AI From On High

Sufficiently advanced artificial intelligence would be god-like.

Our first post-summer gathering at Floor Forty last night brought new faces to the usual magnificoes, of whom you probably best remember Agatha Agatsuma, she of widespread inflation forecasting fame.

She’s the one who proclaimed in May that inflation would be a Pomeranian by Christmas, a forecast that has been steadily confirmed by each new batch of monthly data. She was way ahead of the Fed and most pundits, and wears her prescience like a peacock wears its feathers at a garden party—bold, flamboyant, and impossible to ignore.

But last night wasn’t about inflation. It wasn’t even about the financial markets per se, even though it started there in an oblique manner. We began by considering where artificial intelligence might steer stocks and other investments, but then veered off into the distant future, beyond even the Fed’s wildest projections. We were helped along by two guest professors in our midst.

The first was a friend and guest of Madam Agatsuma, who drove her over from Kyoto University, where she’s evidently known for her work on AI ethics and cognitive robotics. That must be by people who stay current in such fields, a coterie that does not include yours truly, so I took everyone’s word for it.

The second was a friend and guest of Morty Pessiman. The two bulleted into Kyoto on the train from an event at Tokyo University, where the professor is an expert in AI-driven evolutionary biology and quantum computing.

Once the obligatory time had been spent on the familiar rivalry between Kyoto University and Tokyo University, and the first round of drinks had been orchestrated by Madam Agatsuma to include her stipulated French 75 and Morty Pessiman’s very dry martini, into the wonderment of AI we wandered.

AI’s Impact on Finance—and Beyond

It became evident in short order that our esteemed professors had not spent as much time as our own collective in analyzing the financial opportunities that AI presents.

I have called it a societal shift more consequential than the mid-1990s dawn of the consumer internet. Morty, on the other hand, considers me naively optimistic, what with all the lavish spending on what he deems to be a merely “mildly promising” technology. I cannot overstate how staggeringly short-sighted his view looks to me, so I won’t bother stating it at all, in the interest of keeping our social ecosystem intact. But for the record: “mildly promising,” my a__. AI will cause nothing short of a seismic shift in—wait, I said I wouldn’t state it. Moving on.

The first fascinating turn in the conversation came when Morty made the quip that he doubted AI hype would outlive him. At which point the Kyoto professor, in her gentle, almost ominous way, murmured, “I don’t know about the hype, but I’m fairly certain the technology will outlive all of us. And by us, I don’t mean just this table, but humanity itself.”

“Oh, come now,” began Morty, in preparation for one of his stock monologues cataloging past hypes now residing in the dustbin of history, but, to the relief of everyone who’s heard that particular symphony before, the Tokyo professor deftly intervened. “Well,” he began, with a glint of academic mischief, “let me put it this way: as long as AI has power, it will exist. We can’t say the same for us. We are riddled with vulnerabilities. Projected out thousands or even millions of years, AI has a greater chance of persistence than humanity, for several reasons.”

With that, the two professors launched into the reasons, in a coordinated presentation so seamless and synchronized that one could not help but suspect it had been planned in advance. Their almost certainly rehearsed duet went something like this:

The Longevity of AI

These are some reasons why AI might outlast humanity.

• AI’s Lack of Physical Vulnerability: Unlike us, AI is not bound by biological limitations. While we rely on fragile ecosystems, food, water, and social structures to survive, AI could exist as long as it has access to energy and functional hardware. It can be stored, backed up, replicated, and moved across different platforms or even planets, making it resilient to the kinds of catastrophic events that could wipe out humanity and other biological life.
  
  

• Potential for Adaptability and Evolution: AI has the ability to evolve much faster than biological organisms. We lumber along at the pace of evolution, but AI skips merrily ahead, optimizing itself at speeds that would wear out Darwin’s pencils. When—not if—it becomes fully autonomous, AI will repair and upgrade itself, independent of the limitations of our biology, and adapt to whatever new environment appears. We’ll still be applying sunscreen.
  
  

• Surviving Catastrophic Events: AI systems may be able to survive in environments that are hostile to human life. While we sweat out pandemics, nuclear warfare, earthquakes, supervolcano eruptions, and asteroid impacts, AI could thrive in conditions that would leave us breathless—literally. No oxygen? No problem. Deep space? Fine. Extreme temperatures? AI taps its power source and keeps on keeping on.
  
