From Harness to Saddle: Where AI Value Will Ride

Sequel to Revenge of the Data Mesh: AI and the Fourth Epoch of Data Management

As Xenophon, write how “saddles” will commoditize AI harnesses — because they carry people instead of loads.

ChatGPT Prompt

Power versus Mastery

I have lived among horses, soldiers, merchants, and rulers. I have seen men destroyed by power they believed they possessed, and kingdoms preserved by men who understood how to guide power without being thrown from it.

The young often mistake strength for mastery.

They see the horse and admire its speed.
They see the engine and admire its intelligence.

But civilization is not built upon power alone.
It is built upon the means by which power becomes governable.

In my own day, the Persians possessed great numbers, great wealth, and great beasts. Yet many who owned horses could not ride them well. They burdened them with carts and loads, as oxen are burdened, believing that possession of force was equivalent to command over it.

But the peoples who mastered the saddle discovered something greater.

A horse that merely pulls increases labor.
A horse that carries a rider increases judgment.

The distinction is not small. It is the distinction between transport and cavalry, between agriculture and empire.

Commoditizing Harnesses

So too, in your age, men have discovered a new kind of force: artificial intelligence.

At first, they treated it as a beast of burden.

They built harnesses for it.

These harnesses are clever contrivances:

  • memory systems,
  • orchestration graphs,
  • guardrails,
  • verification engines,
  • tool routers,
  • state machines.

Such things are necessary. I do not mock them. No cavalryman despises reins.

But one must understand the nature of a harness.

A harness exists to constrain force so that it may pull a load without chaos.

Its purpose is reliability.

And reliability, once understood, becomes common knowledge.

The yoke becomes standardized.
The wheelwright copies the wheelwright.
The blacksmith copies the blacksmith.

Soon no man pays dearly for a slightly superior harness.

Thus it shall be with these systems you call frameworks and orchestration engines.

They solve universal engineering problems.
Universal engineering problems become commodities.

Yet many in your age confuse the harness for the product itself.

This is an error born from proximity to invention.

Civilizing Force

The true transformation begins only when intelligence ceases merely to pull work and begins instead to carry human intention.

This requires not a harness, but a saddle.

For a saddle is not simply leather upon a horse.

It is an agreement between human judgment and nonhuman power.

It allows the rider to guide without wrestling.
To steer without carrying the full burden.
To travel farther than he could walk.
To wield speed without surrendering control.

The saddle civilizes force.

Governance Is Profit

Thus the important question for your age is not:

“How intelligent are the machines?”

Nor even:

“How reliable are the machines?”

But rather:

“How shall institutions ride them?”

For institutions are not built merely from capability.
They are built from delegated authority.

A merchant delegates to his factors.
A general delegates to his captains.
A king delegates to his governors.
A republic delegates to its magistrates.

Every civilization is a structure for transferring judgment while preserving accountability.

This is why the greatest problems of your new intelligence are not technical, but political.

  • Who may authorize action?
  • Who bears responsibility?
  • When may the machine act alone?
  • When must it defer?
  • Who may halt it?
  • Who may audit it?
  • Who may overrule it?

These are not questions of computation.

They are questions of governance.

And therefore the enduring wealth of your era shall not belong solely to those who breed stronger horses.

Nor even to those who forge stronger harnesses.

It shall belong to those who build the trusted saddles through which human beings and institutions learn to ride intelligence safely.

For men will always pay to reduce labor.
But they will pay vastly more to extend judgment.

The winners of your age shall become the custodians of delegated agency itself.

They shall own the permissions,
the interfaces,
the memory of institutions,
the chains of approval,
the audit trails,
the operational trust.

In this way, they shall become not merely merchants of software, but governors of machine authority.

This pattern is ancient.

The horse transformed war only when cavalry arose.
Writing transformed kingdoms only when bureaucracy arose.
Coinage transformed trade only when law arose.

Power alone does not change civilization.

Governed power does.

