I really liked the cover generated for my previous post, Why I Built OpenZeroCode: From Learning Coding Agent Harnesses to Building My Own Reins for the AI Era, so for fun I asked ChatGPT to write a story and generate a sequence of storyboard images.
The Mechanical Horse of the Late Edo Period
Near the end of the Edo period, there was an unremarkable little workshop outside the castle town.
Its owner was Masato.
He was neither a swordsmith nor a carpenter.
Every day, he stuffed gears, copper wire, paper talismans, and strange mechanisms that nobody understood into the belly of a wooden horse.
People in town laughed at him.
“Horses are for riding. What good is a wooden horse?”
“Even if it really moves, could it possibly be faster than a living horse?”
Masato did not answer.
Because what he wanted to build had never been just a horse that could run.
Around that time, something called a “mechanical steed” had begun appearing across the country.
You only had to write an instruction on a slip of paper and place it inside the horse’s belly, and the mechanical steed would move on its own.
Someone would write:
“Deliver this letter to Kyoto.”
The horse would race toward Kyoto.
Someone else would write:
“Move my rice to the warehouse.”
The horse would pull the cart away by itself.
Everyone who saw a mechanical steed for the first time was amazed.
Some people even said:
“Soon we won’t need grooms anymore.”
But before long, problems began to appear.
Someone wrote only “Deliver it to Kyoto” without saying where in Kyoto, so the mechanical steed kept circling outside the city gate.
Someone ordered it to “take the fastest route,” so it cut straight across farmland and ruined an entire season of rice.
Another person told it to move some goods. The next morning, he discovered that it had moved the neighbor’s sacks of rice as well.
The mechanical steeds were certainly fast.
But nobody knew what they would do next.
Soon, a saying began spreading through town:
A mechanical steed is not a divine horse. It is simply a horse that listens too literally.
When Masato heard this, it only made him more certain about what he was building.
The problem was never the horse.
The problem was that people threw instructions at it without first preparing proper reins.
So Masato opened his wooden horse again.
The first mechanism he added inside its body was called the “Memory Box.”
The Memory Box remembered the routes its owner had taken, the decisions that had been made, and the places where things had gone wrong.
Next, he added a second mechanism called the “Toolbox.”
Inside were maps, seals, an abacus, keys, and all kinds of tools. The mechanical steed no longer simply ran forward. It could choose the right tool for the task.
The third mechanism was called the “Observation Mirror.”
After every stretch of travel, the mechanical steed first had to look carefully at what was happening ahead before deciding what to do next.
Finally, Masato fitted a red rein around the horse’s neck.
It was not an ordinary rein for controlling direction.
Whenever its owner pulled it, the mechanical steed had to stop and wait for confirmation.
The people around him did not understand.
“If it can move by itself, why does it still need reins?”
Masato answered:
“Precisely because it can move by itself.”
A few months later, the domain lord heard about this strange mechanical steed and summoned Masato to the castle.
The lord asked him:
“What can your horse do?”
Masato replied:
“It can now deliver letters, move goods, read maps, and help people organize their accounts.”
The lord looked somewhat disappointed.
“The other mechanical steeds in town can do that too.”
Masato replied:
“My lord, everyone else is building a faster horse.”
“I am building a way for people to control one.”
The lord then gave him a task.
There was a problem with the accounts at a warehouse south of the castle, and he wanted Masato to send his mechanical steed to investigate.
Masato did not simply write, “Solve the problem.”
He first placed the warehouse location, old account books, a list of guards, and the investigation rules into the Memory Box.
Then he told the mechanical steed:
Read the account books first.
Then compare them with the goods.
If the records do not match, record the discrepancy first. Do not change anything on your own.
If silver is involved, wait for human confirmation.
The mechanical steed set off.
It checked the account books first, then entered the warehouse and counted the goods.
When it discovered that three crates of rice were missing, it did not immediately arrest anyone or change the numbers on its own.
It simply stopped and sent a message back through the reins.
When Masato arrived, he discovered that the three crates had not been stolen. A few days earlier, they had been sent out as emergency flood relief, but nobody had updated the records.
Another mechanical steed might already have arrested the guards as thieves.
Masato’s did not.
Because it knew what it was allowed to do, and it also knew what it should not decide for itself.
The lord watched all of this and remained silent for a long time.
At last, he understood that the truly important question was not whether a horse could run by itself.
It was whether its owner had prepared enough maps, memory, tools, and rules for it.
From that day on, Masato’s workshop stopped building only mechanical steeds.
Some people used them to manage freight.
Some used them to organize official documents.
Some asked them to help trading companies calculate purchases, while others used them to manage the everyday work of an entire residence.
Mechanical steeds gradually evolved from simple errand runners into assistants capable of operating all kinds of mechanisms.
Later still, they began helping merchants handle business, reply to customers, and analyze accounts.
Only then did the people in town realize that Masato had never really been building a horse.
He had been building a mechanical system of his own.
Many years later, someone asked the elderly Masato:
“Why did you build that mechanical steed in the first place?”
Masato looked at the faded red rein hanging on the wall and smiled.
“At first, I only wanted to understand why it could run.”
“Later, I realized that the real question was not how to make it run faster.”
“It was whether, once every horse in the world could run on its own, we would still be able to decide where we wanted to go.”
Conclusion
After hearing such profound words, I feel as though I have heard some words. That is one of the impressive things about AI today: it can give you a result extremely quickly, but somehow it still feels like something is missing—a bit of soul.