All right. So, I’m in this place where I feel like the singularity is going to be boring. And this is not a unique or novel thought because about a year ago, Sam Altman wrote a blog post called “The Gentle Singularity.” So, what that indicates to me is that he honestly, like, originally thought that we were going to be in for, like, a fast takeoff or a hard takeoff scenario. And the fact that he felt the need to write on his blog—which he only writes a couple blog posts a year—that like, “Oh, actually things are going to be slow and it’s going to be gentle and it’s going to be a little bit on the boring side.”

And here’s, like, here’s what happened and here’s what inspired me to make this video in particular right now: is that particularly over the last month, uh, with Claude Code and CodeEx and agentic coding basically getting past the point where it can code autonomously for 24 hours—for, in one case, allegedly a week. Um, and then, uh, Claude Co-work, if you haven’t heard, is basically the next edition of Claude Desktop, was coded up in 10 days, and 100% of the code was written with Claude.

So, what really seems to have happened is people are realizing like, “Oh, wait, is this—is this AGI?” And they’re like, “Well, is that even the right term for it?” Because we’re in this regime of what Ethan Mollick calls the “jagged frontier,” meaning that artificial intelligence on some dimensions surges ahead of human capability by a long shot. And then in other cases, it’s just not as good, namely on contextual memory. Uh, and of course with recursive language models, the contextual memory might have been solved. And so everyone’s like, “Wait, maybe—maybe the AGI term was the wrong term all along.” And what we’re really seeing is like the first thing where it’s like, “Okay, this is now doing something that—that no human team could possibly do.” Shipping an entire, like, you know, I don’t know if it’s enterprise-grade, but an entire shippable app in 10 days. You can’t really do that. I mean, yeah, you can—you can—you can have a hacking session, you know, and—and—and launch a proof of concept, but not something that is ready to launch to the masses.

And so when we—when it’s like, “Okay, cool, that’s all great,” and that really changed people’s minds and it—it—it advanced the Overton window, but then like, did the world change? No. You know, it’s like, and—and in the grand scheme of things, most people don’t care. It’s like, “Great, you know, the nerds have more software. Yay. Woo software devs. You’re all divas,” and that sort of thing. Like, I don’t know if you realize this, but software engineers do not have the best reputation in the rest of the world because they all think that they’re Tony Stark. And you know, when they say, “We’re going to solve the world’s problems,” it’s like, “Okay, what about energy? What about coal? What about geopolitics?” And it’s just like, their definition of everything is actually much narrower than actually everything.

And so, you know, for people that are skeptical, it’s like, “Okay, yeah, we have unlimited code coming.” I call it the “fire hose of code.” We’re going to have—we’re going to get to a point where we’re going to reach closer to a saturation point because as—as other people pointed out, I think it was actually Elon Musk, he basically said like, “The optimal amount of code that humanity needs is, you know, a thousand times more than we have today just because software is so useful.” Um, and so since software is so useful, we need more of it. So therefore, we’ve invented something that the best thing it—it can do right now is write more software. Uh, which that’s fine. You know, there’s nothing wrong with that.

The Evolution of Computation

And taking a big step back, um, in hindsight, it’s going to be very obvious that the reason that we’ve invented computers was to run AI and that everything—every computational technology we’ve done for the last, you know, century or so has basically just been the leadup to running AI. Because if you look at every FLOP that we run, every—every calculation that we run, the lion’s share of those calculations are going to be artificial intelligence. So what I mean is like, you know, fast forward 20 years. So the year is 2046 and you look at, “Okay, how many CPU cycles or GPU cycles or processor cycles are spent on human tasks?” And it’s going to be like less than 1%. And 99% is going to be on AI, like AI running stuff.

And so in—in—in the grand scheme of things, we invented computers and we invented software to generate AI. And there’s going to be this really weird thing where it’s like, “Wait, you mean you used to write computer code by hand and you did that for a few decades?” And then for the rest of human history, the rest of human existence… I mean, yeah, there’s going to be—there’s always going to be nerds who want to understand the code and the math and—and—and doing it on their own. But also, as more time goes by, people use more and more abstract, uh, programming languages.