And so I say to you:

The age of the harness is already ending.

The age of the saddle has begun.

Appendix I: The Recurring Transition

Throughout history, civilizations repeatedly mistake the source of power for the source of durable value.

At first, societies marvel at the raw capability itself:
fire,
horses,
steam,
electricity,
computation,
intelligence.

Entire industries arise around extracting, amplifying, and stabilizing these forces.

Yet over time, value migrates upward.

It leaves the engine and settles in the systems that make power governable.

The pattern recurs with remarkable consistency.

Agriculture

The first agricultural societies discovered that animals could multiply human labor.

  • The ox and horse were the engines.
  • Yokes, plows, and carts were the harnesses.
  • But the enduring value emerged in:
  • land systems,
  • irrigation,
  • logistics,
  • taxation,
  • and organized agriculture itself.

The wealth did not ultimately belong to the maker of the yoke.
It belonged to the civilization that learned how to organize labor around animal power.

Cavalry

Ancient societies first used horses to pull loads and chariots.

Then came the saddle and stirrup.

This changed everything.

  • The horse ceased merely to pull labor.
  • The horse began carrying judgment.

Mounted command transformed:

  • warfare,
  • communication,
  • administration,
  • empire.

The key innovation was not more horsepower.

It was governable horsepower.

Navigation

Ships created unprecedented mobility.

But ships alone did not create maritime civilization.

The enduring systems were:

Again, value migrated upward:
from the vessel itself
to the institutional systems that governed movement.

Industry

The steam engine transformed production.

Then came:

  • rail systems,
  • schedules,
  • logistics,
  • factories,
  • national infrastructure.

The enduring fortunes were not built solely upon engines.
They were built upon coordination systems.

Industrial civilization was fundamentally an exercise in governing mechanical power at scale.

Electricity

Generators created electrical force.

But civilization changed only once electricity became operationalized through:

  • transmission grids,
  • switching systems,
  • appliances,
  • factories,
  • lighting,
  • urban infrastructure.

Electricity became valuable when ordinary people no longer needed to understand generation itself.

The complexity disappeared beneath usable systems.

Computing

Microprocessors created computational power.

Then came:

  • operating systems,
  • middleware,
  • networking,
  • databases,
  • enterprise software.

Again, the greatest fortunes accrued not to raw compute itself, but to the systems through which institutions coordinated computation.

The operating layer became more valuable than the processor.

The Internet

TCP/IP created global connectivity.

But enduring value accumulated around:

  • search,
  • marketplaces,
  • identity,
  • platforms,
  • social graphs,
  • application ecosystems.

Protocols became infrastructure.
Governed interaction became the product.

Cloud Computing

Cloud infrastructure industrialized compute.

Then came:

Servers became commodities.

Operational governance became the durable value layer.

Artificial Intelligence

Foundation models generate intelligence.

Harnesses stabilize intelligence through:

  • orchestration,
  • memory,
  • tool use,
  • verification,
  • guardrails,
  • observability.

These are necessary advances.
But they solve universal engineering problems.

And universal engineering problems eventually standardize.

The next layer is the Saddle.

The Saddle governs:

  • delegated authority,
  • escalation,
  • permissions,
  • oversight,
  • workflow continuity,
  • institutional trust.

The important transition in AI is therefore not merely:
from models to orchestration.

It is:
from contained intelligence to governable agency.

Across every era, the same transition recurs:

  • First comes raw force.
  • Then comes engineering discipline.
  • Finally comes institutional governability.

The first phase rewards invention.
The second rewards engineering.
The third rewards governance.

This is why harnesses repeatedly commoditize.

Harnesses solve universal technical problems:

  • reliability,
  • coordination,
  • execution,
  • transport,
  • routing,
  • stabilization.

But saddles solve contextual human problems:

  • trust,
  • delegation,
  • accountability,
  • authority,
  • workflow,
  • institutional continuity.

And contextual systems preserve value.

