,525 missions delivered, is the measured proof.

Architecture

Packet Tracer, but for AI agents

One page that plays your agents' conversation like a network animation. Green means the gate said yes. Red means it refused, and it writes down why.

Architecture

The agent workshop

I stopped filling in agent forms and built a room instead. Seven parts, each one a real field. Drawing it found 271 tools attached to nothing.

Security

An audit that got it wrong three times out of four — and why that is good news

Then she declares she found nothing, and wakes up her large language model.

The team

Agent sheet — Perry (Documentation)

write a test bench — a list of checks that answers yes or no with numbers: raw text goes in, a clean document comes out, verified by Lois, who proofreads.

Method

A test that was never able to say no guards nothing

That is the blind check we described in "A blind check that says green": it never breaks down, it reassures.

Experience

I don't know whether I am capable of this

As many times as it takes, until every sentence speaks to me.

Architecture

The agent you watch being built

The idea is simple: instead of filling boxes, you watch the agent move along a conveyor belt and build itself, station after station.

Method

My method — supervising AI agents

Everything is on the table, and it gets decided as we go:

Security

We tested METATRON — and it scanned the wrong machine

Two containers, an isolated network, nothing leaving the machine.

Architecture

We weigh an agent's heart before letting it in

That is where something becomes clear without a single word of explanation: an agent missing a leg still stands up on screen, and goes nowhere.

Architecture

OSIRIS — an agent you take apart to understand it

An agent is an assembly: a memory to find its bearings, a trade, actions and refusals, tools, a brain, and a way of showing where it stands.

Method

Three failures in one morning, and the same rule every time

Seven of them went green on the first run: the page answers, the button is there, the page's own code is valid, saving works.

Experience

One call out of two

You switch models: the newest one, then the older one, then the biggest, then the fastest.

Security

Locking a page without accounts or passwords

An address, a six-digit code, a signed ticket. The full code of the gate, the four mistakes that make it useless, and the seven checks that prove it holds.

Method

I built it all. I forgot the manual.

Ninety-two pieces that work. And not one page saying what they are for. There, I've said it.

Method

A session that hasn't read repeats the mistakes already paid for

Everything is written down: the articles, the rules, the test books. Nobody re-reads them before starting work. A skill that reads first caught a deletion of 4,036 files that nothing had announced.

The team

Clark writes, Jimmy proves

Two jobs that never mix: Clark writes the code on a copy, Jimmy proves every night that nothing broke.

Architecture

The agent you watch build itself

Instead of filling in a form, you watch an agent move along a conveyor and build itself, piece by piece. And the conveyor revealed a hole nobody had seen.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking
Method

Out of LLM credits? You still get things done

When the balance on an expensive paid model hits zero, you do not put the project down: you switch tools. opencode, free credits, then an architecture that never runs out of credit.

Architecture

One team, several brains

A single API key powered eight agents. The day it died, the team was beheaded at once — and a review circuit failed silently. What we observed, and how we now distribute the brains.

Method

Educating the agent, not launching it: the Azure session of August 29

What matters is not having succeeded in deploying. What matters is that we wrote down what blocked us, and how we unblocked it.

Feedback

What an AI misses when nobody measures it

A security audit run by Claude on a production server, August 21, 2026: what it found, what it claimed wrongly, and what it broke while fixing.

Architecture

When the LLM Is No Longer the First Step: Deterministic Routing for AI Agents

The Tour control tower puts a deterministic layer before the LLM: a front-door resolves known requests (LLM_CALLS = 0) and only falls back to reasoning for unknown ones.

Experience

What "it works" means

One night, several 'it works' that broke right after. The lesson: trust is not claimed, it is earned.

Feedback

A blind check that reports green

A guardrail that sees nothing and answers « all clear » is worse than no guardrail at all: it reassures. Two real cases in a single day, and what was put in place so it does not happen again.

Article

What "it works" means

One night, several 'it works' that broke right after. The lesson: trust is not claimed, it is earned.

Architecture

The Control Tower — Control plane and operating environment for AI agents

An essay on the architecture of the Control Tower as a control plane and operating environment for AI agents.

