Mirror, Mirror

In 1976, Belgian anthropologist Jean-Pierre Dutilleux filmed first contact with the Toulambi tribe in Papua New Guinea.

At one point, he handed them a mirror.

They had likely seen reflections before in water or polished surfaces. But a small handheld object that perfectly reproduced their image felt like sorcery.

Some recoiled in terror.

Some stared at themselves for long stretches.

Some seemed curious.

A few experimented.

With a basic understanding of optics, the mystery and fear would have faded. The mirror’s utility would have become clearer.

Our current moment with AI feels similar.

Some people fear it while others obsess over it.

Some dismiss it.

Many have never touched it.

A few are learning how to use it in incredible ways.

There are legitimate reasons to belong to each group.

Safety concerns are as real as the potential breakthroughs.

AI is over-hyped in ways and vastly under-appreciated in others.

Just as understanding basic physics would demystify the mirror, having even an elementary idea of how AI models work can provide a clearer perspective on its future impact.

Over the next several issues of The Leap, I want to build that understanding using plain language. No AI engineering degree required.

My goal is to provide you with a framework to help you intentionally transition into an AI world both personally and professionally.

The best place to begin is simple.

What is AI?

And just as important, what is it not?

What AI Is

At its core, modern AI is a prediction system.

When you type a prompt into a language model, it is not pausing to think. It is calculating. Based on patterns learned from enormous amounts of text and other data, it estimates what sequence of words, images or numbers is most likely to come next.

It does this one small piece at a time, repeatedly.

That is the mechanism.

Try this: what's the first word that comes to mind after "Old ______"?

You might say "man" or "woman".

But if I say "Old MacDonald ______," nearly everyone responds with "had", to finish the nursery rhyme lyric, “Old MacDonald had a farm.”

You predicted “man”, “woman” or some other word on the first try based on what your experience has taught you is likely to come next.

The exercise was easier when presented with “Old MacDonald” because your confidence in the likelihood of the next word increased since you know the song.

AI works exactly the same way and its sophistication comes from scale.

Models are trained on vast patterns of human art, language, argument, explanation, humor, and structure. When it produces an answer, it is drawing on those patterns and assembling a response that statistically fits the data it absorbed during training.

It is not generating ideas exactly the way a human does. It is generating the most probable continuation or outcome based on the question or prompt provided by the user.

AI is extremely good at recognizing patterns and reproducing them in useful ways.

It should be no surprise that AI is often used for image recognition, coding software, all forms of writing and performing basic conversation functions such as customer service.

Those activities, at their core, are pattern matching and recognition.

That is its power.

What AI Is Not

Because AI’s output feels fluent, we project onto it.

We say it understands.

We say it reasons.

We say it thinks.

It does none of those things in the human sense.

Recently, scientists announced they had mapped just one-millionth of a human brain. The sample was smaller than a grain of sand, about one millimeter squared.

In that tiny fragment they found 57,000 cells, 150 million synapses, and over nine inches of blood vessels. The estimated storage capacity of that microscopic slice was more than 1.4 million gigabytes. A high-end laptop might hold 1,000.

And that was a speck.

Every person you meet is walking around with a biological system of staggering complexity between their ears. A living network shaped by emotion, memory, hormones, sensation, and lived experience.

AI has none of that.

It has no awareness of itself. No internal narrative. No sensations. No goals. No beliefs. No fear of being wrong.

When it produces an answer, it is not drawing from experience. It is assembling patterns.

It resembles understanding because it was trained on enormous volumes of human expression. It has learned what understanding sounds like.

Fluency creates the illusion of comprehension.

Confidence creates the illusion of certainty.

When people fear AI, they often imagine something intentional. When people trust it blindly, they assume something authoritative.

Both reactions come from the same mistake.

We confuse pattern replication with cognition.

AI is powerful. But it is not a mind. At least not yet.

The Next Step

What the Toulambi first experienced as sorcery when they saw the mirror was only physics.

Although AI can feel like magic, it’s not. It is mathematics.

Right now, we are in that first-contact phase. Some recoil. Some stare. Some dismiss. A few experiment.

But fear and fascination both fade with understanding.

You do not need to become an AI engineer, but you should have a basic understanding of how models work.

When you understand that AI is a prediction system, not a mind, you stop reacting emotionally and start evaluating strategically.

You begin asking better questions such as:

• Where is pattern recognition valuable in my work?

• Where does human judgment still matter most?

• Where can this tool amplify me instead of replace me?

This kind of clarity is leverage.

Next week, we go one layer deeper.

If AI is a prediction system, what exactly is a large language model (LLM) like Chat-GPT or Claude?

How does it break language apart?

How does it generate responses?

And why does it sometimes get things confidently wrong?

The mirror is no longer mysterious.

Next week we’ll examine how it is built.

My goal with The Leap is to provide you each Saturday with the knowledge, tools and lessons learned to help you get started and keep going toward building your future. 

Whether you are making the leap to startups, solo-entrepreneurship, freelancing, side hustles or other creative ventures, the tools and strategies to succeed in each are similar.