Thus the enduring economic layer of AI is unlikely to be:

  • the model itself,
  • the orchestration framework,
  • or the memory architecture.

Those become infrastructure.

The enduring value will accrue to the systems that determine:

  • who may delegate authority,
  • how responsibility is transferred,
  • when escalation occurs,
  • how oversight is preserved,
  • and how institutions remain capable of steering machine intelligence without directly operating it.

Civilization advances when power becomes governable.

The saddle is the abstraction through which governance becomes operational.

Appendix II: Who Was Xenophon?

Xenophon of Athens (c. 430–354 BC) was many things:

  • soldier,
  • cavalry commander,
  • historian,
  • student of Socrates,
  • political exile,
  • and one of the earliest thinkers about leadership, management, and organized power.

He lived during the long decline of classical Athens after the Peloponnesian War, in an age when Greek civilization was learning painful lessons about coordination, discipline, and imperial overreach.

Unlike many philosophers, Xenophon was not primarily a theorist of abstraction.

He was a practitioner.

He commanded troops.
He organized retreats.
He managed estates.
He trained horses.
He advised rulers.

He understood systems because he had survived them.

The Soldier Who Learned Governance

Xenophon first became famous through the Anabasis, his account of leading ten thousand Greek mercenaries trapped deep inside the Persian Empire after the death of their employer, Cyrus the Younger.

Surrounded by enemies and abandoned without leadership, the Greeks faced annihilation.

Xenophon emerged not through formal authority, but through operational competence.

He helped transform a disordered mass of frightened soldiers into a functioning organization capable of:

  • coordination,
  • retreat,
  • supply management,
  • negotiation,
  • morale maintenance,
  • and adaptive leadership under extreme uncertainty.

The Anabasis is often remembered as an adventure story.

But beneath it lies a deeper insight:

Civilization survives not through raw force alone, but through disciplined systems of delegated judgment.

Xenophon and the Horse

Among Xenophon’s lesser-known works are:

These texts are not merely technical manuals.

They reveal a worldview.

Xenophon understood that horses were not simple tools.
They were powerful, semi-autonomous intelligences requiring:

  • trust,
  • discipline,
  • training,
  • responsiveness,
  • and mutual adaptation between rider and mount.

A frightened rider destroys a horse.
A brutal rider exhausts it.
A careless rider is thrown.

The objective was not domination through force.

It was governable partnership.

This makes Xenophon an unexpectedly fitting guide for the age of artificial intelligence.

For AI resembles the horse more than the plow.

It is:

  • fast,
  • adaptive,
  • powerful,
  • partially unpredictable,
  • capable of extending human reach,
  • but dangerous when improperly governed.

Harnesses can constrain such power.

But saddles allow humans to ride it.

The First Management Thinker

In Oeconomicus, Xenophon explored household and estate management.

Yet the Greek word oikonomia meant something larger than finance.

It referred to the governance of systems:

  • people,
  • labor,
  • property,
  • responsibility,
  • and delegated authority.

Xenophon repeatedly returned to the same question:

How does a human being extend intelligence and judgment beyond the limits of one individual mind?

His answer was never “through force alone.”

Instead, effective leadership required:

  • incentives,
  • trust,
  • hierarchy,
  • discipline,
  • training,
  • accountability,
  • and operational visibility.

In modern terms, Xenophon was thinking about:

  • control planes,
  • governance systems,
  • delegation architectures,
  • and scalable coordination.

Two thousand years before software.

Why Xenophon Matters Now

Most technological revolutions begin by worshipping capability.

The early builders admire:

  • speed,
  • power,
  • novelty,
  • scale.

But mature civilizations eventually discover that the central problem is governance.

Not:
“How powerful is the system?”

But:
“How can human institutions safely operate it?”

Xenophon belonged to a civilization making precisely this transition.

The horse had already transformed mobility.
But societies were still learning how to operationalize mounted power:

  • militarily,
  • politically,
  • economically,
  • institutionally.

The saddle was not merely a piece of equipment.