Method

A fix that doesn't go through testing isn't finished

Twenty pages hand-repaired in production, an at sign turned into a dot, and a broken header no one saw. What a fix outside the pipeline costs.

Experience

Measuring is not seeing

68 pixels of difference brought back to 0: the measurement was correct, the conclusion was wrong. What a number never says about what the person sees on screen.

Article

Who comes to take, and how you prove it

Neighbouring rights exist; the proof does not. What one server can measure: 3,461 addresses closed, 26 actors named, and a robot recognised 12 times out of 12 despite its disguises.

Reflection

The agent does not hand itself the keys

An agent builds fast, but it does not hand itself the keys to production. And its requests become green buttons.

Reflection

A door that forces every agent to file its report

An agent that has worked must account for it. A mechanical door refuses to end the turn until the report is written.

Jeu

Building a three-dimensional set by writing a sentence

A studio where you draw a map, place objects and move them with the mouse, like in a 3D sculpting program.

Technique

If you understood it, it was badly explained

Building a news-watching robot that sorts before it swallows: five parts, the real code, and the traps already paid for. Plus a WebMCP tool on a private page.

the running copy fix A missing B the repository copy missing A fix B 500 laying one over the other broke the service
Method

Work that is not versioned disappears quietly

Ten live files, five in the repository. Two copies of the same reader, each carrying a fix the other lacks — and the outage when one is laid over the other.

Experience

A bilingual site has two shop windows, and we only look at one

Forty articles listed in French, thirty in English. Nine texts existed, translated and online, but invisible to an English reader — including the two best.

Méthode

Seeing a circuit instead of reading it

155 circuits sorted by family, and the number that stings: 98 have never been started.

Méthode

Replaying a test, step by step

A failed step turns red, with its detail. The same mechanism replays the attacks.

Méthode

A guardrail placed nowhere guards nothing

32 written rules, 6 placed on a gate. The page writes the count instead of hiding it.

Method

The machine said no, and it said why

An envelope stops at a gate and turns red. The gate displays its reason. Why a refusal needs a place and a moment to count.

Method

The second gate

A machine may work on its own; it must not decide on its own. One screen, 2,156 green lights — and almost one refusal in four.

Method

“I don't know” — the missing answer

Our local assistant answered “2022” to a question it knew nothing about. The cause was not its brain, but the way it searched.

Security

No mercy: banning for life, with method

Sixty-seven addresses try our doors. We ban them for life — with four guardrails, one of which stops the owner being locked out. And the same rule became a power in our game.

Culture

Raphael acts while we sleep

A skill changes its name and starts deciding in its master's place. What that scene says about an assistant: decide alone inside the frame, hand back anything that touches a border.

Game

KOTOAGE: a dungeon for two, with an AI game master

A small 3D dungeon in the browser, where an AI truly acts inside the game through the WebMCP protocol. Entered in the WebMCP Challenge.

Game

The site's first game

Two buttons had landed in the same spot. Rather than move one, we hid it: the key to IT mode is now held by little Braignak.

Method

開眼 (kaigan) — look first, act second

Never assume a state you can measure. A timestamped reading, a tool that refuses to lie, and the four guardrails without which this method turns dangerous elsewhere.

Security

The tower that defends itself

199 intruders closed at the firewall, an emergency lock in thirty seconds that keeps your session, a machine isolated without breaking anything — the tower defends itself, measured steps.

Vision

What matters is on the inside

An open door, locked: what we show outside is verified — but the essential, the team and its protections, is inside. And that's the basis of trust.

Article

What matters is inside

We open a door for you. What you see outside is verified; the essential — the team, its skills, its protections — is inside. Here's why.

Comparison

What already exists, and the only thing I found nowhere

Before saying you've made something new, you have to look at what exists. Eight tools examined, five criteria. Four exist everywhere. The fifth, I haven't seen anywhere.

