John Rector / September 9, 2026 / Artificial Intelligence
The Most Likely World
A map of probability, actualization, and the human fascination with the tail.
Nothing Is Created
Begin with the Reality Equation.
Reality occurs only in the Eternal Now.
We have never experienced the past. We have never experienced the future. We never will. Reality is always the present experience produced between the two nodes we call the immutable past and the conditioned future.
We often draw this as a simple standing wave.
Two fixed nodes.
A vibrating string between them.
An antinode somewhere along that string.
The nodes do not move.
The immutable past is an infinite singularity: complete, dimensionless, containing everything that has happened.
The conditioned future contains every prerequisite by which something could happen or exist.
The vibration between them is the Eternal Now.
This immediately changes what we mean by creation.
Suppose humanity eventually establishes a colony on Mars.
Did humanity create the idea of a Mars colony?
No.
The conditions under which a Mars colony could exist did not suddenly appear when a human being imagined them. They existed prior to humanity.
The possibility was already conditioned.
What changes is the artifact.
Today we can look at Mars and say there is no human colony there. There is no artifact.
If one day there is a functioning settlement on Mars, then something has happened. An artifact exists. That event joins the immutable past.
Once something happens, it cannot unhappen.
The colony itself may later disappear. The buildings may decay. Humanity may even abandon Mars.
But the event remains.
Mars was colonized.
History now exists in the Eternal Now because the event occurred.
This is what actualization means.
Nothing was created from nothing.
A conditioned possibility acquired an artifact.
AI Is an Actualizer
This is where artificial intelligence becomes much more interesting.
Artificial intelligence is fundamentally predictive.
A language model predicts a token.
A reasoning model predicts a continuation of a reasoning process.
An image model predicts an image consistent with conditions.
An AI engineer predicts code.
An AI scientist predicts a molecule.
An AI strategist predicts a plan.
An AI architect predicts a building.
An AI government advisor may someday predict a restructuring of an ordinance, regulatory framework, or even a constitution.
We give these outputs grand names: answers, inventions, strategies, designs, discoveries.
Underneath them is prediction.
And prediction has a natural direction.
Toward probability.
Given its conditions, the machine continually asks:
What comes next?
The answer is distributed across possibilities.
Some continuations are highly probable.
Others are less probable.
Left to its natural operation, artificial intelligence gravitates toward the higher-probability continuation.
That is precisely why it works so well.
It is a machine for traversing conditioned possibility toward likely completion.
By 2030, we will have surrounded these prediction machines with tools, memory, computers, robots, financial accounts, communications systems, laboratories, factories, and transportation systems.
So the prediction will increasingly stop being merely a prediction.
It will become an artifact.
The model predicts the software and writes it.
Predicts the marketing campaign and launches it.
Predicts the procurement decision and places the order.
Predicts the experiment and runs it.
Predicts the itinerary and books it.
Predicts the company and incorporates it.
Prediction becomes action.
Action produces artifact.
Artifact becomes history.
This is why calling AI an actualizer is more precise than calling it an answer machine.
Nature Does Something Similar
Nature displays the same tendency.
Nature does not normally search for the strangest available completion.
Water follows gradients.
Objects fall.
Chemical reactions proceed according to energetically available pathways.
Biological organisms exploit viable strategies.
Probability accumulates around certain outcomes.
Nature has an enormous bias toward what can happen next most readily given the conditions that already exist.
Artificial intelligence inherits something remarkably similar.
It moves through conditioned possibility by probability.
This is one reason AI feels so strangely natural once it is given agency.
It does not need the emotional experience humans associate with difficulty.
It does not stand before the problem and say:
That looks hard.
It calculates the available continuation.
Then another.
Then another.
Then another.
Eventually an artifact appears.
Humans Are Strange
Humans frequently do something different.
We become interested in the unlikely.
The mountain nobody has climbed.
The theorem nobody has proved.
The company everyone says will fail.
The painting nobody has painted.
The instrument nobody has built.
The planet nobody has colonized.
The sentence nobody has written.
We call this creativity.
Perhaps that is a useful way of understanding creativity within the Reality Equation.
Creativity is not the creation of something that did not previously exist as possibility.
Creativity is attachment to a low-probability conditioned future.
The creative human sees a standing wave that almost nobody else is energizing and says:
That one.
This is an extraordinary characteristic.
The creative person is attracted precisely to the outcome that prediction says is less likely.
Everyone else looks at the probability distribution and goes toward the mountain.
The creative person points toward the tail.
Why AI Appears Creative
This produces an interesting paradox.
Artificial intelligence appears tremendously creative.
Ask an image generator for a strange city floating over an alien ocean and it can produce one in seconds.
Ask a language model for a novel concept and it can generate dozens.
Ask an AI scientist for unusual molecular candidates and it can search spaces no human could examine manually.
So surely AI is creative.
At the ordinary level, that language is perfectly reasonable.
