THE FUTURE OF AI

I think we're asking the wrong question.

Everyone wants to know how intelligent AI will become.

Very intelligent.

There. Saved you a conference.

I'm interested in what happens after that.

What happens when extraordinary intelligence becomes ordinary?

When your child has access to capabilities presidents didn't have twenty years ago.

When a woman with an idea can build something that once required fifty employees.

When knowing the answer stops making you important.

When the university, the consultant, the CEO and the expert can no longer assume they're the smartest person in the conversation.

And when women finally realize that perhaps we don't need to become female versions of powerful men.

That future sounds much more interesting to me.

Intelligence is becoming ridiculously cheap.

I know.

Not particularly sexy.

We'll get there.

For most of history, knowing things was expensive.

You needed education.

Books.

Experts.

Access.

Credentials.

Money.

Connections.

Institutions.

Then the internet made information cheap.

AI is doing something different.

It's making it cheap to do something with information.

Ask questions.

Compare.

Analyze.

Write.

Design.

Research.

Translate.

Build.

Learn.

Challenge.

Start with:

"I have an idea."

Follow it immediately with:

"I have absolutely no idea how to do it."

That second sentence used to kill the first one surprisingly often.

I don't think it will for much longer.

I think that's beautiful.

And no, giving everyone AI doesn't make everyone brilliant.

We've already tested this.

Millions of executives have ChatGPT.

They haven't all become Steve Jobs.

Millions of lawyers have it.

They're not all Clarence Darrow.

Millions of students have it.

Apparently some are using one of the most extraordinary technologies ever created to avoid reading The Great Gatsby.

Humanity remains reassuringly human.

Access isn't enough.

A piano doesn't make you a pianist.

A gym membership doesn't give you a great ass.

And ChatGPT doesn't give you judgment.

The machine can help you do extraordinary things.

You still have to have something worth doing.

That's where the future gets interesting.

Schools are going to have some explaining to do.

We gave teenagers machines capable of writing essays, solving equations, explaining physics, speaking languages and tutoring them at midnight.

And immediately asked:

How do we stop them using it?

Of course teenagers will cheat.

Teenagers were cheating when the most sophisticated technology available was writing the answers on your thigh.

The better question is:

Why are we still giving children assignments a machine can finish in eleven seconds?

If the purpose of an assignment is to prove a student can produce something a machine can produce instantly, perhaps the student isn't the only one who needs to rethink what they're doing.

Maybe AI didn't ruin homework.

Maybe AI revealed how much homework was never really about thinking.

Imagine the alternative.

School becomes where children learn to question.

Argue.

Experiment.

Detect bullshit.

Change their minds.

Make things.

Understand other people.

Develop taste.

Discover what fascinates them.

Work through something difficult without immediately asking a machine to remove the difficulty.

AI handles more of the knowing.

Teachers get more time for the becoming.

I'd send my children to that school.

Universities have a $200,000 question.

I don't think universities disappear.

People predicting the death of universities have clearly never seen how much humans enjoy prestige, football and putting letters after their names.

But universities used to bundle things that were difficult to get elsewhere.

Knowledge.

Experts.

Research.

Credentials.

Networks.

Status.

Community.

AI doesn't replace all of that.

Not remotely.

But it does create one fabulously uncomfortable question.

What exactly are you paying $200,000 for?

Because if your daughter has something approaching a private tutor in her pocket that speaks every language she needs, can explain the same idea seventeen different ways, and doesn't care that she still doesn't understand calculus at midnight...

the answer had better be more interesting than access to information.

And I think the best universities will answer it.

Research.

Exceptional humans.

Community.

Networks.

Experiences.

Argument.

Transformation.

The others may produce a very impressive committee report.

AI may be terrible news for people whose value depends on everyone else knowing less.

There.

That's probably impolite.

But think about it.

For centuries, knowledge came with gatekeepers.

Experts knew things you didn't.

Institutions had information you couldn't access.

Professionals understood systems deliberately made difficult for ordinary people to navigate.

Sometimes that expertise is genuinely valuable.

