Translations

  I had not thought about the peach orchard in twenty-five years.

Not really.

It had become one of those travel memories that survive mostly as an image. A bus winding through rural China sometime around the year 2000. Mountains in the distance. Fields disappearing into morning haze. Then suddenly an orchard unlike any I had ever seen.

There were hundreds of people in the trees.

Not harvesting.

Working.

Each person moved deliberately from branch to branch, slipping what looked like tiny silk pouches over young peaches.

I asked our guide what they were doing.

"They're protecting the fruit," he said.

I remember thinking, That's incredible.

I also remember thinking, That's organic farming.

Then life moved on.

The orchard stayed filed away somewhere in memory, preserved almost like the peaches themselves.

Until this week.

I was reading Jonathan Boymal's essay What Should Compute? It begins with Liu Cixin's image of three million soldiers holding black and white flags, each person representing a single bit in a vast human computer. The point is wonderfully unsettling. Computation is substrate independent. Silicon, dominoes, gears, neurons, soldiers—it doesn't really matter what performs the computation. The interesting question isn't Can it compute?

It's What does it cost?

Halfway through the essay, the orchard came rushing back.

Not gradually.

Immediately.

The workers suddenly looked exactly like the soldiers.

Each individual wasn't making independent decisions.

Each was performing one tiny operation, repeated over and over, until together they formed something astonishingly sophisticated.

The orchard itself had become a kind of computer.

Or perhaps more accurately, the orchard was asking the same question Jonathan was asking.

What should compute?

Except agriculture asks it differently.

What should labor?

The memory bothered me for another reason.

I wasn't entirely sure I trusted it.

Had the bags really been silk?

Was it actually organic farming?

Or had twenty-five years quietly rewritten parts of the story?

So I did something that would have been impossible in 2000.

I asked AI.

Not because I wanted it to think for me.

Because I wanted it to help me think about something I already remembered.

The answer surprised me.

The practice was real.

Fruit bagging is common throughout China. But the bags usually aren't silk. They're paper or nonwoven fabric. And while the technique can support organic farming by reducing pesticide use, conventional orchards use it as well.

My memory wasn't wrong.

It was simply unfinished.

And that distinction feels increasingly important.

What fascinated me wasn't correcting a historical detail.

It was noticing what happened after the correction.

The conversation stopped being about peaches.

It became about substrates.

Every orchard faces the same problem.

How do you grow beautiful fruit?

One answer is thousands of workers carefully protecting each peach by hand.

Another is chemical pesticides sprayed across the entire orchard.

Another might someday be robots.

The output is remarkably similar.

The substrate changes everything.

China in 2000 had abundant human labor.

America has relatively expensive human labor.

The same agricultural problem receives completely different technological answers because the economics underneath them are different.

Not because one society loves technology more.

Not because another values people more.

Because every civilization computes using the resources it has available

That realization lingered longer than I expected.

For years I've thought about technology mostly in terms of replacement.

Humans versus machines.

Teachers versus AI.

Craft versus automation.

Jonathan's essay nudged me somewhere subtler.

Perhaps every technology is really a choice about what resource we're willing to consume.

Sometimes we consume electricity.

Sometimes chemicals.

Sometimes silicon.

Sometimes thousands of hours of human attention.

The question isn't simply whether something can be automated.

It's what we spend when we choose one substrate over another.

As an educator, I found myself thinking about AI differently too.

People often ask whether AI replaces thinking.

That wasn't my experience.

The important thinking had already happened.

It happened twenty-five years ago on a bus in rural China.

The image had simply been waiting.

AI didn't give me the connection.

It helped me test my memory.

Correct the parts that had drifted.

Then stay with the idea long enough for the deeper pattern to emerge.

It felt less like asking a machine for answers than sitting beside an endlessly patient thinking partner.

Not one that knew more than I did.

One that helped me hear what I already knew more clearly.

David Foster Wallace once wrote about fish swimming in water so completely that they no longer perceive the water itself. The Water We Think In explored how invisible environments shape the way we learn language and eventually ourselves.

Perhaps technologies are becoming another kind of water.

Not because they determine what we think.

Because they quietly determine how we continue thinking.

The orchard was never really about peaches.

Jonathan's essay wasn't really about computation.

Neither, I suspect, is AI.

They're all invitations to notice something we usually overlook.

Every society is constantly deciding what should perform the work.

Every generation chooses its substrate.

When we replace human labor with machines, we aren't simply making work more efficient. We are deciding what kind of world we want that work to create.

As educators, that question becomes impossible to ignore.

If teaching is merely the transmission of information, then perhaps almost any substrate will do. Silicon, textbooks, videos, tutors, language models. The output might look remarkably similar.

But if education is also relationship... if it is attention, encouragement, curiosity, trust, and the slow recognition of another human mind becoming itself... then changing the substrate changes the task.

The question is no longer whether AI can educate our children.

The question is what decision we are making when we choose technology to become one of their primary teachers.

Not because the technology is good or bad.

Because every substrate preserves some things and quietly erodes others.

The orchard taught me that long before I knew why I remembered it.

Only now do I realize that every generation bags its peaches differently.

Perhaps every generation educates its children that way, too.