The Age of Intelligence

FIELD NOTES SCHOOL'S OUT SATURDAYS
THE AGE OF
INTELLIGENCE

Two years of interviews sort into four buckets. Not one of them is about what we teach.

07.30.26ISSUE 077 MIN READBY JASON PADGETT

For over two years I have been interviewing educators, ed tech people, and administrators about AI implementation on the podcast. Hundreds of hours of it now. When I look at where those conversations actually go, they sort into four buckets, and only four.

01Safety and security
02Teacher tool use
03Student tool use
04Government and policy

All four are real and I am glad we are having them. Look at what all four have in common. Every one of them asks how to handle the thing. Not one of them asks what the thing does to the content of an education. Two years of serious conversation and the curriculum has not been on the table.

I think that is because we are treating AI like a lamp. Something you carry into a room and set down. Put it on the teacher's desk and it lights the teacher's work. Put it on a student's desk and it lights theirs. Set it by the door and the policy people can see it well enough to write a rule about it. Same lamp, moved around, different light depending on where it lands.

AI is not a lamp. It is electricity.

Electricity never got a spot in the room. It got run through the walls, and then everything in the building changed, including the parts nobody thought were electrical. The school day got longer. Night school became possible. The bell got wired to a clock. Nobody offered a course in electricity to make any of that happen.

General purpose cognitive technology is the plain name for what AI is. Writing was one. Print was one. Electricity was one. None of them ever became a class. They became conditions, and every subject inherited new questions whether the master schedule acknowledged it or not.

So here is the fifth bucket, the one I have not heard anyone open. Seven departments. What each one keeps, and what each one has already inherited.

> WHAT EACH ROOM INHERITS
ENGLISHComposition for machines beside composition for people. Technical writing, documentation, and prompt construction taught as what they are, which is writing with a specification.
PHYSICSHow physics behaves inside a machine against how it behaves in the room. A world model can predict a bouncing ball perfectly and still know nothing about the ball being wet. Add the energy and water a single query costs, why anyone would put a data center in orbit, and where quantum actually fits.
BIOLOGYPut a neural network beside a nervous system. How does learning work in a silicon based embodied AI compared to the carbon based animal in the next seat. Both are pattern machines, both improve by being wrong, and only one of them runs on glucose and sleep.
COMPUTER SCIENCEConcepts and systems ahead of code, because the machine already writes clean C++ and Python. The student should leave able to explain how machine learning, convolutional nets, predictive models, and multimodal language models fit together, and then aim that stack at a real problem in a business or a lab.
HISTORYIntelligence revolutions, with the gains and the damage both on the board. The Luddites had a real grievance and still lost. So did the firms that watched the computer arrive and never pivoted. A student should be able to argue both sides of what a revolution pays and what it costs.
GOV & ECONWhat an economy looks like once work is automated. Universal basic income. A government that simply provides healthcare and necessities. What capitalism, socialism, and communism each turn into when labor stops being scarce.
PHILOSOPHY NO LONGER OPTIONAL
Consciousness, creativity, and what intelligence is, next to the questions students are already living inside. Anthropomorphism. Sycophancy in a model built to please you. AI personhood. Who owns a machine mind and what that ownership entitles them to. Machine bias measured honestly against human bias. These left the university seminar years ago. They are in a seventeen year old's pocket tonight, and nobody has assigned them.

Now picture a K through 12 system that teaches all of that on purpose, age appropriate the whole way up. A first grader learns that a machine can be confidently wrong, which is the most useful thing anyone can know about one. A fifth grader compares how she learns to how a model learns. A sophomore reads an energy bill for a data center. A senior can hold her own in an argument about personhood and then go build something with a multimodal model on a Saturday.

That graduate walks into the age of intelligence already fluent in it. Not braced for it, not warned about it, fluent. Their future, not our past. That line is on the show for a reason and this is the version of it I care about most.

The wiring is the work, and it starts small. One department, two lines. What we keep. What we already inherited. Any chair can have that on paper by Friday, and every chair who does makes the fifth bucket a little harder to ignore.

> ONE REPLY QUESTION

Which of the four buckets does your district live in, and what would it look like to open a fifth?

Two sentences is plenty. I read all of them, and I quote the good ones with permission. If you would rather say it out loud, the podcast is where these four buckets got named in the first place.

SUBMIT YOUR RESPONSE HERE
School's Out Field Notes. Graphic reportage for educators. AI, school, work, and what students need next.
AGIPODCAST.AI/FIELD-NOTES ARCHIVE SUBSCRIBE