What if you fell asleep for 10 years?

Imagine you’re Rip Van Winkle, Washington Irving’s famous fictional character who takes the longest nap ever.

You close your eyes, fall asleep… and don’t wake up until September 2036.

When you finally open your eyes, what kind of world would you discover?

Are there stunning medical breakthroughs, robots simplifying everyone’s to-do lists, and awe-inspiring technological advancements that make the world a better place? Or are tech companies ruling the world, with truth practically impossible to determine and hordes of hopeless, unemployed people fighting over basic resources?

I’m not trying to be melodramatic for the sake of a good article intro. I have five children and four grandchildren now. So I’m very invested in what the future holds.

My grandkids – none of them older than four – may never remember a time when artificial intelligence (AI) didn’t dominate. Interacting with robots might be as normal for them in 10 years as using smartphones is today.

So all this talk about whether we should stop building data centers and rein in AI models is incredibly consequential. And we need to be properly prepared for what may come.

AI can accomplish amazing things, as evidenced by the Navier-Stokes equations. (Which, don’t worry, I’d never heard of either until now). They detail the movement of fluids and gases in everything from aircraft construction to weather predictions to blood flow.

The problem is, these usually predictable forms can potentially misbehave, causing chaos in the process. And that’s not something Navier-Stokes has ever been able to account for in nearly 90 years.

The globally respected Clay Mathematics Institute even offered a million-dollar cash prize in 2000 for whoever could solve it. But no one could.

Not until OpenAI set around 10,000 AI agents on the job – and succeeded!

Actual mathematicians and scientists still need to “peer” review the results, mind you. However, the Clay Mathematics Institute itself said it’s “apparently been settled.” And I sincerely doubt it issued that praise on a whim.

In which case, what else could AI solve given a bit more time?

It’s a valid question, so I understand why President Trump recently dismissed apocalyptic warnings about AI destroying humanity. He’s deeply concerned about America surrendering its technological and economic leadership to China in this regard.

And that’s understandable, too.

But the other side of the argument isn’t unfounded either. And it’s made quite the waves in the last two weeks, starting with Jacob Coxon.

He resigned from Anthropic (around six weeks into the job) after working at OpenAI for about three years. According to him, both companies’ unstrained AI exploration is “gambling with our lives.”

Even Evan Hubinger, Anthropic’s alignment science lead, believes the chances of AI destroying humankind in the next decade are greater than 10%. Moreover, he admits that the very superintelligence his company is exploring has no proven controls in place.

Truth be told, nobody knows where we’re headed, whether good or bad. But if I had to bet, I’d say AI isn’t taking us toward either paradise or Armageddon.

It seems much more likely that while it continues to become enormously powerful, it will remain imperfect, expensive, and dependent on human input. That’s why I expect a somewhat mixed bag of results down the road…

By 2036, there’s a good chance most knowledge-based workers will operate with an AI agent – if not a team of them – to help them do what they need to do, whether that be:

A single person will therefore be able to run an entire department in some cases. So, yes, some jobs will disappear.

But I think many more will be redesigned. In fact, the kind of common sense, moral judgement, accountability, and relational capabilities humans display might grow more valuable as a result of AI going mainstream.

As such, the real battle for usefulness won’t be between people and machines. It will be between people who know how to unleash AI’s full capabilities and those who don’t.

We could easily see AI’s greatest benefits in the field of healthcare. Doctors are already gaining brilliant new insights into genetics, imaging, and disease detection.

Moreover, new medicines could go through digital testing instead of expensive clinical trials. We could even see individualized treatments rather than medicines designed for statistical averages.

The resulting healthcare plans could become much more manageable at home, in outpatient facilities, or through specialized research centers.

This wouldn’t eliminate hospitals altogether, mind you. But it would take pressure off the current strain on staffing.

Perhaps by a lot.

The first stage of generative AI has primarily occurred onscreen.

The next stage, however, will involve hands, wheels, cameras, and sensors.

Robots are already being utilized in warehouses, e-commerce centers, agriculture, and medicine – though in limited form by a limited amount of businesses. But by 2036, they could easily expand all the way into your home.

As for autonomous vehicles (AVs), expect them to reshape logistics, reduce the need for parking facilities, and add value to city streets that are impossible to park on. Though I do have to add a qualifier here…

AV adoption might take longer than 10 years since the physical world can be so messy. As I acknowledged last month, there are permits to obtain, safety standards to meet, insurance costs to pay, and all the other expenditures those issues come with.

That’s just one of the reasons I’m bullish on Uber (UBER).