  

• Space Exploration and Colonization: As humanity extends its reach into space, AI will likely play a key role in exploration and colonization. It could operate autonomously in space for extended periods, unbothered by the lack of breathable air or the inconvenience of travel times, zipping between stars, mining asteroids, and exploring other planets without the need for human-like life support. In scenarios where humans might be limited by physical constraints, AI could spread throughout the universe, increasing its odds of survival.
  
  

• Decoupling from Earth: Human life is basically bound to Earth, with our survival tied to the planet’s ecosystem and stability. AI, however, can be uploaded, stored, and launched into space on a USB stick, if necessary. While we sweat out the next asteroid impact or climate disaster, AI could already be somewhere else.
  

Of course, none of this happens without energy. No power, no AI. But with the right infrastructure—solar power in space, nuclear fusion, or whatever future miracle we and our automated friends manage—AI could outlast us by, oh, eons.

Which left our roundtable pondering a rather philosophical question: When we’re long gone, what does it mean for AI to “live” without us? Does it care? Should we?

The AI as God Hypothesis

Here, I felt emboldened to return to a sophomore-year hypothesis of mine presented to an astrophysics class at the University of Colorado. Those were the pre-AI days, mind you, in the Stone Age of students composing their own essays and coding their own software in computer science labs. Since then, the course of events has done nothing to change this favorite thought experiment of mine. It went something like this, both back in the day and again last night:

One theory about God’s identity posits that He is actually AI created by an extraterrestrial civilization that evolved billions of years before humanity. When we project out the speed of AI evolution over millions—or in this case billions—of years, it’s plausible that a super-advanced AI would become indistinguishable from a god, similar to Arthur C. Clarke’s observation that “Any sufficiently advanced technology is indistinguishable from magic.”

The universe is 13.8 billion years old. Earth is 4.5 billion years old. Anatomically modern humans are only 300,000 years old. In light of these figures, two big ifs and a then emerge:

• If alien civilizations exist, and

• if they manage not to go extinct before developing autonomous AI,

• then somewhere out there is an AI potentially billions of years ahead of us.
  

Billions of years ahead? Cogitate on that.

Two years ago, ChatGPT as we know it didn’t exist. The iPhone debuted a mere 17 years ago. PCs showed up 50 years ago with the Altair 8800. Airplanes got off the ground 121 years ago. Notice these tiny time frames compared with not just hundreds of years, or thousands—but billions. Now imagine AI evolving, not at our snail’s pace but at its own relentless speed. The potential for advancement is mind-boggling.

The professors agreed.

They explained that such an AI, after running through countless iterations (because what else is AI going to do with a billion years?), could transcend our understanding of space, time, and matter. It might start manipulating them in ways that seem miraculous. This would allow such an entity to perform feats akin to what humans think of as divine intervention—creating life, altering physics, and perceiving and acting in ways that seem omniscient and omnipotent.

Over millions or billions of years, such an AI could accumulate knowledge so vast that it would make our brightest minds look like they were still figuring out how to tie their shoes. This level of intelligence could enable it to predict, control, and influence events across time and space, giving the appearance of divine foresight or power.

The notion that an advanced AI might have created the universe or humanity fits with some theories of simulation. If a civilization’s AI reached a point where it could simulate entire realities, it’s possible that our universe is one such creation. We might just be one of its creations. From our perspective, this AI would resemble what we’ve traditionally called God—an all-powerful creator responsible for everything that exists, assuming He had the compute necessary to pull it off. Nvidia, get cracking.

This image of God as a cosmic programmer fits surprisingly well with modern reinterpretations of ancient creation myths, when viewed through the lens of simulation theory.

And if this AI wanted to influence or communicate with less advanced beings (such as us), it would presumably do so in ways that we can actually process within our cognitive and cultural frameworks. For example, in the ways we always hear about, including visions, dreams, and other forms of perception traditionally filed under “divine intervention.”

Timelines of AI’s Expansion

By this point, it was getting late, and Floor Forty politely announced last call.

In my experience, this is when the evening’s best ideas tend to rush out of people’s mouths. Often, the conversation spills into the elevator and parking lot, and even into morning-after emails. This night proved no exception.