And so what I mean by abstract, that’s not like, you know, the difference between Schopenhauer and something else. It’s not philosophically abstract. What I mean is abstracted away from machine language. You know, you don’t write in binary. You don’t write in assembly anymore. Some people still do, but most people write in very high-level abstracted languages like Python, which is not even a compiled language; it’s an interpreted language. Um, and so that—that differs from languages like C and C++, which is compiled, meaning that you actually need to have something write it into a binary, an executive—an executable binary, uh, which you know, when you look at the trend, it’s like, we—we—we tend to focus on that abstraction ascension. Then the most abstract language to write in is English, which means that in the future, all software is going to be written in, you know, English or natural language—not ne—not necessarily English. You can write in any language, but you get the idea. Writing in natural language and specifying—specifying things in natural language and then just let the interpreter, you know, figure it out.

I remember my brother-in-law when he first got his hands on ChatGPT, he was trying to think about how do I incorporate this into my stack as a software engineer. And you know, I said, “Well, you know, English is an interpreted language.” So, he’s like, “Oh, it’s like an SDK.” I’m like, “Yeah, you can think of a language model as an SDK. It’s just the functions you have to write with natural language.” And he, you know, he was happy with that explanation. Um, so, you know, there are probably plenty of people who might disagree with that analogy, but it was a good enough analogy for the time.

Projecting the Future

So, you know, from a science fiction perspective—and—and when I say science fiction, I don’t mean like, you know, “This is all fictional. None of it’s happening.” What I mean is like, project into the future, right? If you say, “How does—how does this movie play out?” If you were going to take the world as it is today and imagine 20 or 40 years in the future, which is exactly how science fiction like, um, you know, Ghost in the Shell was written, is like, “Okay, where’s technology going in 40 years?” Um, and so then it’s like, “Okay, well very obviously network is going to be everywhere.” So we have internet saturating everything, and then AI is going to be everywhere and it’s going to be writing code and it’s going to be instantiating everything on the fly.

And you might say, “Okay, well Dave, that all sounds good, but where’s the singularity stuff?” And that’s kind of the point: is that until we get the big energy physics, then your daily life isn’t going to change that much. And you know, because like, “Great, you have more software,” which means that your—your smartphone, you know, will be able to fire up an app on the fly and your desktop will be able to fire up an app on the fly. And of course, you know, you’re going to have—there’s going to be compounding returns. I’m not going to say nothing’s going to happen, uh, because some of those compounding returns means, “Oh hey, you know, you got—you got your DNA sequenced from a, you know, professional source,” and then your home desktop is able to help you model your entire genome and you do it with a natural language interface like they do in Star Trek.

So you’re like, you know, “Computer, analyze my—my genetic code for, you know, yada yada yada, this—this food sensitivity and optimize my health.” And you’ll be able to do all that at home, which is—which is, you know, that’s—that’s how you get to personalized medicine. Because it’s not that you go to a doctor’s office where there’s a human doctor who spends a week monkeying with your genes and then, you know, orders up some custom medications. It’s you analyze your own genetics at home and then you send off an order and you get the peptides that you need to, you know, cure your own disease or live longer or whatever. That’s personalized medicine. And artificial intelligence will help that.

And so then, you know, you send off your genome and the work that your home AI did to, you know, this lab, and their AI will say—it’ll check, it’ll double-check your work and say, “Okay, cool. Yeah, I see what you’re doing. Let me send—let me put together, you know, the—the actual proteins that you need, the actual peptides that you need, and mail it back.” Something like that. I don’t know if that’s actually literally going to happen, but that’s—that’s the kind of thing that cognitive hyperabundance could possibly enable in this future. Um, so there—that’s—that’s like one thing. And yes, that’s very exciting, um, particularly if you have, you know, chronic illness or if you’re getting old and you want, you know, rejuvenation, and so then everyone’s going to be a bio—biohacker at home.

Robots and Energy

So on the one hand, you know, the singularity is mostly about software and intelligence and artificial intelligence. And of course, that will have spillover effects into robotics, and it’ll be—it’ll be a different story when people have robots building their homes and fixing their homes and working in factories. But you’re obviously—you’re honestly not going to see most robots. And what I mean by that is most humanoid robots are going to be in mines. They’re going to be in—in, you know, on farms and not like, you know, cutting things by hand, but driving the tractors for the farmers. Um, they’re going to be in—in factories. They’re going to be in warehouses. They’re going to be in logistics.