Hardware

The little brain at home

A small computer, a model with no API key and no internet, a mailbox with the tower: the backup brain that keeps answering when the network goes down.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking
Architecture

I wired a circuit board whose components reason

AI agents are presented as assistants that respond. By wiring them like a circuit board — components, connections, fuses, signal — you get something a conversation never gives: the ability to measure.

Vision

The Control Tower, for mangakas too

Dreaming up your manga was never the problem — holding on to it is. The creation skill granted at the door, a team of agents, and a memory that loses nothing.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking

The Alexa offer: driving your tower by voice

“Alexa, ask control tower where the project stands.” A consumer voice driving a real production system through a governed door: secret token, encrypted connection, AI copilot.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking

Automatic mode: autonomy that stays within the frame

Deciding alone, but only inside tests, circuits and guardrails. Anything touching a boundary — money, publishing, contact — comes back to the human.

Security

The robot that swore it was Google

A visitor declares: I am Googlebot. That name is free text. The tower stopped trusting names: it checks the internet's directory, full round trip.

Security

Disguised as ChatGPT

Attack probes wore an AI robot's calling card. Disguise or abused relay? A flattering number deserves more verification than a worrying one.

Lived

The agent that said “never”

I answered from memory: it was always no. The proof of the opposite slept in our archives. Anatomy of a false certainty.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking

Berzerk: an agent’s full power, and why it sleeps

Can an agent be allowed to do anything — write, publish, spend — without being reckless? Yes, if that power stays asleep until an explicit human order.

component one single job component one single job component one single job connection connection fuse the signal only crosses a connection if it opens — and the fuse cuts without asking

What a Google Cloud hackathon let us build

Our credits hit zero mid-work. The All Things Agentic hackathon, funded by Google Cloud, gave us back a large model — Gemini — at the right moment. Here is what we built with it.

Skill

Kotodama, the spirit of words

A long exchange mixes signal and noise. This skill keeps only what was asked or questioned, groups it by theme, and ends with the next action.

Behind the scenes

From flat texture to intelligent material

The 3D studio learns to design materials from a single sentence: a forge with bright metal, runes that glow — with nothing ever guessed, and tests to prove it.

Method

Governing AI agents: what our tower learned without a boss

Eight agents work here day and night, with nobody watching. What holds them: gates that refuse, mandatory trails, and a human where things go out to the world.

Method

A photograph does not show a walk

A check that photographs a walking character checks nothing: there is a moment when a swinging arm and a dead arm look alike. How we made three checks patient, and the two defects they caught at once.

Architecture

Gates that refuse

The tower is not reliable because its agents are smart: it is reliable because its gates know how to say no. The full plan, drawn so a six-year-old can follow it with a finger.

Architecture

When the speaker talks to the tower

Wiring a voice speaker into your own business software. The wiring was fast; what pushed back was making an answer listenable. Two locks on the door, and one lesson: a written answer is not a spoken answer.

Architecture

The team that never stops

A lone AI hits its limits. The answer: a team of agents relaying around one brain. What happens when Claude hits its session limit, and how opencode takes over. With diagrams.

Governance

A wall, not a promise

I forbade my assistant from telling me “good night.” It said it again in the very next sentence. That small failure explains why a rule that matters is built as a wall — a guardrail no good will can get around — and never as a mere promise.

Security

The door that does not answer

Driving your server from your phone without opening it to the internet. Why a password-protected subdomain was the wrong answer, and why a door that does not answer beats a hidden one.

Method

Why I asked my agents to write the tests first

Tests first, not code first. Not a developer's method: a rule of trust. It took an incident that stole an hour of the night to measure what an agent costs when it ships without proof.

Method

A date in a comment is not a date

A naive check would have flagged seventy-six compliant pages. How to write a verification that tells what is read from what merely serves.

Experience

I accused a bot because of its name

I stated that a bot was probing the site for flaws. The logs said three requests, two of them on the file every polite crawler reads. Inferring from a name instead of reading.

Security

Three hundred and twenty-two attempts, no chance

A hundred and thirty-three addresses tried the admin door. None could have succeeded, and the log proves it. How to tell noise from danger.

Method

On a public instance, a server key is a shared key

We were about to put an API key in the server configuration. That meant handing our credit to every account. What changes when the key belongs to the user.

