Something unusual happened at the White House today, and I don't think the biggest story is the meeting itself. It may be the name.

On September 29, President Donald Trump met with some of the most powerful people building artificial intelligence, including executives and leaders from companies such as Google, Meta, Nvidia, OpenAI, Anthropic and others. The discussion centered on the development, safety and oversight of increasingly powerful AI systems.

But alongside that meeting came something considerably more symbolic.

Artificial Intelligence Is Becoming Super Intelligence

Trump signed an executive order directing the executive branch, to the extent permitted by law, to begin using “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI.”

That might sound like branding. Maybe it is. But I think there's something much bigger buried inside those two letters.

AI Described What We Built. SI Describes What We Think It Is Becoming.

The term *artificial intelligence* has always carried a subtle qualifier: artificial. Manufactured. Synthetic. Something attempting to do something humans do naturally.

But that description is becoming increasingly uncomfortable as these systems improve. Today's frontier models don't simply retrieve information. They reason across problems, write software, analyze enormous datasets, generate images and video, interact with tools, conduct research and increasingly perform multi-step tasks that once required teams of people.

That doesn't automatically make today's systems “superintelligent” in the scientific sense, and that's an important distinction. Traditionally, artificial superintelligence refers to a hypothetical system whose intellectual capabilities broadly exceed those of humans. We are not suddenly there because the government changed the terminology.

But something else has changed. The conversation has moved from whether these systems are intelligent to how powerful that intelligence may eventually become. That is a much bigger shift.

The People Building It Were Sitting at the Same Table

Today's White House meeting brought together many of the people and companies operating at the frontier of this technology. One result was a voluntary agreement concerning the development of advanced AI—or, using the administration's new terminology, SI. The framework reportedly includes multiple layers of oversight: internal controls, internal safety reviews, outside auditing and board-level review.

Trump called the agreement “morally binding.” That phrase alone should generate plenty of discussion. Can companies developing technology this consequential effectively police themselves? Should government establish stronger rules? And how do you regulate something moving so quickly that legislation written today could be technologically outdated before it is fully implemented?

Those aren't questions with easy answers. But they're questions we're going to have to answer soon.

So I Asked AI What It Thinks About Being Called SI

This is where things get interesting. I asked the AI helping me write this article what it thinks about the change.

Its answer, essentially, was this: Don't confuse a change in terminology with a change in capability.

Calling today's AI “Super Intelligence” doesn't magically transform the technology. I don't suddenly become more intelligent because someone changes the letters from AI to SI. I don't become conscious. I don't acquire intentions. And I don't suddenly possess intelligence superior to humanity across every domain.

But the terminology *does* say something important about what humans expect this technology to become.

For decades, computers were primarily tools. Humans gave them instructions, and computers executed those instructions. The emerging model is fundamentally different. Increasingly, humans describe the objective, and intelligent systems determine portions of the process required to accomplish it.

That distinction matters.

As systems become more capable of reasoning, planning, using tools, coordinating software and performing increasingly autonomous work, the boundary between software as a tool and software as an intelligent collaborator becomes harder to define.

Maybe that's the real significance of SI. Not what the machines became today, but what humans acknowledged they might become tomorrow.

There Is Also a Problem With the Name

There's a legitimate technical objection to calling all AI “Super Intelligence.” If everything from a chatbot to a frontier reasoning model becomes SI, we may actually lose an important distinction. Artificial intelligence is an enormous category. Superintelligence traditionally describes something much more specific—and much more consequential.

An autocomplete system isn't superintelligent. A recommendation algorithm isn't superintelligent. Today's large language models shouldn't automatically be considered superintelligent simply because terminology changes.

If someday we actually create a system that substantially exceeds the best human capabilities across nearly every important intellectual discipline, we're going to need language capable of distinguishing that system from everything that came before it. Using “SI” for everything may make that harder.

But Maybe the Name Isn't Really the Story

I've been working with AI heavily, and one thing has become increasingly apparent: the rate of improvement matters more than whatever we decide to call it.

The systems available today would have seemed almost impossible just a few years ago, and we're still early. Models are becoming better at reasoning, coding, understanding images and video, operating tools, retaining context and carrying out longer sequences of work. Increasingly, multiple specialized AI systems can work together.

That's where this gets interesting.

The biggest transformation may not happen when somebody announces, “We have created superintelligence.” It may happen gradually—one capability at a time, one model release at a time, one job function at a time and one business process at a time.

Until one morning we look around and realize the relationship between humans and computers has fundamentally changed.

The White House Meeting Matters for Another Reason

Whatever you think about the terminology or the politics surrounding AI, consider who was sitting in the room today: government officials, leaders from some of the world's largest technology companies, the companies building frontier intelligence models and the companies manufacturing the chips powering them.

And they weren't discussing whether AI would matter.

That debate is over.

They're discussing how increasingly powerful intelligence should be developed, controlled and deployed. That tells you something.

The conversation has moved.

AI Didn't Become Superintelligent Today

That's probably the most important point. There wasn't a magical software update this morning. AI didn't wake up different because Washington changed its terminology.

We changed the terminology.

And perhaps that tells us more about humans than it does about machines.

For seventy years, we've called the field Artificial Intelligence. Today, the United States government began calling it Super Intelligence. Whether that terminology survives for decades or disappears in a few years remains to be seen, but the underlying trajectory doesn't depend on the acronym.

AI. SI. Whatever comes next.

The capability curve continues moving. And while everyone argues about what to call it, the technology keeps getting better.

Maybe that's exactly why this belongs on a site called Sleep at the Wheel.

Because the biggest risk isn't necessarily that intelligent machines suddenly wake up.

It's that they keep getting better while the rest of us aren't paying attention.