The vast majority of robots are probably going to be invisible behind the scenes. So, again, the singularity is going to be kind of boring because it’s just like, “Oh, you know, who—who works in that Amazon warehouse nearby? It’s 100% staffed by robots. There’s no humans there.” There’s—I mean, there’s maybe a few humans there to, y—you know, keep the lights on, but then eventually those humans are going to be replaced by robots. Because if you have a—if you have a human that can like, you know, go help another robot, then why not just have a robot that can go help the other robot or make sure the power stays on and that sort of thing?

You know, and that again… so like that’s—that’s very interesting and it—it’s kind of dystopian, but it’s also kind of boring because then you don’t get to see it. You know, I mean, yeah, Amazon might do tours like, “Oh, look at our, you know, we [have] 10 million, you know, humanoid robots scattered across the country and they operate 24/7.” And so now, you know, your Amazon, uh, fulfillment timeline, you get everything within an hour. And it’s like, “Okay, cool.” You know, it was—with humans in the loop, it was 2 to 4 hours is the fastest you could get an Amazon fulfillment. And with 100% robots, it’s an hour or two. Okay, great. You know, what’s—what’s the big deal? What’s the big difference?

And then you might say, “Oh, we have infinitely more software.” Okay, great. So the—the—the thing that’s really going to move the needle for most people, I think, is the energy front. And again, uh, in hindsight, it’ll be very obvious that all the energy that we’ve figured out how to generate is going to be powering the AI. And what I mean by that is that, you know, solar, nuclear fission, nuclear fusion, other renewables, whatever, doesn’t matter. Um, artificial intelligence, at least on the current trajectory, is going to be the largest single, uh, consumption source of electricity. And so then it’s like, “Oh well, we need nuclear fusion and we need solar everywhere just to power the AI that we want to bootstrap the automation layer of our entire civilization.”

And so in hindsight—hindsight being 20/20—in, you know, 20 or 40 years, we’re—we’re going to say like, “Okay, how much, you know, if you break it down into like domestic use, industrial use, and AI…” If you make a pie chart of human energy consumption, it’s going to be like, you know, 30% to 50% is industrial, the vast majority is going towards AI workloads, and then like 1% to 5% is going to be residential use. You know, it might not be that extreme. It might be more extreme than that. Like I’ve—I’ve—I’ve had conversations with people where—where they think that it could be like that hum—that human use, like you know, heating and cooling your home is like less than 1% of the—of the future energy grid demand and that the vast majority is going to be AI workloads.

Now, of course, that assumes that we don’t make the chips more efficient. I do think that it’s possible that thermodynamic computing and photonic computing and all sorts of other kinds of computing paradigms make it more efficient. Because here’s the thing is: when something starts consuming that much—that—that many resources, that is then a signal to, you know, the investors and the startup funds to say, “Hey, let’s find a more efficient way of doing this.” So, but even then it’s like, “Okay, well, what does that really change for you on—on a short timeline?” It doesn’t change much. It just means that your power bill is going to be competing with a data center. Um, and it means that, you know, uh, it mean—it means that, you know, it might go up or it might go down. I—I honestly ultimately think that power will be—at least—at least the level that you consume at your home will be trivial. It’ll be like $10 a month.

And of course, Jevons’ paradox kicks in, which that means that, “Oh hey, if electricity is that much cheaper, I’ll just f—I’ll just do more things with it. I’ll grow more food at home,” because then having a—um, having a, you know, a hothouse at my home… Um, and hothouse is not like, you know, hotboxing. What I mean is—is a heated greenhouse. So, you can do a heated greenhouse that’s running 24/7 at home, mean—meaning that you grow your own vegetables and your own microgreens, and guess who runs it? Your home robot does.

Thermodynamics and Space

So there’s going to be lots and lots of little incremental changes, and then the big changes come, however, once we have like Dyson swarms. And this is where I think that it—we will inevitably get here, and—and it is simply because when you look at it from a purely thermodynamic standpoint, when you look at it as—strictly through the lens of entropy, we are going to want to be using more energy than the earth can handle. And so you look at like, what—what’s the maximum amount of energy that the earth’s surface can dissipate? And it’s finite. And you don’t want to cook yourself, right? So what do you do? You offload that entropy, you offload that thermal load into space.