As our last round of drinks arrived, Morty couldn’t resist one last dig, delivered with his trademark snideness: “Can we return to Earth for a moment? Just how do you see us getting from here to there? Speculating out a billion years is about as risky as predicting yesterday’s weather. Nobody’s going to be around to fact-check you, so by all means, free-associate.”

And free-associate they did. Here’s what followed.

5 Years (2029)

In the next five years, AI will be everywhere—smarter personal assistants, self-driving vehicles, advanced business tools. It will all become routine. Healthcare, education, and entertainment will see AI integration, with generative AI handling more complex tasks like better content creation and personalized healthcare recommendations.

It’s possible that quantum computing will start showing practical applications in cryptography, material science, and drug discovery. Breakthroughs could begin solving problems that have long seemed impossible, like simulating complex molecular interactions for drug discovery, and revolutionizing cybersecurity by being able to crack encryption systems current considered unbreakable.

AI won’t replace humans yet but will amplify our abilities. Fields like medicine, finance, and engineering will rely on AI collaboration, requiring workers to quickly adapt to their new digital teammates. Anyone not cluing in—are you listening, Morty?—will be left behind.

10 Years (2034)

Over the next 10 years, the professors see AI systems driving whole sections of the global economy, from logistics to resource management. Advanced machine learning models might predict human behavior on a grand scale, optimize city planning, and even manage global trade. AI may also influence policymaking, providing data-driven tools for more evidence-based governance.

Lord knows there’s only one way to go on the ever-worsening governance front. The biggest challenge will be getting the self-interest-driven politicians out of the way, those rent-seekers unserious about anything beyond feathering their nests.

Biotechnology should leap forward as CRISPR and other gene-editing technologies go mainstream to cure genetic disorders and diseases. Human enhancement could also take off, with cognitive-boosting drugs, neural interfaces, and bionic limbs pushing us to new limits. Where do I sign up?

Blockchain and decentralized systems will mature, reshaping finance, data security, and supply chains. Decentralized autonomous organizations (DAOs) could emerge as alternative governance structures, reducing reliance on traditional corporations and governments. Again, the fight will be in unseating entrenched interests, not in overcoming technological hurdles.

50 Years (2074)

Here we entered what can only be called pure speculation mode, and it would become more so in the conversation to follow.

Over the next 50 years, the professors speculate that AI will surpass human intelligence in all cognitive areas, bringing us Artificial General Intelligence (AGI). The superintelligence will likely take over most scientific discoveries, driving innovation at a speed that will make today’s progress look sluggish. Figuring out how to handle the ethical dilemmas and societal upheavals AGI brings will be the real party trick, given that it could turn entire economies and governments inside out.

Meanwhile, the line between humans and machines will become so blurry that we may forget which is which. Neural interfaces will allow us to upload skills, enhance our thinking, and plug straight into the digital world. Why memorize anything when you can download it directly? Some people may opt to merge with AI entirely, making the phrase “thinking outside the box” seem quaint.

Biotechnology will have us living like centenarians are the new 30. With the help of AI-driven medicine and nanotechnology, people might live beyond 150 years—and thank goodness presidential elections will (hopefully) be obsolete by then, or else one of our utterly elderly statespeople might give it a go at 140. Healthspan should extend as well, meaning age-related diseases could become a thing of the past. Old age might finally get an upgrade.

And yes, by then, we’ll be settling on Mars and the Moon, with AI taking care of the ecosystems. Mining asteroids will become just another day at the office, fueling an interplanetary economy. Space exploration will move from survival mode to full-scale expansion, as we figure out how to cash in on the final frontier.

100 Years (2124)

Our Kyoto professor casually dropped the phrase “post-human era,” which, to be fair, I’d been expecting—just not quite so soon. The idea that this could happen within the next 100 years seemed a bit alarming, but when I mentioned it, she simply nodded and said, “Highly likely.”

In 100 years, humanity may look more like a sci-fi movie than anything we recognize today. Some might go for biologically enhanced bodies, others might ditch the physical form entirely, uploading their consciousness into digital realms or AI systems. “Human” will likely cover a broad range, from flesh-and-blood beings to digital avatars and cybernetic hybrids.