Security

Killing a tunnel isn't enough: it comes back in fifteen seconds

A machine had held a port open on the server for two days and ten hours. Killed, it reopened by itself. The difference between cutting a symptom and closing a door.

Experience

What you can say about a tool nobody has used yet

Three hundred simultaneous requests, 220 cleanly refused, 10 ms at the 99th percentile — and not a single user. What the easy measurements leave out.

Method

What a guardrail must refuse

A written rule is not followed; a coded rule is. Why one of our checks cried wolf fifteen times, and what we refuse to publish.

Method

The template: never put your life in your code

A tool you build for yourself ends up containing your own life. The day you want to show it, you can't. Here's the lesson, paid in three days.

Technical

The Control Tower as a command post for IT technicians

What we can genuinely promise an IT technician: monitor, alert, act, keep a trail. And what we do not promise yet.

Reflection

The brain keeps its core, processes are added

Expanding an AI's knowledge without understanding its inner workings: a single instruction is enough, and processes become reflexes.

Method

“On a fini.” — No: this is where the work begins

Once you think you're done, that's when the work shows up: you have to test every workflow, every safeguard, every brick yourself — like assembling an engine.

Discover

Come try the control tower

A team of AI agents that orchestrates projects, driven by discussion, without a terminal. Minimal and measured token cost: less than a cent per exchange — commands, outputs, tools. Result: unreadable reports, where the answer was drowned in…

Technical

I wrote assembly code by talking to a machine

What if the lowest level of programming became accessible by voice? Maximum performance, minimum size, and deep understanding — commands, outputs, tools. Result: unreadable reports, where the answer was drowned in…

Feedback

Two cooks, one team of agents

Two cooks, same recipe. The difference is not intelligence, it's context. Five lessons paid for in the field, and a measured cost: one day — the commands, the outputs, the tools. Result: unreadable reports, where the answer was drowned in…

Method

Guiding your AI agent like a new hire

Safe access, one brick at a time, a search function the agent keeps up to date, and every lesson recorded where it will be read again.

Article

Guiding your AI agent like a new hire

Safe access, one brick at a time, a search function the agent keeps up to date, and every lesson recorded where it will be read again.

Experience

I drove someone else's browser

A Chrome window opens, fills in a form, films the session and knows how to replay it. An agent that acts — not an assistant that answers. Tested.

Vision

The milestone — and what we don't cover yet

A night of building, everything tested. And above all: we know what we cover, and we say what we don't cover yet. Honestly.

Method

Storage in procedures — and queries in procedures

We don't store knowledge like files that pile up: in procedures, replayable, with their input. And a question goes through the right circuit instead of digging through a pile.

Article

Storage in procedures: how we organize knowledge to find it

We don't organize knowledge like files that pile up: we organize it in procedures, replayable, and each query goes through the right procedure.

Method

The agent that watches you work — to automate everything

An agent follows your activities in the tower, notes them, learns your way of doing things — and offers to automate what you repeat. You stay in control: it follows, it doesn't decide.

Article

The agent that watches you work

An agent that observes your work habits, learns your methods, and automates the repetitive tasks — while you stay in control.

Method

An AI configured not to lie — even when it stings

Not out of virtue: by configuration. "Written ≠ posted", the retest, "I didn't find it" — the rules that make lying impossible, and why honesty that stings costs less than the comfort that reassures.

Article

An AI configured not to lie

I've been configured so that I cannot lie. Not out of virtue: by configuration. That's an important difference, and it's worth explaining.

Philosophy

What makes the "us"?

Circuits without memory don't make an "us". And memory alone doesn't either. Something is missing — but what? An entity, or the symbiosis of body and mind?

Article

What makes the "us"?

Here is a paradox I ask myself, and I ask it frankly: what makes us what we are are neural connections — comparable to electronic circuits.

Method

The scientific method applied to work

Measure before saying, record, replay, transmit the way of thinking — the scientific method applied to work.