And you know, so you can—you can either beam the heat from the surface to space, or you can just move the workloads to space. So this is where I made that video a little while back where I—I called it like, “reverse Trantor.” So instead of having 5,000 layers of, you know, a built-up ecumenopolis, then what you do is instead you just take that surface area and put it outside in the solar system. Why? Because then you have uninterrupted sunlight. Uh, and so then you—you have—the satellites are—are, you know, receiving some of the heat and they’re doing some of the processing and the workloads for—for artificial intelligence. But then they can also be growing food, right? You know, they can be growing food, they can be recycling stuff because if they have unlimited access to—to, uh, thermal energy from the sun or—or photonic energy from the sun and unlimited ability to eject excess into deep space without, you know, humans being in the way or an atmosphere being in the way.

But again, you don’t see that. You know, it’ll be like, “Okay, Elon Musk and SpaceX and whoever else is going to be launching hundreds, if not thousands, of—of—of ships a day,” but then eventually there’s going to be enough people and robots and infrastructure in space that you don’t even need to keep launching them. That you have all the foundries in space, you have all the robots already in space building more. And so, like the von Neumann probe idea, we’re going to use that to colonize our own solar system first. But again, you don’t see that. You’re not going to see that. You’re going to see, you know, a few—a few Starship launches and then you’re going to, you know, if you have a good enough telescope, you’ll be able to see some solar panels on the moon. Um, but most of those satellites and stuff, they’re going to be in very high Earth orbit or maybe even Lagrange points, meaning that they’re going to be like, you know, a sixth of the way around our orbit ahead of us, or way further out where the James Webb Space Telescope is.

So you’re—you’re just not going to see a lot of the infrastructure, and a lot of the intelligence is going to also fade into infrastructure. And—and of course like, yes, there will be visible things, you know, like more holograms and more, you know, humanoid robots wandering around the streets and driverless cars and that sort of thing. Um, and you know, but some of the stuff is going to play out over such a long time horizon that you’re not going to notice it.

Longevity and Normalcy Bias

And the—the biggest thing is going to be longevity and rejuvenation. Now obviously, if you go visit your grandpa in, you know, 10 years and suddenly he looks younger than you do, it’s like, “Okay, grandpa, where’d you get the peptides?” You know, and it’s like, “Oh, well, you know, my AI cooked up this thing and I feel better than I have in 40 years.” Uh, which I think—I think 10 years is pretty reasonable for that kind of thing to happen. Certainly within 20 years, we’re going to have the intelligence long before, you know, the—the full event horizon of the singularity.

Um, and you know, but your—when—when you go from old to young or from sick to healthy, your brain actually has evolved to update its—its self-model very quickly. So, here’s—here’s what I mean by that is like, “Okay, imagine that you’re a caveman and you break your arm,” and so you feel like crap for, you know, a couple weeks as your arm is healing, but as soon as your arm has healed, evolution wants you to get back to normal behavior. So you basically forget how miserable you were. You—your—your—your mental model updates and you’re like, “Oh, well, I can use this arm again, so back to the hunt.” You go back to your normal behavior very, very quickly.

And so your normalcy bias or your self-model updates so quickly because it’s not—it’s no longer adaptive to have sickness behavior or injury behavior because then you starve or you don’t—you don’t contribute to the clan anymore. Um, and so we have all these neural machineries—or maybe that’s the wrong… like neural characteristics: normalcy bias and—and, uh, quick regression from sickness behavior so that everything that is around you feels normal in the moment. As long as it—as long as things take more than like 3 months to change, you know, let alone 3 years to change, it just feels normal the—every single step of the way.

Because it’s like, “Oh, what happened in—in, you know, late December 2025 or early January 2026 was—oh, now we live in a world where Claude Code can operate autonomously for a week.” Great. Okay, cool. That then becomes normal. And it becomes normal faster and faster. So it’s like, on the one hand, I do criticize people for saying like, “You don’t have any neural machinery to intuitively understand exponentials,” but you do have neural machinery that allows you to basically say, “This is the new normal,” and then you get ready for the next change.