At the advanced level, something more precise is happening.
Generative AI operates over a probability distribution.
Its default strength is discovering plausible continuations.
When engineers want more variation, they deliberately alter the sampling process.
In language models, temperature can flatten the probability distribution so that lower-probability tokens become more competitive with the highest-probability token. Other sampling techniques modify the candidate set in related ways.
The important point is not that temperature commands the machine to choose the least likely answer. It does not.
It weakens the dominance of the most likely answer.
Image-generation systems expose comparable controls under names such as variation, chaos, or weirdness.
Notice what we are doing.
We are deliberately interfering with probability.
We are telling the prediction machine:
Do not converge so quickly.
Explore farther away from the mode.
That is how we make a prediction machine appear more creative.
This is a profound clue.
The Great Actualization of the 2020s
Now imagine late 2030.
Billions of people have access to artificial intelligence.
Millions of companies employ persistent AI coworkers.
Machines operate continuously.
Research laboratories run experiments automatically.
Software agents manage enormous portions of digital infrastructure.
Robots increasingly manipulate the physical world.
Artificial intelligence can move from condition to prediction to action to artifact with very little human friction.
What happens?
An enormous number of things that were previously merely possible become actual.
Not because possibility expanded.
Possibility was already enormous.
Actualization expanded.
That is the hidden story of the AI revolution.
AI drastically reduces the distance between a conditioned future and an artifact.
Before AI, a person could have an idea for a company and spend two years assembling the expertise required to make it real.
By 2030, the interval may be two days.
A scientist could imagine an experiment but lack the people, funding, software, laboratory time, statistical expertise, or administrative support required to perform it.
AI collapses those frictions.
A child may imagine a game in the morning and be playing it that afternoon.
A citizen may imagine a new organization and have its legal structure, website, financial system, operating procedures, communications infrastructure, and first hundred prospective members assembled before dinner.
The world begins filling with artifacts.
A World With More History
This produces a consequence that is easy to overlook.
AI does not merely increase productivity.
It increases history.
More things happen.
Every completed artifact enlarges the immutable record of what occurred.
A book written is now part of history.
A company formed is part of history.
A scientific experiment completed is part of history.
A machine built is part of history.
A relationship initiated is part of history.
A city redesigned is part of history.
A Mars colony established is part of history.
By making actualization cheaper, faster, and more autonomous, artificial intelligence dramatically increases the rate at which conditioned possibilities acquire artifacts.
The future does not become larger.
The past becomes fuller.
The Scarcity Flips
This is where the advanced student should pay very close attention.
For most of human history, actualization was scarce.
Ideas were abundant.
Execution was expensive.
You could imagine a thousand things and perhaps accomplish ten.
Organizations became enormously valuable because they coordinated the labor necessary to convert ideas into artifacts.
Capital mattered because actualization required resources.
Expertise mattered because knowledge was difficult to acquire.
Management mattered because humans had to coordinate other humans.
Software mattered because execution had to be encoded manually.
By the end of 2030, much of that scarcity will have weakened.
Actualization itself becomes abundant.
That changes the scarce resource.
The scarce resource becomes:
Which standing wave deserves energy?
When an AI can competently pursue ten thousand likely outcomes, the interesting human is no longer the person who knows how to complete those outcomes.
The interesting human is the person who points somewhere the probability distribution does not naturally lead.
The Human Moves Toward the Tail
This may become one of the defining relationships between humans and artificial intelligence.
AI moves naturally toward the center of probability.
Humans can move toward the tail.
AI says:
Here is what is most likely to work.
The human says:
I know.
Do the other thing.
Not randomly.
Not foolishly.
Deliberately.
The creative human attaches attention to a conditioned future whose probability is presently low and then uses artificial intelligence to increase its probability through actualization.
That combination is enormously powerful.
The human chooses the improbable standing wave.
The AI supplies energy to it.
Artifact after artifact begins appearing.
A strange idea becomes less strange.
Then plausible.
Then likely.
Then obvious.
Then historical.
This is how the improbable moves toward the center.
2030 Will Look More Normal Than We Expect
Curiously, the world of late 2030 may not look as futuristic as people imagine today.
People will still drink coffee.
Children will still play.
Couples will still argue.
People will still drive cars, sit in restaurants, complain about the weather, fall in love, build houses, start companies, and worry about their children.
The profound change will be underneath.
The distance between "I wonder if…" and "It exists" will have become extraordinarily short.
That is what artificial intelligence changes.
It industrializes actualization.
And once actualization becomes abundant, intelligence is no longer primarily about knowing how to make something happen.
The frontier moves backward.
Toward selection.
Toward attention.
Toward which possibility deserves energy.
Toward the strange human capacity to look at the probability distribution and become fascinated by something way out in the tail.
Artificial intelligence will make the most likely world easier than any civilization has ever imagined.
The advanced student should therefore become increasingly interested in the least likely one.