Absorption / Field Note 01
What Gets Absorbed Disappears
The product remains. The process leaves consciousness. That is the deeper transformation hiding inside AI.
Uncertainty
Prediction
Outcome
Released attention
The object does not vanish.
The demand for attention does.
AI is not first a worker. AI is first a synthetic prediction machine.
It does not manufacture the outcome in the world. It predicts the outcome.
When the outcome is made of language, this distinction can be difficult to see because prediction and production happen almost simultaneously. The model predicts the next token. Software applies the recursive rules of language—grammar, syntax, structure, sequence—and stores the actual artifact. Token follows token until a complete language product appears.
The model contains a compressed representation of a combinatorially vast—effectively unbounded—language space. The article was not sitting inside the model waiting to be retrieved. The model predicts a path through that space, one token at a time.
That is technically important, but it is not the central event.
The central event is what happens to human attention once the outcome becomes predictable.
01The product remains
The burden was never the typing.
Today, a person may spend six hours writing an article.
The article occupies the person’s attention. The writer must decide how to begin, what belongs in the middle, which example will work, whether the argument is clear, and how the conclusion should resolve the opening.
The physical typing is not the real burden.
The burden is uncertainty.
The next sentence has not yet been determined. The writer must continuously predict it.
Tomorrow, the same person may describe the intended article, inspect the predicted result, make a few judgments, and move on.
The article does not disappear. Quite the opposite. Vastly more articles will exist.
What disappears is the human attention previously required to produce each one.
Automation means that a machine performs an action previously performed by a human.
Absorption means that an activity becomes sufficiently predictable that it no longer requires sustained human attention.
The outcome remains in the world.
The process disappears from consciousness.
02The predictable recedes
What becomes predictable becomes invisible.
Human beings do not attend equally to everything around them.
Attention is pulled toward mismatch, surprise, uncertainty, and unresolved conditions.
You do not continuously attend to the chair supporting your body. You attend to it when it breaks.
You do not attend to every breath. You attend to breathing when it becomes difficult.
You do not attend to the spelling of every familiar word. You attend when a word looks wrong.
The predictable recedes. The surprising advances.
RealityActual ÷ Expectation
When Actual ≈ ExpectationReality → 1
SurpriseLog(1) = 0
In the Reality Equation, there is one Reality value at each micro-instant: Actual divided by Expectation. When Actual closely matches Expectation, Reality approaches one. The natural logarithm of one is zero. There is no meaningful surprise and therefore no new demand for attention.
This does not mean nothing happened.
The chair still supported you. The lungs still breathed. The sentence was still written.
It means the event did not rise into consciousness.
Human attention is normalized accumulated surprise. AI is beginning to make many intellectual outcomes predictable enough that a human no longer has to attend to every step.
03Three absorptions
The outcome grows.
The routine recedes.
CASE / CASH
The bank teller and the ATM
For decades, a significant portion of a teller’s day was spent dispensing paper currency. A customer entered the bank, presented identification, requested money, and waited while the teller completed the transaction.
The automated teller machine absorbed that work. It did not eliminate cash. More people gained access to cash, at more locations, during more hours of the day.
The predictable withdrawal left the teller’s field of attention. The disputed charge, unusual account problem, suspected fraud, and consequential financial decision remained.
CASE / IMAGE
The radiologist and the image
A radiologist does not manufacture the diagnosis. The radiologist predicts the meaning of the image. Some patterns are familiar. Others are ambiguous, rare, contradictory, or consequential.
When an AI system becomes reliably useful at identifying familiar patterns, it can absorb part of that prediction work. More images can be examined. More screenings can be performed. More subtle patterns can be compared against prior cases.
The routine pattern recedes. Medical attention moves toward uncertainty and exception. The clinical outcome does not disappear. Attention is redistributed.
CASE / CALCULATION
When “computer” was a job
Before electronic computers, human computers performed calculations for engineering, navigation, astronomy, insurance, military operations, and scientific research. Their work required concentration, training, and time.
Then machines absorbed calculation. The occupation largely disappeared. Calculation exploded.
Once computation no longer required a human being to consciously execute each operation, civilization could place it everywhere—and attend to an entirely new layer of problems.
04The remains register
Absorption is not annihilation.
The ATM did not merely perform a teller’s action. It removed a stable pattern from the teller’s field of attention.
The computer did not merely perform yesterday’s calculations faster. It changed what human beings were able to notice.
Use the register below to separate the outcome that stays in the world from the routine uncertainty that stops bidding for consciousness.
What disappears is not necessarily the product, the need, or the value. It is the stable pattern’s claim on human attention.
Human attention is finite. Every unresolved task is bidding for it.
05The competition for attention
The important thing can still lose.
At any moment, countless things are present in the observable field, but only a small number can command conscious attention.
A business owner may care about strategy, employee development, customer relationships, family, community, beauty, ethics, purpose, and the future.
But if payroll is wrong, a proposal is due, a contract must be written, and twenty customer emails remain unanswered, those mismatches dominate attention.
The other concerns do not cease to exist. They are snuffed out by larger surprises.
The owner does not ignore purpose because purpose is unimportant. Purpose loses the competition for attention.
A teacher may care deeply about a struggling student, the meaning of education, the intellectual development of the class, and whether the material is changing lives.
But grading, scheduling, reporting, preparing documents, and answering routine questions may consume the available attention first. The subtle concern is still present. It simply cannot become loud enough to be heard.
AI changes this competition.
When the proposal becomes predictable, it no longer demands six hours. When the standard contract becomes predictable, it no longer occupies the attorney’s entire afternoon. When the routine lesson plan becomes predictable, it no longer consumes the teacher’s evening.
When the report, summary, response, analysis, and presentation become predictable, they begin to recede from attention.
The language products remain.
Human attention moves.
06What was snuffed out
What becomes visible after the old work disappears?
The most important consequence of AI may not be what it produces.
It may be what becomes visible after the old work disappears.
The deeper question is not, “What job will AI replace?”
The deeper question is, “What has been present all along but unable to command attention?”
Perhaps it is the customer whose real problem never fit neatly into the form.
Perhaps it is the student who needed to be seen rather than graded.
Perhaps it is the patient whose case did not resemble the familiar pattern.
Perhaps it is the strategic question buried beneath constant reporting.
Perhaps it is philosophy, theology, family, beauty, nature, community, or meaning.
These things are not created by artificial intelligence. They were already there. They were simply losing the competition for attention.
Yesterday, the human being attended to the article because the article was unresolved. Tomorrow, the article will be predicted. The person will attend to something else.
That “something else” is the real future of work.
07The unresolved boundary
AI does not eliminate work.
It is tempting to imagine that once a task is absorbed, the human being is left with nothing to do.
History suggests the opposite.
When calculation became abundant, humanity found more things to calculate.
When cash distribution became abundant, more transactions occurred.
When medical prediction becomes abundant, more people can be screened and more conditions can be examined.
When language outcomes become abundant, organizations will produce more agreements, explanations, documentation, education, analysis, and communication.
But abundance changes the location of human effort.
Some existing occupations will shrink. Some will disappear. New occupations will emerge.
But the deepest change will happen before any job title is invented.
The predictable work will stop bidding for attention.