What is Then?
Yesterday. Today. Tomorrow.
What Is Then?
We understand now.
Now is where we are. It is what we can see, measure, experience, and respond to.
We understand before, too—or at least we try to. Before is recorded in photographs, documents, memories, data, decisions, successes, and mistakes.
But what exactly is then?
Then is different.
Then hasn't happened yet.
And in an age of artificial intelligence, we may be becoming increasingly tempted to treat then as though it already has.
We Have Become Very Good at Predicting
Modern technology is built partly around anticipation.
Weather models tell us what conditions may look like tomorrow. Algorithms estimate what we might buy. Navigation systems calculate when we'll arrive. Businesses forecast demand. Financial models project future performance.
AI takes this considerably further.
Give an AI system enough information and it can generate scenarios, identify patterns, estimate outcomes, and suggest what might happen next.
It can create a remarkably convincing picture of then.
But a picture of the future isn't the future.
Prediction is not knowledge.
That distinction is becoming increasingly important.
Then Is a Possibility, Not a Place
Suppose you're redesigning a website.
You can study visitor behavior, identify confusing sections, reorganize information, improve navigation, simplify the writing, and create clearer calls to action.
You might even use AI to analyze all of it.
Eventually, though, you have to publish the new site.
And then something happens.
Real people encounter it.
They interpret it.
They click things you didn't expect them to click.
They ignore something you thought was important.
They ask questions you never anticipated.
They behave like people.
That is then becoming now.
And reality gets a vote.
AI Can Simulate the Future—But It Cannot Visit It
This may be one of the easiest things to forget about artificial intelligence.
AI can generate a plausible future.
It can generate ten plausible futures.
It can compare them, describe their advantages and disadvantages, and help us prepare for each one.
But it hasn't been there.
Neither have we.
AI works from information that already exists. Even when it produces something genuinely surprising, it is operating from patterns, relationships, instructions, and information available in the present.
The future contains something none of those systems fully possess:
events that haven't happened.
A competitor changes direction.
A customer reacts unexpectedly.
A new technology appears.
Someone has an idea nobody considered.
A cultural preference shifts.
A seemingly insignificant event changes everything that follows it.
Then changes.
That's Not a Weakness
Uncertainty can sound like a problem to be eliminated.
Perhaps it isn't.
If everything about the future could be calculated perfectly, there would be little room for discovery.
There would be no genuine surprise.
No unexpected breakthrough.
No idea that changes the assumptions behind the original question.
Some of the most important developments occur precisely because reality doesn't behave according to the plan.
The unknown future isn't merely a limitation.
It is where possibility lives.
The Goal Isn't to Predict Everything
This changes how we might think about AI.
Perhaps we shouldn't primarily ask AI to tell us what will happen.
We can ask it to help us think about what could happen.
What's one possible outcome?
What's another?
What are we assuming?
What might change those assumptions?
What haven't we considered?
What would we do if the opposite happened?
Where is our plan unnecessarily dependent on a particular prediction?
Those questions don't ask AI to eliminate uncertainty.
They use AI to help us become better prepared for it.
Then Eventually Becomes Now
Every plan eventually encounters reality.
Every prediction eventually meets an outcome.
Every forecast eventually becomes something we can compare with what actually happened.
Until that moment, then remains something peculiar.
It is not a fact.
It is not a destination we can inspect in advance.
It is a collection of possibilities shaped by what we know, what we choose, what other people choose, what changes—and what we never saw coming.
AI may help us imagine those possibilities more clearly than ever before.
But perhaps wisdom lies in remembering the boundary.
Before gives us evidence.
Now gives us choices.
Then gives us possibilities.
And no matter how sophisticated our technology becomes, there will always be some distance between imagining what comes next—
and discovering it.