Are we going to miss the artificial intelligence train too?

When we look at the world of artificial intelligence, we are still asking the same questions: Is OpenAI ahead, or Anthropic, or Google?

I think we are looking in the wrong place.

Because the artificial intelligence race is no longer just a race to see who can build a better model. The business is heading toward something much bigger.

The equation today is roughly as follows:

Model + computing power + energy + distribution + capital + state.

Whoever can possess all of these at the same time will be the powerful player of the artificial intelligence era.

What happened in Washington in recent days was actually an indicator of this. Executives of major artificial intelligence and technology companies are now appearing not only in Silicon Valley but also in Washington.

We shouldn't be surprised.

Because artificial intelligence is no longer just a technology issue.

It is an energy issue.

It is an economic issue.

It is a national security issue.

It is a defense issue.

And increasingly, it is a foreign policy issue.

The interesting part is that while artificial intelligence companies say, “This technology could be dangerous, we must be careful,” they are simultaneously spending billions of dollars to build larger data centers, more processing power, and more powerful models.

On one hand, they talk about hitting the brakes, while on the other, they are pressing the gas pedal to the floor.

One of the important reasons for this is money.

Artificial intelligence is a very expensive business.

You need chips. You need data centers. You need energy. You need water. You need engineers. And for all of these, an incredible amount of capital is required.

That is why artificial intelligence companies need to keep their future story alive.

Here we come to the AGI issue.

AGI, the idea of artificial intelligence approaching the general mental abilities that humans can demonstrate in different fields, has been one of the industry's biggest promises for years.

I am not convinced that we will reach AGI, let alone superintelligence, by simply scaling up existing large language models.

But I also don't think this diminishes the importance of artificial intelligence.

Because for a very important contribution that artificial intelligence can make to humanity, it does not need to be AGI.

I call this “Move 37.”

In 2016, Google DeepMind's AlphaGo system faced Lee Sedol, one of the world's best Go players. In the 37th move of the second game, AlphaGo made such a move that experts were astonished. Because it was a move that seemed extremely unlikely for a human to choose.

Later, it was understood that the move was one of the critical decisions that changed the course of the game.

I think the real importance of artificial intelligence lies here.

Its ability to see things we cannot see.

We didn't do everything perfectly while building civilization.

We couldn't have anyway.

We created law, and we built the economy on top of it. We developed technology, and we built other systems on top of that. We created institutions. We wrote rules. Then we added new rules on top of the old ones.

What emerged was a civilization that looks a bit like a Jenga tower.

One block is a bit to the right.

The other is a bit to the left.

But the tower is still standing.

Now artificial intelligence is coming, and it is starting to notice that some of the blocks we have been looking at for years without seeing are in the wrong place.

That is why I think we sometimes ask the wrong question regarding artificial intelligence safety.

We ask, “Will artificial intelligence threaten our safety?”

Perhaps we should first ask this:

Have we really built a safe system?

In some cases, artificial intelligence may not be creating new vulnerabilities as much as it is making visible the vulnerabilities we have failed to notice for years.

But immediately after this comes a much more difficult question.

Ethics.

Let's say artificial intelligence found a solution we couldn't see.

Who is it good for?

It might be good for me.

It might be good for a company.

It might be good for a state.

But is it good for society?

Is it good for humanity?

The “Zeroth Law” that Asimov later added to his famous robot laws touched exactly on this problem: A robot may not harm humanity.

Sounds nice.

But the second question arises immediately:

Who will decide what is in the interest of humanity?

I think this will be one of the most important philosophical issues of the artificial intelligence era.

Because humanity loves dualities when thinking.

Black or white.

Good or bad.

Me or society.

Freedom or security.

Human or machine.

Perhaps the new Move 37s we should expect from artificial intelligence will emerge right here.

Solutions where society does not have to lose while I win.

Solutions where people are not harmed while a company wins.

Solutions where the individual does not lose their freedom while the state grows stronger.

In other words, solutions where we can say “both this and that” instead of “either this or that.”

So, should we be afraid of the speed while all this is happening?

I don't think so.

Humanity has already been constantly accelerating.

First, we ran.

Then we rode horses.

Then we got into cars.

Then we got on planes.

Today, fighter pilots can perform under accelerations that the human body would never encounter in its natural life.

And humans have adapted to this.

If we define intelligence as the ability to adapt to changing conditions, adapting to the upcoming technological change will be the new test of human intelligence.

Now let's come to Turkey.

I met the internet in 1991, the year I started Boğaziçi University.

At that time, there was a metaphor I used frequently when trying to explain the internet:

A train is coming.

Whatever we do, let's get on that train. Let's jump into the last carriage if necessary, but let's not miss it.

We missed it.

We missed the internet train.

After that, we also missed the social media train.

We could not build the world's largest internet companies. We could not produce the social networks used by the world. We largely became users of technologies developed by others.

Now artificial intelligence is coming.

Moreover, it is coming much faster than the internet.

This time, our chance of catching up by running from behind is lower.

That is why it is not enough for Turkey to say, “Let's also build an artificial intelligence model.”

We need a leap.

First, we need data.

I am not just talking about artificial intelligence that speaks Turkish.

I am talking about artificial intelligence that knows Turkey.

We need high-quality training data that can understand Turkish, our culture, our history, our literature, our world of thought, our daily life, and our concepts.

Second, we need computing power.

Because artificial intelligence does not work with slogans.

It works with computers.

Third, we need energy.

Massive data centers consume massive amounts of energy. Therefore, we cannot think of the artificial intelligence policy of the future independently of energy policy.

Turkey needs to turn its renewable energy capacity, especially solar, into a strategic advantage here.

Another issue is quantum computers.

It might sound a bit distant today.

But in the past, warships, planes, or aircraft carriers were indicators of a country's strategic power.

One of the strategic infrastructures of the world ahead will be computing capacity. We do not have the luxury of seeing quantum technologies as just a distant scientific subject that universities are interested in.

But above all of this, there is perhaps something more important.

Culture.

And philosophy.

Because artificial intelligence is not just code.

What you teach artificial intelligence is also what artificial intelligence is.

For this reason, Turkey does not need only a “domestic and national artificial intelligence” slogan.

Turkey needs an artificial intelligence ecosystem based on its own data, culture, energy capacity, computing infrastructure, capital, and human resources.

We missed the internet train.

We also missed the social media train.

This time, we may not have time to run after the artificial intelligence train.

Perhaps it is time to stop trying to catch the train and learn to leap instead.