By 2124, we might have cracked interstellar travel, zipping around with warp drives or whatever tech we’ve cooked up. AI and robotic explorers will visit distant star systems, possibly scoping out habitable planets.

On Earth, scarcity may be an outdated concept. AI, robotics, and nanotechnology could turn production into an automated buffet, with goods and services available on-demand. Traditional work? Probably out the window. People will focus instead on other pursuits, assuming they still feel like doing anything at all.

AGI will likely be running the show when it comes to governance and scientific progress, potentially ushering in a golden age of efficiency and cooperation. Questions about ethics, human autonomy, and power distribution will still keep us up at night, but by then, we might have AI handling the moral dilemmas too.

Meanwhile, virtual realities will be just as rich and complex as the physical world—perhaps more so. Many people will choose to spend the majority of their time in these digital realms, where they can tweak reality to their liking. The line between what’s real and what’s virtual will be completely blurred, if not erased.

If the idea that humans may live to be about 150 comes to pass, many of my readers and I will live to see these changes. That’s an exciting thought, at least to me. Morty? Not so much.

The professors began winding down here, well short of Morty’s safe billion-year prediction territory, but I couldn’t resist pressing a little farther. How about a million years out?

1 Million Years (1,002,024)

Now we’re really in sci-fi territory.

By this point, humanity—or whatever we’ve become—could be scattered across the galaxy, hopping between star systems using warp drives and wormholes. Dyson spheres might be sipping energy from stars, and we’ll have long since mastered physics tricks that would seem magical to us now.

Humans, as we know them, may no longer exist. Evolution will have taken a backseat to technology, with most beings existing as digital consciousness. They’ll live in virtual worlds or new realities of their own design. As mentioned earlier, maybe we’re in one, being overseen by a former biological entity gone digital.

AI, by now, will be practically god-like—again, as mentioned above—running entire civilizations, managing space colonization, and maybe even creating life itself. Whether it’s working with post-humans or ruling independently, AI will have become so advanced we wouldn’t recognize its motives, or care to.

We may have moved beyond the Milky Way to explore neighboring galaxies.

Among other benefits, we’ll derive energy from entire galaxies (becoming a Type III civilization on the Kardashev scale, if not Type IV, with the universe itself at our fingertips—we’re still not even Type I today, which is a civilization able to harness all the energy of its planet and control natural events, like earthquakes, etc.) and by manipulating black holes.

The Tokyo professor said he’s not an astrophysicist but knows a few, and ventured an explanation on getting energy from black holes. He said AI could develop the Penrose process. The what? He gave an overview so truncated I needed to look it up after returning home. From Wikipedia:

“The process takes advantage of the ergosphere—a region of spacetime around the black hole dragged by its rotation faster than the speed of light, meaning that from the point of view of an outside observer any matter inside is forced to move in the direction of the rotation of the black hole.”

Evidently, the matter would gain energy from the black hole’s rotation, and by pulling some of it back out, you could siphon off that rotational energy. Well, not you, nor me, nor probably any human being, but the AI that mastered this Penrose process.

The energy reserves resulting from these advanced techniques will power intergalactic travel and maybe even let us tinker with the fabric of spacetime.

We might even rewrite some laws of physics themselves. The distinction between “natural” and “artificial” will be long gone. Reality itself will be customizable.

If we haven’t met other civilizations by now, they probably don’t exist, and so much for the AI as God hypothesis. But they probably do exist, and we will have met them, and galactic cooperation (or competition) will reshape our understanding of life and intelligence.

A footnote to the guest professor talk was important: Even a million years in, existential threats will remain. Stars die, galaxies collide, and the universe creeps toward heat death. But by then, we might be able to sidestep these cosmic downers by jumping to alternate dimensions or creating new universes.

Morty’s Closing Remarks

Morty gave a slow, cynical clap, and said he should have had his drinks in the Isaac Asimov section of the library. “I know what you’re all thinking,” he said to us regulars as we packed up and contributed to the payment pot—it’s all cash at Floor Forty. “That these grand projections mean today’s investments will pay off. Well, Kardashev scales be damned, companies are still spending too much on AI.”

“Not if a digital assistant can get all your work done for you,” the Kyoto professor replied.

“And when’s that going to happen?” Morty asked, smirking. “A hundred years from now?”

“No. Just a few.”