But you know, you do those incremental changes over the span of… you know, because it’s—it—this isn’t slowing down, right? It’s not—it’s not like, you know, 2026 is the last year. We’re going to achieve more during 2026 than we achieved in 2025. We’re going to achieve more in 2027 than we did—than we do during 2026. That’s going to continue, but every step of the way, because there’s enough time for every human to kind of update their priors and update their beliefs, it’s going to feel boring every—every—every step of the way. Um, because the moment that you get used to something, you’re like, “Well, don’t take that away from me. I want more of that.” You know, like when—once we got reasoning models, it’s like, “Great. I can’t imagine—I can’t—I can’t remember what it was like to not have reasoning models. How did we get by? The AI was so stupid.” And then of course, you know, the next thing with recursive language models or whatever else comes next, you know, agents, we’re going to say, “How did we get by with just chatbots?” Well, now we have agents.

And so you update your mental model of “how is the world working today” and anything less than that feels boring, and then you also get used to the current, you know, scheme of things very quickly. Now of course, a lot of people—like only a third of Americans—u—use AI on a regular basis and the other two-thirds are just kind of like in the dark. Um, so some of them are going to be in for a rude awakening, but you know, a lot of people are aware of it and it’s… I don’t know, it just… I have felt this—this building sensation of… it’s not even the gentle singularity; it’s the boring singularity is what we’re going to get. What we’re already in.

Because when—when Claude Code and—and CodeEx and all those things started really surpassing what any human team could possibly do, everyone’s like, “So, we’re in the singularity now, right?” And they’re like checking with each other, like, “Are we in the singularity now?” And it’s, “Yeah, okay, so why don’t I—why don’t I feel suddenly like, you know, king of the world or—or like, you know, the birth of a digital god?” And it’s just because that’s not how our brains work and that’s also like… it… like, physics is still the gating factor. And so like, you know, time steps still matter, energy still matters, physical distance still matters, and no amount of intelligence can erase those things.

Physics and the Golden Age

Now, I will say that physics did erase distance in the—at least in terms of communication. Because if you were to tell someone, you know, like from 200 years ago, where the fastest ways to get information around was horse, trains, and boats, it’s like the—the global bitrate of the world was like a few bits per second. And now the global bitrate of the world is like measured in petabits in terms of how fast we can move information from literally one side of the planet to the other. That would have seemed physically impossible 200 years ago. So never say never on physics. We can usually find something to exploit, um, that—that appears like magic. Um, but physics might also be running out. And so what I mean by that is: there’s no new energies, right? Like we—we’ve—we’ve discovered most forms of energy, most particles. So there’s—there’s less surprise coming from physics, which means we might be past the golden age of physics discoveries. Like the—the Oppenheimer and Einstein era might have been the steepest part of the curve and everything has been slowing down since then.

Now that’s not to say that there’s nothing else to—to—to discover. Like we still haven’t fully characterized dark matter and dark energy. Um, but does that mean that those are a unique form of energy? Not necessarily. Does it mean it’s a unique form of—of matter? Not necessarily. It’s just that we don’t fully understand the physics of the universe. But at the same time, it’s been a long time since we’ve discovered a brand new particle. And the—the rate of discovering particles has slowed down dramatically. And it’s been a long time since we’ve discovered a fundamentally new kind of energy. And we might not discover any new—new types of energy.

So again, are we, you know… is it—is it that most science fiction was written… like the golden age of sci-fi was written during the—the—the, uh, like the period of the fastest change in terms of our fundamental understanding of the universe, and we’re still kind of in that expectation, right? Like the—the—when you look at—at stuff like Star Trek where the idea is like, “Oh, we’re just going to keep making discoveries for centuries.” But that might not be the way things work. It might be that—that, you know, there’s a finite number of things to discover in the universe, um, and/or—or finite systems. You know, every system has emergent possibilities, but it might be that we are already living through the fastest rate of change and it’s only going to slow down from here.

