The Steering Wheel Is in the Middle

On a Saturday night in September, Jan and I watched The AI Doc: Or How I Became an Apocaloptimist on Netflix. When it ended, she said we had finally watched something thought-provoking. The film swings between terror and hope and lands where these films land: we all have to do our part to steer toward a good outcome.

I agree. But the whole time, I kept hearing a sentence I have carried for twelve months. In 2009, the biologist E.O. Wilson said: “We have Paleolithic emotions, medieval institutions, and godlike technology.”

Most people quote that line as a lament. After thirty-six years on factory floors, I read it as an engineering spec. It names three layers of one system. Each layer has strengths. Each has failure modes. You steer a system by learning what each layer does well and what it does badly, then building each one up or down until the whole thing holds.

Two of those layers will not move for us. Our emotions change over millennia. The technology will not wait. The institutions in between are the layer we can rebuild within a working lifetime. The steering wheel is in the middle.

That’s the argument.

The Paleolithic layer

Our emotions were shaped around fires, in small groups, over hundreds of thousands of years. A factory runs on them. People learn by watching someone they trust. They take pride in work others can see. They stay until two in the morning to fix a line for the crew they will face on the next shift. Belonging and apprenticeship are how capability gets into human hands. Lukas, the German tradesperson I hold up as the standard, has roughly 25,000 hours of formation by age thirty. That is Paleolithic wiring inside a structure that feeds it.

The same wiring carries fear, and the worst fear on a floor is replacement. A worker who thinks the robot is coming for his job stops teaching the new hire. He stops volunteering the fix. His biology is busy protecting him.

I know this from the inside now. Not long ago, after thirty-six years on factory floors, my job ended. The first nights brought the old questions of the fire circle. Who am I without a crew? Who will need me tomorrow?

Then came the three-in-the-morning questions. Did I not move fast enough? Should I have broken things and moved people? Did I fail to make the formation work visible in time? Was I too bold, speaking in public about formation before the work at home had taken root? Was I so absorbed in building Bloom that I paid too little attention to everything else a job asks? I don’t have the answers. I carry the questions.

Not one colleague from the floor called. What brought me back was Jan, and the seven other founding principals of our institute, people who chose this work with me.

So build this layer up where it forms people: belonging, apprenticeship, mastery others can see. Build it down where it freezes them. Bloom’s three-year employment runway, with wages rising as capability is confirmed, exists to take the fear of replacement off the table long enough for formation to happen.

The medieval layer

Our institutions descend from guild charters and double-entry books. They endure. They outlive their founders and coordinate thousands of strangers who will never share a fire.

They also see only what their ledgers record, and they wait only as long as their clocks allow. Formation takes years. Most institutions count in quarters.

When a company pays engineers to teach a robot, those hours become an asset. When it forms a person, those hours land as labor variance someone has to explain. Nobody blames the untaught robot. Everybody blames the untaught worker. Our best estimate is that companies deploy about thirty percent of the capability they have already paid for. The rest sits where the books can’t see it.

The blindness doesn’t stop at the plant gate. The same ledger that can’t see the operator can’t see the scheduler who carries a plant’s logic in her head, or the controller who knows which numbers to distrust. The CFO is Wilson’s medieval layer in person, usually a decent person working with an instrument built for another century.

Kant saw this more than two centuries ago. He wrote that a good state could be built even for a nation of devils, as long as they had understanding. He did not ask people to become better first. He asked for better institutions. He also drew the line those institutions still cannot see. Everything, he said, has either a price or a dignity. What has a price can be replaced by an equivalent. What has dignity cannot. A ledger built for price will always ask what a person can be replaced by. The institution we need asks what a person can be formed into.

Build this layer up with management accounting that puts formation on the books, the way lease accounting finally put leases there. Build down the reflex to treat people as a cost to be minimized.

The godlike layer

The film covers this layer’s power well. What I can add is how its failures look up close.

On the glass line at Lansing Delta Township, a robot lays the urethane bead and people set the glass. One night a vision camera misread an option code. A quality checker at the end of the line spotted a rear glass with an antenna the build didn’t call for. Abdelhak Radjai, then a team leader on that line, walked the line back and found four wrong back glasses. His crew cut the glass out, cleaned the urethane off each body, and ran new glass through the bead robot. Then two people hand-installed it at line height, with no winch, without smearing the headliner or the trim. Minimal downtime. Nobody hurt.

The robot did exactly what it was told. Formed people fixed what it could not see.

Technology amplifies whatever the other two layers hand it. Give it formed people and institutions that can see, and it multiplies capability. Give it frightened people and a cost-only ledger, and i