It was a civilizational abstraction.

It transformed horsepower into governable agency.

That same transition now confronts artificial intelligence.

And so Xenophon stands as an unlikely patron thinker for the AI age:
the philosopher-cavalryman who understood that enduring power comes not from possessing force, but from learning how to ride it.

Appendix III: Making Saddles for Humans in the Age of AI

By Thierry Hermès

When I first opened my workshop in Paris, the world was changing.

The old aristocracy still rode by carriage and horse, yet industry was already beginning to reorganize Europe around steam, rail, and mechanical power. Many craftsmen believed the future belonged to those who produced stronger equipment at lower cost.

They were mistaken.

The true opportunity lay not in competing with horsepower, but in refining the human experience of governing horsepower.

Thus we did not merely make harnesses.

We made saddles.

This distinction is important.

A harness exists primarily for the animal.
A saddle exists primarily for the rider.

The harness distributes force.
The saddle distributes trust.

The harness allows labor.
The saddle allows command.

A poor saddle exhausts both horse and rider.
A proper saddle allows them to move as one intelligence.

This was never merely a matter of comfort.

It was a matter of governance.

For the wealthy families and officers who came to us did not truly purchase leather. They purchased:

  • confidence,
  • control,
  • endurance,
  • safety,
  • elegance under motion,
  • and the ability to travel farther without losing command of themselves or their mounts.

A well-made saddle transforms raw capability into usable civilization.

This lesson now returns in your age of artificial intelligence.

I observe with fascination that many builders today resemble the blacksmiths and wheelwrights of my own century.

They labor tirelessly upon:

  • engines,
  • frameworks,
  • memory systems,
  • orchestration layers,
  • and mechanical reliability.

Such work is honorable.

But it is not where enduring value will reside.

For eventually every harness becomes standardized.

The market cannot preserve large fortunes forever in straps and buckles alone.

Infrastructure diffuses.
Techniques spread.
Competitors imitate.
Craft becomes process.

Yet human beings will always pay dearly for trusted systems that allow them to wield power safely.

This is why the greatest luxury houses survived industrialization.

Not because they controlled factories.

But because they understood the emotional and institutional dimensions of power.

A saddle is intimate.

It sits precisely at the boundary between:

  • human judgment,
  • and nonhuman force.

It absorbs shock.
It translates movement.
It prevents panic.
It enables trust.

Without such mediation, the rider tires quickly and loses confidence.

This is precisely the condition facing your artificial intelligences.

The problem is no longer merely whether these systems are capable.

They are already capable.

The problem is whether:

  • executives,
  • governments,
  • workers,
  • institutions,
  • and ordinary people

can learn to inhabit this new intelligence without fear, exhaustion, or loss of control.

That requires a new generation of saddlemakers.

The successful companies of your era will not simply produce more intelligent engines.

Nor will they win merely by engineering stronger harnesses.

They will win by creating trusted interfaces through which human beings can:

  • delegate authority,
  • preserve oversight,
  • maintain dignity,
  • and remain psychologically capable of steering machine intelligence.

The finest saddle is one the rider eventually ceases to notice.

It disappears beneath fluency.

The horse feels natural.
Movement feels intuitive.
Trust becomes unconscious.

This is the highest form of infrastructure:
that which vanishes into human confidence.

In time, your users will not ask:

  • which orchestration framework powers the system,
  • which memory architecture it employs,
  • or which model version it runs beneath the surface.

They will ask:

  • whether it understands them,
  • whether it protects them,
  • whether it carries responsibility well,
  • whether it preserves trust,
  • and whether they feel safe placing institutional authority in its care.

This is no longer a question of engineering alone.

It is a question of civilization.

And so I say this as one who spent a lifetime crafting saddles for another age of accelerating power:

Human beings do not merely seek capability.

They seek a way to inhabit capability without being destroyed by it.

Those who understand this will build the enduring houses of the AI era.

Leave a comment

Blog at WordPress.com.

Up ↑