What Is All This Intelligence For?

A note before this starts: it opens with the day my son Atticus was stillborn. It's the hardest few paragraphs in this whole series, and I didn't want to write them gently. If you need to stop and come back later, this will still be here — and so will Part 2.+ + + + + +

On June 10, 2021, my son Atticus was stillborn.

There are sentences that divide a person's life into a before and after. That's certainly one of mine.

I remember the ultrasound — how still he was. I remember how quiet the room became as we realised the hell we were entering. I remember my wife's face, staring at that screen, and how badly I wanted to look away. I remember helping her into the next room. We couldn't talk.

Then our doctor, whom we'd met for the first time that day, came in and told us that the worst thing we had ever lived through was about to get worse, because our son still had to be born.

I remember trying to be present for my daughters as they realised they'd never meet their baby brother. I remember shutting myself in a closet at some point that day, crying tears from a place I didn't know existed. Something in me was breaking.

That kind of pain changes your relationship to the future.

For a long time I thought I was grieving Atticus. I came to understand that I was grieving possibility itself — the conversations that would never happen, the version of fatherhood I would never meet, the parts of me that would only have come alive in relationship to him.

Something else happened too, slower and a lot more difficult to name. In the weeks and months after, I started noticing something in ordinary rooms — meetings, checkout lines, conversations in the hallway. I began to see that almost everyone I met was carrying something, whether they said so or not — a different shape than mine, but the same kind of weight. I couldn't unsee it after that. It changed how I listened. It changed what I thought a conversation was for.

In the years that followed, grief, deconstruction, and AI all began to converge. I kept telling my students to be brave, knowing I was still organising too much of my own life around safety. Eventually I understood that my daughters weren't only listening to what I said about courage. They were watching what I was willing to build with mine.

Around the same time, I began to see AI not only as a tool but as a question about the future itself: what kind of intelligence are we building, and what kind of people will it help us become?

Humanum Huis began at that intersection: grief, imagination, responsibility, and a future my children will have to live inside. It also begins with the simple belief that this intelligence should serve life. And it begins with a warning that AI does not tell us who to become. It shows us who we already are.

The AI future has been branded. But the human dream has not.

This is a founding document. A reckoning, yes, but also an invitation. And it is for the leaders, builders, educators, parents, artists, and students who feel the AI conversation getting louder and thinner at the same time — who suspect something essential is being skipped, but haven't yet had language for it.

Here is some language. By the human dream, I mean something simple and difficult: that pain at scale should go down, possibility at scale should go up, and intelligence should serve life instead of merely accelerating what already exists. A future worthy of the people who have to live inside it. A dream isn't serious until it changes what we build, measure, fund, teach, reward, and refuse.

That dream isn't being told. And silence, here, isn't neutral.

The Dream No One Is Selling

Look at what we've been sold. Speed. Productivity. Disruption. Capability. Inevitability. These ideas have extraordinary brand power: the confidence of a category that already knows how to sell itself. The race to build and profit from AI is being narrated with tremendous confidence and almost no transparency about what the dream actually is.

Now I really do want to be fair about this, because the people building these systems have slowly started answering the question, some of them at length and with real feeling, and their answer is mostly a story about what the machines will do for us, cured diseases and abundance and time given back. Lately the same people have started arguing in public about how fast the machines should be allowed to go. What I still haven't heard with anything like the same force is the other half, the half that depends on us, what we will have to see, decide, practice, and refuse so that any of it arrives as something better than acceleration. I have heard almost no one ask how fast people are becoming ready, and that is the pace I care about most.

Every day that dream goes unbranded, the other one — dreamless acceleration, power without purpose, capability without conscience — consolidates more cultural territory. The ideas that protect human flourishing have been out-branded by the ideas that exploit it. That's not inevitable. Marketing problems can be solved.

When the human idea loses, the cost isn't theoretical. It shows up in classrooms, companies, homes, hospitals, childhoods, and nervous systems.

The future will not become humane simply because it becomes intelligent.

