Pete Hegseth went to Quantico this week to talk about the future of war.

Some of what he said made sense.

The battlefield is changing. Fast.

Drones are everywhere. Artificial intelligence is moving into targeting, surveillance, logistics and command systems. Cheap autonomous weapons can destroy equipment costing millions of dollars.

Ukraine has already demonstrated that.

Hegseth wants the Pentagon to catch up.

Good.

But then came the part that ought to make every soldier, sailor, Marine and airman stop for a minute.

Hegseth announced Project Meridian, a 120-day effort to look beyond today’s battlefield and determine what weapons and technologies the United States will need for the wars of the future.

And he named three men to help lead it:

Elon Musk.

Palmer Luckey.

Newt Gingrich.

Not one of them served in the United States military.

Not one of them has fought in a war.

Think about that.

We are going to study the future of warfare without putting a veteran of warfare among the three men Hegseth singled out to lead the study.

That’s not an argument against bringing civilians into the room.

Bring them.

Musk knows rockets, satellites, communications and manufacturing. SpaceX has become deeply entwined with American national-security infrastructure.

Luckey knows autonomous weapons. His company, Anduril, is building exactly the kind of systems the Pentagon says it needs.

Gingrich has spent decades thinking and writing about government, national security and technological change.

They belong at the table.

But there is another kind of knowledge.

It comes from getting shot at.

And that’s the knowledge missing from those three chairs.

WAR ISN’T A POWERPOINT PRESENTATION

A drone looks different on a manufacturer’s screen than it does when somebody is trying to kill you with one.

A communications network looks terrific until the enemy jams it.

Artificial intelligence looks wonderful until it identifies the wrong target.

A logistics system looks efficient until the bridge is gone.

And an autonomous weapon looks revolutionary until the battery dies, the GPS disappears or the enemy figures out how to spoof it.

Soldiers have a technical term for this.

War.

Hegseth himself seems to understand the problem.

In the same speech, he announced plans for a new four-star Autonomous Warfare Command — AutoWarCom — intended to push drones, robotics and artificial intelligence throughout the joint force.

The command is supposed to be established by October 2027, subject to the congressional action necessary to create it.

Until then, the Pentagon will use something called Project Agincourt.

That may be the more important experiment.

Agincourt is supposed to put operators and entrepreneurs together, shorten the distance between the battlefield and the procurement system, and let troops tell engineers what actually works.

That’s exactly what should happen.

Because the guy carrying the equipment usually discovers its weaknesses long before the guy who bought it.

WE HAVE BEEN HERE BEFORE

I remember sitting at an AFCEA/USNI conference in San Diego in 2012 listening to retired Marine Gen. James “Hoss” Cartwright talk about taking the human out of the airplane.

His argument was simple.

Why build an aircraft around a human being?

A pilot needs oxygen.

A pilot needs an ejection seat.

A pilot can tolerate only so many Gs.

A pilot has to eat, sleep and eventually come home.

Take the pilot out and you can build a very different machine.

Park it somewhere. When you need it, fire it up.

That sounded futuristic in 2012.

It doesn’t anymore.

The Air Force is developing Collaborative Combat Aircraft designed to operate alongside crewed fighters. The Navy is working on unmanned aircraft. Across the services, autonomous systems are moving from experiments toward actual combat formations.

Last year, I wrote that it was time to rewrite the infantry manual because drones had already changed ground combat.

Now the Pentagon is essentially saying the same thing.

Hegseth said the lesson from Ukraine isn’t that one particular drone wins wars. The lesson is that military forces have to adapt constantly because the technology and countermeasures keep changing.

He’s right about that.

The enemy gets a vote.

And nowadays the enemy may get a software update Tuesday morning.

SO ASK THE PEOPLE WHO HAVE BEEN THERE

This is where Project Meridian bothers me.

At the same time Hegseth is proposing another four-star command, he is pushing a roughly 20 percent reduction in generals and admirals.

There is certainly fat in the Pentagon.

There always has been.

The military doesn’t need a general standing behind every colonel holding a PowerPoint clicker.

But those stars also represent something Silicon Valley cannot manufacture.

Experience.

Some of those officers commanded troops in Iraq.

Some fought in Afghanistan.

Some operated ships in the Persian Gulf.

Some flew combat missions.

Some watched expensive Pentagon programs work beautifully during testing and fail miserably after meeting the enemy.

That doesn’t make generals automatically right.

History provides a warehouse full of generals who were wrong.

But combat experience deserves a chair at the table.

So does enlisted experience.

Especially enlisted experience.

If we’re designing the battlefield of 2040, find the staff sergeant who spent months dodging drones.

Find the Navy chief who kept a ship operating after something important broke.

Find the pilot who watched the electronic-warfare picture disappear.

Find the infantry lieutenant who discovered that the $100,000 piece of equipment headquarters loved was too heavy to carry up a mountain.

Then put Musk and Luckey across the table from them.

Now you’ve got something.

Engineer meets operator.

Silicon Valley meets mud.

Theory meets incoming fire.

That’s how good weapons get built.

THE MACHINE ISN’T THE WAR

There is another danger here.

We are becoming fascinated with the machine.

Autonomous drones.

Artificial intelligence.

Robot ships.

Unmanned aircraft.

Super intelligence.

All of it matters.

None of it changes the oldest rule in warfare.

Somebody still has to accomplish the mission.

Technology changes how wars are fought.

It doesn’t repeal friction, fear, exhaustion, bad intelligence, broken equipment, lousy weather, confused orders or human stupidity.

Clausewitz didn’t have drones.

He would have recognized all the rest.

Hegseth’s Autonomous Warfare Command may turn out to be one of the most consequential Pentagon reorganizations in decades.

Project Agincourt may finally force the procurement bureaucracy to listen to the people using the weapons.

And Musk and Luckey may have valuable ideas about technologies the military hasn’t even imagined yet.

Fine.

Listen to them.

But before Project Meridian starts designing the wars our children and grandchildren may have to fight, somebody should drag another chair into the room.

Put a warfighter in it.

Then ask the simplest question in the Pentagon:

Will this thing actually work when somebody is shooting back?