That the—and—and I’m not the only person who thinks this—that the progress is going to follow a sigmoid curve. Now that doesn’t mean that other things aren’t following an exponential curve or a super-exponential curve, like computational processing. But that computational processing doesn’t—doesn’t ultim—fundamentally alter physics. It doesn’t alter biology. Even though we can now map all 200 million proteins and we can soon synthesize new proteins, again, you’re still working within the hydrocarbon regime. You’re not fundamentally changing how reality works.

Limits and Diminishing Returns

And so like, I’m kind of mentally bracing for that, like, law of diminishing returns from here on out. Now that’s—again, that’s not to say that… like, yes, I do think that artificial intelligence and computing is going to continue on a super-exponential trajectory, but eventually that trend will give out and we’re going to realize intelligence doesn’t solve literally everything. It solves a lot. But we’re going to—we’re going to realize that the rate of change that we would imagine, you know, going to, like, flying cars and those sorts of things… like, it comes down to joules and watts. You know, it—there’s—there’s always going to be limits in terms of—excuse me—in terms of what you can do with batteries and motors and that sort of thing.

Now, of course, continuous ongoing improvement feels normal. It feels boring. Even though batteries get, what, 8 to 12% more efficient every single year, that compounding return doesn’t feel like magic to us because we’re used to it. We’re used to the idea that like, “Oh yeah, you know, batteries weren’t good enough for electric cars for a long time and now they are. And batteries weren’t good enough for drones for a long time and now they are. And batteries weren’t good enough for, you know, flying VTOL bikes and now they are.” And so like, it’s those slow incremental changes that just don’t feel magical. .

And I think that’s the big difference because like, yes, things will continue to compound and continue to change, but we’re never going to wake up one day and—and have this feeling like the singularity has arrived and we have all transcended. I mean, unless something truly spectacular happens, like we have a universal breakthrough of mind uploading and we prove that there’s a ghost line and that you can actually empirically demonstrate your soul is now living in that machine over there. That could be a really remarkable change. But I don’t know of anything in physics that would allow that to happen. Um, I could be wrong, you know, and I don’t mean to be like, you know, Negative Nancy over here, but, you know, I don’t know how that could happen. .

And I don’t think that science and engineering is anywhere near discovering anything like that, if it even is possible. Same thing with faster-than-light travel. You know, there’s—there’s a handful of theories about how fi—faster-than-light travel might be possible, but the engineering and science is nowhere near ready for that. Like it’s not, you know, like with—with fusion, people, you know, people say fusion is always 30 years away. Faster-than-light travel… we might live in a universe where faster-than-light travel is just not possible. We might live in a universe where you can imagine things that literally cannot happen. .

But having lived through the golden age of science fiction where it’s like, “Wow, you know, like the internet, who knew that that could happen?” And so we—we—we have this cultural belief system. It—it—and it—and it is a cultural operating system where things that we couldn’t imagine suddenly become possible. But I do think that it—it is that we will be entering into a new regime where people realize just because you can imagine something doesn’t mean that it’s physically possible—faster-than-light travel being a—a prime example. .

We can also imagine animals that are impossible. We can imagine space whales: not physically possible. We can imagine Godzilla: not physically possible. So, I think the key thing is like, yes, if you—if you consult your imagination, that’s science fiction, which is fun. I’m a sci-fi author myself. Um, but I think that—I think that the—the unspoken belief, the zeroth principle of like, “There’s just so much more to discover than we will ever discover,” which might be true, but there are also limits. And I think that that’s why we’re in for a boring singularity. .

Um, so yeah, thanks for listening and watching and have a good day. Sorry to rain on your parade.

Join us,
Our team at Living & Loving Inquiry

Mike R Jay & Gary Gile
Founders @ The NEW LeadU

Subscribe @ leadu.ai/news

PS: For clarification –

1. If you just want notice of the LeadU blog posts subscribe @ leadu.blog
2. If you want the blog content by email, our weekly newsletter, and breaking news, subscribe @ leadu.ai/news.
3. Disclaimer

Mike R. Jay is a developmentalist utilizing consulting, coaching, advising and helping… emergent from dynamic inquiry as a means to cue, scaffold, support, lift, and protect; offering inspiration to aspiring leaders who are interested in humaning where being, doing, having, becoming, contributing, relating, guiding to produce resilience and wellth help people lead generative lives.

© Generati