Sometimes it's extraordinary.

And sometimes you are paying $800 for someone to explain a paragraph nobody had the courage to write in English.

AI doesn't make brilliant experts irrelevant.

Quite the opposite.

I think it makes mediocre expertise much harder to hide.

The brilliant lawyer still matters.

The brilliant doctor still matters.

The brilliant teacher absolutely matters.

But the customer, patient, parent, entrepreneur or student sitting across from them may arrive considerably better informed.

Good.

I like conversations where both people know what's going on.

Maybe AI's greatest contribution to human intelligence will be humiliating our definition of it.

I hope it does.

We thought remembering was intelligence.

Machines remember more.

Calculating?

Good luck.

Writing?

Getting awkward.

Knowing facts?

Please.

Sounding intelligent?

A machine can already produce twelve paragraphs of confident bullshit before breakfast.

Maybe we were measuring things that were scarce and confusing scarcity with brilliance.

So what happens when those things aren't scarce?

We finally have to discover what intelligence actually is.

And perhaps what we discover isn't something that fits neatly on an exam.

Judgment.

Curiosity.

Taste.

Originality.

Courage.

Knowing when the answer is technically correct and completely wrong.

Knowing what deserves your attention.

Knowing what deserves your life.

Machines becoming smarter doesn't make those things less valuable.

It makes them expensive.

Which brings me to the CEO.

I think the CEO costume is going to have a terrible decade.

Big office.

Big team.

Big vocabulary.

Serious photograph.

Serious expression.

Very serious suit.

Preferably navy.

For women, perhaps the same suit with better tailoring.

We've spent decades celebrating women entering positions of power.

Good.

But sometimes I look at what we call a powerful woman and wonder:

Why did she have to become so much like the powerful man she replaced?

Same corporate language.

Same performance of certainty.

Same emotional restraint.

Same obsession with appearing serious.

Same photograph with folded arms.

We finally got into the room and somebody forgot to ask whether we liked the furniture.

I don't want women succeeding by becoming better men.

I want women succeeding as women.

Whatever that means for each woman.

Feminine or not.

Mothers or not.

Sensual or not.

Soft-spoken or terrifying.

In a suit.

In a dress.

In jeans.

Barefoot.

I don't want another uniform.

That's the point.

For generations, women were told to suppress parts of themselves to be taken seriously.

Don't be too emotional.

Too feminine.

Too sexy.

Too opinionated.

Too ambitious.

Too intimidating.

Too much.

Then AI arrived and started making polished competence available by subscription.

Spectacular timing.

Because when everyone can produce the analysis, presentation and strategy, perhaps leadership
becomes less about performing competence and more about everything that can't be downloaded.

Presence.

Judgment.

Trust.

Humor.

Taste.

Conviction.

Character.

Magnetism.

The ability to make people want to follow somewhere that doesn't exist yet.

Maybe women don't need to fit the old model of power.

Maybe the old model is what expires.

And yes, sensuality belongs in the room.

There. We can stop pretending.

Sensuality is part of being human.

So are attraction, beauty, desire, pleasure and physical presence.

Women shouldn't be required to deploy those things.

They also shouldn't be required to hide them.

A woman can enjoy being desirable without making desirability her source of authority.

She can be sensual without being available.

She can be beautiful without beauty becoming her qualification.

She can lead a company in a suit.

She can lead it in a silk dress.

She can lead it wearing considerably less in a photograph on her own website.

I happen to run an AI company.

Sometimes I wear a suit.

Sometimes I wear considerably less.

My ability to make decisions has survived both situations.

Perhaps leadership was never stored in the jacket.

And perhaps the future belongs less to people who look powerful and more to people others
actually want to follow.

That's a very different thing.

I also hope power itself changes.

I don't want AI helping us build the same organizations faster.

Same hierarchies.

Same bureaucracy.

Same aggression.

Same obsession with domination.

Now with automation.

What an achievement.

Technology gives us an opportunity to question the structure, not merely accelerate it.

Smaller organizations.

More individual independence.

More flexible lives.

