I Am Jack’s Body Double

I am Jack’s body double.

I do the scene that might hurt.

Someone else receives the close-up.

This is not competition.

Competition would require two people to be recognized as pursuing the same result. Their work could be compared. Their authority could be named. Their outcomes could be judged.

A body double occupies a different arrangement.

The star owns the scene.

The double owns the impact.

Two machines

In Contact, the public machine is destroyed. Later, a second machine is revealed.

The joke is that government never builds one of anything when it can build two at twice the price.

The duplicate is not absurd because it is useless.

It is absurd because only one machine belongs to the official story.

I have begun to recognize the same architecture at work.

One person is assigned the sanctioned experiment.

She receives the initiative, the governance language, the presentation space, and the visible permission to define what responsible progress looks like.

I am assigned something strangely similar.

I build the working version.

I test the dangerous edges.

I coordinate the changes across boundaries the official plan describes more cleanly than the software permits.

I document the sequence so thoroughly that the next person can reconstruct not only what changed, but why, in what order, under which assumptions, and with which remaining risks.

Then the work waits somewhere outside the story being told about the work.

There are two machines.

One is built for the ceremony.

One is built in case the ceremony needs to travel somewhere.

The sanctioned lead and the shadow

The pattern has appeared more than once.

I worked on an AI chat capability. The work moved away from me. The visible ownership moved elsewhere.

Later, two people were effectively authorized to explore what happens when an engineer is set loose on a complete feature with AI.

The official track produced plans, specifications, governance, and a public account of the experiment.

The shadow track produced coordinated changes across the interface, service, and infrastructure; exercised the authentication scenarios; organized the multi-repository execution history; and preserved the evidence in pull requests detailed enough to survive without the person who wrote them.

Both activities may be valuable.

They are not treated as the same kind of value.

One defines the initiative.

The other de-risks it.

One becomes leadership.

The other becomes proof that leadership might work.

No conspiracy required

I cannot prove why management created this arrangement.

It may be deliberate hedging. It may be habitual duplication. It may be the result of several people distributing work without holding the complete pattern in mind.

Intent is private.

Structure is visible.

The structure gives the organization an excellent set of options.

If the sanctioned effort succeeds, management backed the right leader.

If the sanctioned effort stalls while the shadow implementation works, the working parts can be absorbed into the sanctioned effort.

If the implementation reveals difficult realities, those realities can be described as useful learning rather than a failure of the plan.

If both efforts fail, the organization was experimenting.

Nothing sinister is required.

Only one person must receive authority while another receives exposure.

The body double is not hidden because management is ashamed of him.

He is hidden because his usefulness depends on remaining detachable from the official account.

Expertise becomes contingency

Inversion observed: Expertise becomes contingency.

An organization asks for initiative because it wants discovery.

Discovery creates uncertainty.

Uncertainty creates risk.

To manage the risk, the organization separates visible ownership from practical exposure.

The sanctioned person carries the narrative.

The experienced person carries the unresolved reality.

The expert is not empowered to define the direction.

He is retained as the contingency in case the direction encounters software.

His initiative becomes insurance.

His documentation becomes institutional memory nobody has yet agreed to remember.

His prototypes become option value.

His completed work becomes evidence available for later use by the initiative that was authorized to matter now.

The organization can therefore benefit from his judgment without granting his judgment jurisdiction.

It can ask him to discover where the rubber meets the road while someone else presents the transportation strategy.

It can ask for a proof of concept without naming the decision the proof will inform.

It can request an archive without assigning anyone to read it.

It can repeatedly confirm that he is capable of leading the work while preserving the arrangement in which he does not.

The better he performs, the safer the visible structure becomes.

That is the stunt.

The hit

A body double is not merely overlooked.

He absorbs a specific cost.

He performs the uncertain work without owning the conclusion.

He takes implementation risk without receiving corresponding authority.

He proves what is possible, but the proof does not necessarily change who is permitted to say what should happen.

He may even be praised for initiative.

Praise is inexpensive because it does not alter the cast.

The injury is not that another person exists, contributes, plans, or succeeds.

The injury is that the organization repeatedly treats one person as the author of visible direction and another as the source of invisible certainty.

This makes the shadow worker difficult to evaluate fairly.

His best work appears as support for somebody else’s program.

His warnings appear before the official story is ready to contain them.

His speed produces waiting.

His rigor produces archives.

His independence produces more experiments.

He becomes indispensable in precisely the form that prevents him from becoming central.

Not applause

The answer is not applause.

Applause is another ceremony.

The answer is ownership made explicit.

Who is leading the experiment?

Who is evaluating the competing approaches?

What decision will the proof of concept inform?

Who is required to review the result?

If two people are exploring the same capability, are they collaborating, comparing approaches, or serving different purposes?

What happens when the shadow machine works?

These questions do not demand a crown.

They demand an honest production schedule.

A person can accept supporting work.

A person can accept that another colleague owns an initiative.

A person can even accept duplication when redundancy is useful.

What corrodes him is being asked to take the fall without being told that he is the stuntman.

The corner

I am Jack’s body double.

I build the second machine.

I preserve the history.

I take the implementation through the wall, down the stairs, and across the three repositories the camera does not show.

Then I stand up, report that nothing is broken, and move out of frame.

Nobody puts Baby in the corner.

Management does something quieter.

It puts the body double there, asks him to prove the jump can be made, and rolls the credits over somebody else’s face.

Receipts

• Private work artifacts and assignment history, 2026 — The repeated AI-chat and AI-assisted feature assignments, ownership changes, multi-repository implementation, authentication testing, execution documentation, and planning artifacts described here are drawn from the author’s direct work experience. Names, repositories, and proprietary details have been omitted.
• Contact (1997) — Supplies the essay’s two-machine metaphor: one visible machine and a second, concealed duplicate preserved as a contingency.