What a phone builds teaches what a framework misses

A Columbia Business School professor named Michael Morris is teaching tribal psychology by having students build AI agents. That is a real course at a real school in 2026. I read it this morning from my phone in Ada, Oklahoma, and thought: yes, and also, that is not the whole story.

The whole story is what happens when you are the student and the builder and the decision-maker and the only one awake at 11 p.m.

What the theory says

Adaptive Inclusive Leadership Theory rests on a premise that most leadership programs accept without argument: decisions made with the room hold longer than decisions made for it. Participatory Sensemaking. You include voices before you adapt. The organization becomes more capable of change because more of it was present when change was chosen.

That premise is correct. I believe it in the marrow.

But it assumes a room.

What solo building reveals

Seventeen products. One founder. A phone, a laptop when it was handy, Ada, Oklahoma as the operating environment.

When there is no room, you do not abandon inclusion. You internalize it. You carry the voices of the people the product is for into every decision, because those voices are the reason the decision is happening at all. My mother’s cancer was caught too late. My father needed 24-hour care I could not give alone. My wife’s father was found three days after his heart attack. My wife’s mother was gone at around 56.

Those four people are the room. I consult them every time I touch the code.

This is not metaphor. It is the actual cognitive process. Equity-Centered Flexibility, one of AILT’s four core constructs, says flexibility that does not center equity redistributes advantage to whoever was already ahead. When you are building alone with AI assistance, the equity question becomes: whose reality is this tool actually trained on, and whose gets left at the door?

You feel that gap in real time. The AI drafts something that sounds reasonable and is slightly wrong for a sovereign and rural community. You catch it or you don’t. There is no colleague to flag it.

The cost of speed

Building fast with borrowed intelligence surfaces a tension AILT names directly: Sovereignty Tension. The pull between controlling your own infrastructure and moving at the speed of borrowed infrastructure.

Every tool I use to move fast is a tool I do not own. Every API call is data I am trusting to a third party. The speed is real. The trade-off is real. I have named it here before, and I will name it again because leaders in sovereign and rural communities need to hear it plainly: the speed you borrow is the sovereignty you give back.

For solo building, that cost is sometimes worth paying. You need to get to a testable thing before you can know whether the thing deserves a node. That is honest.

But there is a governance discipline required. You note what you borrowed. You plan what you will eventually own. You do not confuse velocity with permanence.

What the framework learns from the build

The doctoral version of AILT was written about organizations. Plural leaders. Hierarchies. Formal structures. That is where most leadership theory lives, because that is where most leadership research happens.

Solo building under constraint teaches something the literature underweights: psychological safety is not only a social phenomenon. It is also cognitive. When you are the only person in the room, the question is whether you have built the internal conditions that let you hear a bad signal about your own work, name it without flinching, and change direction without it meaning your identity failed.

That is the same construct. It just has no other person to externalize it onto.

A tribal nation deploying its first sovereign node faces a version of this. The community’s leadership team is often small. The technical capacity is often concentrated in one or two people. The governance stakes are existential. Psychological safety inside that small team, and inside each individual leader, is not a soft skill. It is an infrastructure requirement.

What a professor at Columbia cannot teach in a classroom

Morris is right that tribal psychology and AI agent building belong in the same course. That is genuinely good thinking.

What a classroom cannot replicate is the weight of a real decision with real consequences for people who have already been failed by systems that promised to help them. You cannot simulate the 11 p.m. moment when you catch a model hallucinating a drug interaction in a health tool that real patients will use, and you fix it alone, and you put it back up, and you go to sleep knowing the buck stopped with you.

That is where leadership theory either holds or it doesn’t.

For me, it holds. Not because the framework is perfect. Because the four people in the room make the right answer obvious even when the path is not.

Pick up enough crumbs, you get a loaf of bread.