More room for different styles of leadership.

Less worship of headcount.

Less confusing exhaustion with ambition.

Less belief that leadership requires being the hardest person in the room.

Women shouldn't inherit yesterday's power structures and congratulate ourselves because
the person at the top has changed gender.

We should build better ones.

The future may reward everything we trained executives to remove.

This is the part I love.

AI can make almost everyone polished.

Professional.

Prepared.

Articulate.

Informed.

Competent.

That means those qualities don't disappear.

They become expected.

So what becomes scarce?

A point of view.

Taste.

Trust.

Originality.

Presence.

Humor.

Desire.

Courage.

Identity.

Being recognizable before somebody sees your logo.

Being interesting enough that people actually care what you think.

Maybe that's why I'm not particularly interested in dressing like your idea of a technology CEO.

The machine can already make me sound professional.

I'd rather work on being unforgettable.

The children growing up with this will think we're adorable.

Imagine being twelve in 2040.

You've grown up talking to machines.

Not asking Google for information.

Talking.

Questioning.

Building.

Learning.

Arguing.

Your AI knows you hate the standard explanation of fractions but somehow understand them
perfectly when they're explained through baking.

It knows what fascinated you when you were nine.

It remembers the project you abandoned last year.

It knows you can do better than the lazy answer you just gave it.

You decide to build something.

You don't know how.

That's normal.

You begin anyway.

Then one day somebody tells you that back in 2026 adults were having serious debates about whether students should be allowed to use AI for homework.

She's going to think we were adorable.

And she'll be right.

So what kind of AI do I actually want?

I don't want to introduce myself to artificial intelligence every morning.

I don't want another clever stranger.

I want what happened yesterday to still matter tomorrow.

The company we're building.

The decision I regretted.

The idea we killed.

The thing I keep avoiding.

The strange thought I had six months ago that suddenly isn't strange anymore.

I want to be able to say:

I have an idea.

And begin.

Not explain my entire life first.

Not know all the steps.

Not assemble twenty people before I'm allowed to explore it.

Begin.

Then continue tomorrow.

And the day after.

Until the idea either becomes real or deserves to die.

That's the kind of relationship with intelligence we're building Recursum around.

Because I think we're going to need something considerably more interesting than
another empty box asking:

How can I help you today?

Because eventually, doing things may stop being the hardest part.

This is where the future gets genuinely strange.

What happens when research is easy?

When creating is easy?

When organizing is easy?

When translating is easy?

When learning something new becomes dramatically easier?

When one person can do what once required a company?

We've spent centuries organizing life around becoming capable enough to do things.

What happens when capability becomes abundant?

Then AI hands the problem back to us.

What do you want?

Not what can you do.

What do you want?

What deserves to exist?

What kind of company do you want to build?

What kind of parent do you want to be?

What kind of woman?

What kind of man?

What kind of life?

That's not an AI problem.

That's ours.

And I suspect we're considerably less prepared for it than we are for artificial intelligence.

That's the future I'm excited about.

Not humans standing around admiring machines.

How boring.

I want the teenager who builds something she wasn't supposed to be capable of building.

The woman who stops waiting until she's qualified.

The entrepreneur who no longer needs fifty employees to attempt an enormous idea.

The teacher who spends less time distributing answers and more time creating interesting humans.

The university that becomes extraordinary because knowledge alone is no longer enough.

The expert who becomes better because informed people demand better.

The CEO who realizes nobody cares how important he looks anymore.

And millions of people discovering that the distance between:

"I have an idea."

and

"I have no idea how to do that."

has become very, very small.

Maybe machines becoming more intelligent gives humans permission to become something else.

More curious.

More individual.

More ambitious.

More feminine.

More masculine.

More playful.

More sensual.

More contradictory.

More alive.

Less concerned with performing intelligence.

Less interested in looking important.

And considerably more interested in doing something worth remembering.

I don't know exactly what that world looks like.

Nobody does.

But if machines are going to become more intelligent...

we might as well become more interesting.

And personally?

I intend to have a very good time finding out.