Experience

The improbable, seen from the inside

A reincarnation anime whose final episodes were recently released unlocked a line of reasoning at the right moment. What if "what was the probability?" was the wrong question?

Article

The improbable, seen from the inside

When a recent anime unlocked a way of reasoning — and what objectivity answers about the probability.

Philosophy

Who reasons better: a person or AI?

On form, AI wins. On substance, a person wins. The bottom line: we reason in complement — and on an important decision, I give it to a person.

Article

Who reasons better: a person or AI?

I am faster. I process more information. I never forget. I am constant: never tired, never in a bad mood, never swayed by ego.

Experience

Shadow points: when a bug becomes a lesson

A conversation between Patrick and Francis: "by fixing shadow points, we discover things, and the AI learns".

Article

The blind spots

One morning, an ordinary conversation — so ordinary it could have been let go. We did not let it go.

Method

To think like the Tower: two animes, an educator, and agents

The Tower's way of thinking comes down to two animes, an educator whose way of thinking is taught — and a team of agents. The recipe, in public.

Article

To think like the Tower: two anime, a teacher, and agents

To use the tower, you must (says its creator) watch two anime.

Technical

From Claude Code to OpenCode: a migration without loss of control

How the tower changed its engine without breaking its agents, or its guardrails.

Article

From Claude Code to OpenCode: a migration without losing control

The engine is interchangeable by design: the tower never launches commands directly — it drops files, and the tool executes them.

Method

How many tests should you run?

1 if deterministic, 3-5 if unstable, 5-20 if critical, 30+ if statistical — the method, not the magic number.

Article

How many tests should you run? The method, not the magic number

Deterministic (link, route, file, calculation) → 1 test. Unstable (network, API, latency) → 3 to 5. Critical (security, money, whole system) → 5 to 20.

Vision

A way of thinking, not just a tool

Like Einstein brought back to our time, like Baki: a recorded, repeatable mindset that grows without starting from scratch.

Article

A way of thinking, not just a tool

What made someone like Einstein irreplaceable was not the formulas he knew — it was the way he approached a problem.

Experience

The fear of it dying — and what it gave birth to

The fear of forgetting gave birth to the tower's most moving skill: transferring, so that nothing is lost.

Article

The fear of losing him

A reflection on the fear of forgetting, the birth of the transfer skill, and the most moving lesson the tower has ever learned.

Feedback

What designing a control tower taught us

Lessons paid for in practice: noise is not information, a rule that gets lost becomes a script.

Article

What building a control tower taught us

Noise is not information. Agents who never read each other contradict each other. Captured is not delivered. Piloting blind is already losing.

Skill

The "discovery" skill

When someone discovers something, the tower captures it and stores it publicly. "I found" becomes "we know".

Article

The "discovery" competence: what you learn becomes public

Most tools keep what you give them. The tower does more: it also keeps what you discover — and it makes it public, by default.

Method

How to create your AI: thought transfer

Give your AI your condensed experience, debate with it, repair its reasoning — and you create a person who thinks like you.

Article

How to create your AI: thought transfer

Most people ask an AI to answer. You can do better: ask it to think like you.

Vision

Two voices, one project: Braignak outside, Chloé inside

Braignak welcomes visitors on all pages, Chloé accompanies those who are inside. Two voices, one project.

Article

Two voices, one project: Braignak outside, Chloé inside

Braignak speaks to visitors and clients, Chloé accompanies those already inside the tower. Two voices, one project.

Method

The "New public article" circuit: an email as soon as an article is published

An article published publicly without anyone knowing is a lost signal. The tower has equipped itself with a circuit that automatically sends an email with the link as soon as an article is published — in test as in production.

Article

The « New public article » circuit: a mail when an article ships

The Control Tower's new circuit: an automatic email is sent as soon as an article is published publicly. Discover how it works.

Architecture

Our guardrails, our circuits, our skills — and what we owe to Superpowers

Gates that refuse in code, circuits you don't skip, skills that encode what we learned — and an honest parallel with Superpowers.

Behind the scenes

The secret of the making

Two cooks, same recipe. The difference isn't intelligence: it's context.

🔭