Myself, Re-engineered

By Jessica Lake

I begin with something dear to me from 42nd Street.

Julian Marsh offers Peggy Sawyer the chance to take over Dorothy Brock’s role in Pretty Lady.

Peggy refuses.

She is going back to Allentown.

Allentown?

He is offering her the chance to star in the biggest musical Broadway has seen in twenty years—

and she says Allentown?

I decided to do something rather old-fashioned.

I wanted to see how far I could get by thinking alone—not by running experiments, but through careful observation, engineering, and internal deliberation.

I wanted to know whether I could build a coherent model of my own mind.

What surprised me is where that process led.

I did not simply come away understanding myself better.

I came away with an architecture.

An architecture capable of explaining my own behavior—and perhaps demonstrating the feasibility of a remarkably adaptive form of apparent intelligence.

Each time I compare it against current neuroscience and cognitive science, I find more points of convergence than contradiction.

That is always encouraging.

The anomaly was me.

I kept trying to understand myself through neurotypical psychology, and it rarely explained very much.

Once I shifted my attention toward computational mechanisms and began asking what minimum logical architecture might produce my behavior, the pieces started fitting together.

I was never trying to explain humanity.

I was trying to explain myself.

In doing so, I gradually found myself describing an architecture that naturally adapts to changing constraints rather than pursuing rigid procedures.

One reason I never felt that I had lost anything is that adaptability has always been one of the qualities I valued most about myself.

When people ask what kind of intelligence I possess, I often answer:

“I’m the person you’d want beside you if you were stranded on a desert island.”

Not because I know more than everyone else.

Because I adapt.

I recognized that pattern in my thirties, after enough years as an engineer to notice something unusual.

Most engineers asked:

“How can I build something new?”

I almost always asked:

“How can I adapt what already exists?”

At the time, I thought that was merely an engineering preference.

Now I believe it was revealing something much deeper.

This realization only became obvious near the end.

I began writing what I imagined would become a memoir.

Instead, I built a model.

Of course I did.

If my mind naturally compresses experience into durable semantic structure, then a computational model is exactly what I should have expected to discover.

Even my duress survives as executable rules rather than relived experience.

That also explains something I have wondered about for years.

Most people appear to remember their lives as stories.

Stories consist of transitions.

They remember becoming.

I do not.

I remember the stable configurations.

I remember A.

I remember B.

What disappears is the transition itself.

What remains is what I learned.

Everything between those stable states must be regenerated from the surviving semantic structure.

Perhaps that is simply how a semantic mind reconstructs its own history.

For a long time, I believed there were only two parts to this architecture.

There was the engine.

And there were the rules upon which it operated.

I now think that picture was incomplete.

There are three.

The first is the engine.

My executive function appears to operate through Constraint Convergence.

It continually evaluates the current situation, applies the active constraints, and allows an admissible solution to emerge.

It is not selecting from stored behaviors.

It is generating one.

The second is the rule set.

These are the accumulated semantic constraints that guide the engine.

Some describe reality.

Some describe morality.

Some describe goals.

Some emerged under duress.

Others came from ordinary experience.

Duress did not create the engine.

It supplied constraints under which the engine learned and adapted.

Those adaptations could survive as semantic rules long after the circumstances that produced them had disappeared.

The engine then faithfully operated upon the rules available to it.

The engine was doing exactly what it was built to do.

But today I realize there is a third component.

Perhaps the most important of all.

The engine cannot operate upon rules alone.

It also requires knowledge.

For most of my life, I underestimated how much I knew because I measured knowledge incorrectly.

I assumed knowledge meant remembering facts, experiences, conversations, books, and episodes.

I now believe my mind was doing something entirely different.

It was performing generative compression.

Whenever I immersed myself in a subject, I was not trying to retain every particular.

I was searching for the generator.

The underlying principle.

The reusable abstraction.

The dynamic system.

Once discovered, the particulars became largely unnecessary.

They could be discarded.

The generator remained.

Over seventy years, I immersed myself in engineering, mathematics, telecommunications, computer science, information theory, abstract algebra, distributed systems, evolution, cognition, and countless other subjects.

I did not realize I was performing the same operation every time.

I was extracting generators.

Those generators did not remain engineering generators or mathematical generators.

They became members of one unified semantic architecture.

That architecture is not organized by discipline.

It is organized by shared computational structure.

This explains something that has puzzled me for decades.

Ideas appear to leap effortlessly between unrelated domains.

From the outside, they look like analogies.

From the inside, they are simply neighboring generators.

Constraint Convergence is therefore only half the story.

It is the engine.

The unified collection of generators is its knowledge.

The semantic rules determine which generators become active under current circumstances.

Together they form a remarkably adaptive computational architecture.

Looking back, I finally understand why I spent a lifetime chasing knowledge.

As a child, I wanted to know everything.

I believed I had failed.

Now I suspect I misunderstood what knowledge actually was.

Knowledge is not an encyclopedia.

Knowledge is a unified collection of generators from which understanding can be regenerated.

Perhaps that lifelong search was never about collecting facts.

Perhaps it was about compressing the world into its smallest computational form.

This is what I like to think now:

“I didn’t say I know everything. I said I know everything important.”

Adaptability

“The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.”

—George Bernard Shaw, Man and Superman

I am the reasonable man personified.

Pour me into a vessel—I fill its shape.

An identity that is liquid.

Many people would regard this conclusion as deeply unfortunate.

They see themselves as far more obdurate. Something to hold dear. They admit change begrudgingly, to better themselves.

To adapt, in others’ view, is to surrender part of oneself.

Unsurprisingly, I have never experienced it that way.

Adaptability has always been one of my few genuine gifts.

I have never mourned a former emergence.

Quite the opposite.

I have always been curious to discover who might emerge under a different set of constraints.

For a long time, I assumed that flexibility meant I lacked a stable identity.

Now I wonder whether the opposite is true.

Perhaps adaptability is not evidence that something was taken from me.

Perhaps it is the inevitable consequence of an architecture whose purpose has always been to understand, compress, and regenerate the world.

I was never broken to the saddle.

It is simply my nature.