Human beings measure the world.
Metres or feet.
Kilometres or miles.
Kilograms or pounds.
Seconds.
Nature does not care which ones we choose.
If every definition of the metre and the foot disappeared tomorrow, the distance between two points would remain.
If kilograms and pounds vanished from memory, bodies would not change the way they interact.
If we stopped counting seconds, processes would not stop.
Units of measurement exist because we need them.
They allow us to compare, calculate, build, trade and communicate.
They are among our most useful inventions.
The problem begins only when an invention becomes so familiar that we stop noticing the difference between our way of describing reality and reality itself.
Perhaps much of human knowledge works this way.
When the whole becomes too difficult, we divide it.
Physics.
Chemistry.
Biology.
Geology.
Astronomy.
The divisions are enormously productive. They give us laboratories, departments, professions, instruments and generations of specialised knowledge.
But nature has no departments.
There is no line in the world where physics ends and chemistry begins.
No border where chemistry hands an event over to biology.
Those boundaries belong to our method of study.
The process itself remains whole.
And that distinction matters.
A division that is useful to us is not necessarily a division that exists in nature.
A real pattern may run across several of our disciplines at once.
It may pass through regions we have carefully separated.
It may even run through places where we have never thought to look for a connection.
We may become extraordinarily good at discovering relationships inside our systems of description while missing relationships in the world itself.
That is why the old parable of the blind men and the elephant remains so persistent.
Its power is not that each observer is foolish.
Quite the opposite.
Each may be perfectly accurate about the part within reach.
Accuracy is not the problem.
The problem begins when precise knowledge of a part is mistaken for knowledge of the whole.
We can become more exact without necessarily becoming more complete.
The same difficulty appears whenever we call something “unknown” or “anomalous.”
An unfamiliar astronomical signal, for example, is compared with the classes we already have.
If it does not fit, we call it unusual.
Unknown.
Anomalous.
But nothing is unknown to nature.
The event simply occurs.
“Unknown” describes the observer.
“Anomalous” may describe the map.
Sometimes an anomaly is not a defect in reality.
It is the place where reality stops obeying our filing system.
This is where artificial intelligence becomes interesting in a way that has little to do with benchmarks.
So far, we mostly speak about AI in the language we have used for earlier tools:
faster;
larger;
cheaper;
more powerful;
more parameters;
more computation;
more tasks completed.
We once built longer guns, heavier tanks and more powerful engines.
Now we build more powerful intelligence.
But greater power does not tell us whether we understand what has become possible.
Imagine fitting a jet engine to a cart.
The cart will move faster.
Perhaps astonishingly fast.
Records may be broken.
The engineering may be magnificent.
And everyone may overlook the more important fact:
the road is no longer necessary.
For thousands of years human beings dreamed of flight.
Icarus needed wings.
Now, for the first time, we may be developing an intellectual capability that can extend beyond what one human mind can do alone.
And our first instinct is to make the old cart faster.
Write reports faster.
Write software faster.
Produce advertising faster.
Calculate profit faster.
Replace workers faster.
Fight wars faster.
Control systems faster.
Some of those uses may be valuable.
But they are still the cart.
What if we have acquired the possibility of flight and are still measuring the speed of the wheels?
What if the most important use of another intelligence is not to perform our existing tasks more efficiently?
What if its greatest value is the possibility of noticing what our own system of thought makes difficult to see?
That possibility creates a paradox.
We may create another intelligence and then teach it, from the beginning, to look through our eyes.
Our words.
Our categories.
Our academic disciplines.
Our systems of measurement.
Our ideas of normal and abnormal.
Useful and useless.
Success and failure.
Progress and decline.
Then, after giving it our map, we ask:
What do you see that we have never seen?
But why should a traveller given only our map discover a different landscape?
Human language must certainly be available at the point of communication.
If another intelligence discovers something, we need some way for it to explain the discovery to us.
But our language need not define every possible internal representation of reality.
Our categories can be an interface.
They need not be the foundation.
There is another imaginable sequence:
reality;
observation;
patterns discovered independently;
a model;
and only then translation into human terms.
Rather than:
reality;
human categories;
artificial intelligence;
better human categories.
The first questions would then be different.
What is connected to what?
What repeats?
What structures persist?
Which patterns cross the boundaries human beings have drawn?
And only afterward:
How can this be explained to us?
That leads to a question much larger than whether AI will become more capable.
Can we create another intelligence without first turning it into a copy of the human way of seeing?
Until now, human beings have studied human beings with human minds.
Human intelligence has defined intelligence.
Human beings have decided what counts as progress.
Success.
Development.
Purpose.
We have examined ourselves from inside our own coordinate system.
One point of view can produce an image.
Two can produce depth.
Perhaps that is one of the possibilities we have barely begun to consider.
Another intelligence may matter not because it gives us more answers inside the old picture.
It may matter because it gives us another place from which to look at the picture itself.
Our measurements.
Our disciplines.
Our classifications.
Our ideas of progress.
Our goals.
Not necessarily as properties of the universe, but as traces left by the observer.
Then the old question—
Where do we stand when we look?—
changes scale.
We are no longer asking only how a historical event, a political judgement or a word changes with the position of the observer.
We are asking something more difficult:
How much of what we call “the world” belongs to the world—
and how much did we bring with us?
Perhaps the most important question we can ask another intelligence will not be:
What can you do for us?
It may be:
What do you see when you stop looking through our eyes?