The Mirror

Here is what I have seen watching organisations explore AI adoption: AI reveals who they are very quickly, and then it makes more of whatever it found. The incentives they reward, the voices they suppress, the futures they can and cannot bring themselves to imagine — all of it becomes visible. An organisation built on fear, for example, will tend to use AI to watch and automate. An organisation built on curiosity and care will tend to use AI to listen better and build things worth building.

A mirror can distort. I know that now. The questions we ask shape what becomes visible. The conditions we create shape what people are willing to say. And the person holding the mirror is never completely outside the picture. So the test is not whether the first picture is perfectly true. It is whether the people inside it can look at it, argue with it, correct it, and make it truer.

That is why I have started thinking of AI as a kind of moral exposure event. Used carefully, it can make parts of an organisation's hidden operating system easier to see. Not because it introduces new values into an organisation, but because it removes the lag between what an organisation claims to value and what it actually does.

Most organisations reach for efficiency first because it is the easiest purpose to imagine and the easiest to measure, and it lets an institution move without ever saying out loud what it is for. I don't want to pretend efficiency is small, because an afternoon given back to a nurse or a teacher matters. It just can't be the only question in the room. AI shouldn't only help us move faster. It should help us become clearer about where we're going, and honest about whether it is worth going there.

The organisations that deploy AI without first asking what they are for will simply become faster versions of what they already were. And some of what they already were, we cannot afford to accelerate.

And once AI starts amplifying what is already in a place, you learn quickly that everybody is protecting something. The argument is never really about the technology. It is about what people are afraid of losing if the picture changes. That's why this work can't stop at an honest picture. A group needs something out ahead of them that is worth more than whatever they are currently defending, described well enough that they can argue with it, and when that exists people stop guarding the old way of working, because they finally have somewhere else to put their loyalty.

The Question Beneath the Question

I grew up inside systems that taught me certain versions of power: religious power, masculine power, institutional power, credentialed power, the quiet power of belonging in rooms where others have had to fight harder to be heard. For a long time I moved through those doors without asking why they opened. I benefited from arrangements I now want to change. I've participated in the things I critique. I don't have clean hands. I've mistaken access for merit, certainty for truth, and charisma for care more times than I want to admit.

I'm not telling you this to perform accountability. I'm telling you because the work I am describing requires a kind of honesty that has to start somewhere, and the only honest place I can start is with myself.

The question underneath my life and my work has always been some version of the same thing: what is this power for, finally? In a moment when human beings are acquiring the most concentrated technological, economic, and informational power in history, that question feels less like philosophy and more like a test we cannot outsource. It's also the question most institutions have learned to postpone.

From Grief to Architecture

After Atticus, possibility stopped being abstract. It became embodied, fragile, and urgent. Grief doesn't make me right, but it does explain why I cannot look away.

Nothing I build will replace what we lost, and I don't want anyone to read this work as an attempt at that. What I can do is let the possibility I lost become a way of making possibility somewhere else, for someone else, and that is what this work is trying to do.

The sentence I gave you at the start, that pain at scale should go down and possibility at scale should go up, became the design brief for everything we build. I'll admit that for a long time I wanted it to harden into a framework neat enough to fit on a slide, and every version I built made the work look more finished than it is. So what I carry now is a handful of questions I bring into every room where AI is entering a decision.

Whose burden does this lift, and whose does it quietly add?

What does it help people see that was hidden or misread before?

What does it let them imagine that they couldn't imagine yesterday?

Are the people inside the system coming out more whole or more used?

And what, finally, is all this power for?

Those questions don't come in an order, and none of them outranks the others. But an organisation that can answer only one of them has barely begun to ask what it is doing. That is as close as I can come to saying how grief became architecture, and I expect the architecture to keep changing as people argue with it.

+ + + + + +

This is Part 1 of 4. Part 2 picks up where efficiency stops being a good enough answer on its own. →

CONTINUE TO PART 2

Let's explore Intelligence

If something here is already naming something you're carrying — about your organisation, or about yourself — the first conversation is free, thirty minutes, and it starts with a question, not a pitch. →

BOOK A CONVERSATION