Transcription
We are past the event horizon. The takeoff has started. Humanity is close to building digital super intelligence. And, at least so far, it's much less weird than it seems like it should be.
So, I just read you the first few lines of Sam Altman's new blog post titled The Gentle Singularity. And yeah, that was a pretty wild opener. If you couldn't already guess from the title, he's talking about the singularity; the moment AI surpasses human intelligence and just keeps going. He says we've already crossed the point of no return; the point where we might even get transhumans—not like transhumans, but like beyond human humans. You know what I mean?
Anyway, in this video, I'm going to break down the most important and most insane parts of Altman's blog post, including what he thinks is coming in the next few years, how self-improving AI and robot factories fit in, and what it might mean for jobs, scientific discovery, and basically everything else. It's essentially a full breakdown of where we're headed from the guy literally building the thing that's taking us there. Let's get into it.
All right. So, we're not going to read the whole thing word for word. I'll give you the highlights, pull out the best stuff, and maybe drop in a clip or two of Altman himself speaking about this. So, he starts off by pointing out how not a lot has really changed. Even though we now have AI systems that are smarter than most humans, people are still dying from disease, we still can't go to space easily, and there's a lot we just don't know. He mentions though that the least likely part of the work is behind us. The scientific insights that got us to systems like GPT-4 and 03 were hard-won but will take us very far. So again, he's saying we've already crossed that threshold and the rest of the way is basically momentum. At this point, progress isn't a matter of if, it's just when.
He also writes, "AI will contribute to the world in many ways, but the gains to quality of life from AI driving faster scientific progress and increased productivity will be enormous. The future can be vastly better than the present. Scientific progress is the biggest driver of overall progress. It's hugely exciting to think about how much more we could have." So, I completely agree with this. Yes, there are real risks as we move closer to super intelligence, no doubt, and we'll get into some of those. But the potential for AI to supercharge scientific discovery or even generate entirely new branches of it is just insane to think about. I mean, think about the progress we've made in the last 100 years and then imagine AI compressing that into a decade or a year or even months. We're literally creating digital scientists.
Now, in the next part of the blog, he starts throwing out dates. He says, "In 2025, we've seen the arrival of agents that can do real cognitive work. Writing computer code will never be the same. In 2026, we'll likely see systems that can figure out novel insights. And in 2027, we may see the arrival of robots that can do tasks in the real world." Those are his predictions for the next few years.
Then he moves into the 2030s. He writes, "In the most important ways, the 2030s may not be wildly different. People will still love their families, express their creativity, play games, and swim in lakes. But then, in still very important ways, the 2030s are likely going to be wildly different from any time that has come before. We do not know how far beyond human-level intelligence we can go, but we are about to find out." So, here's what the 2030s might actually look like. He predicts that intelligence and energy, or in other words, ideas and the ability to bring those ideas to life is going to become wildly abundant. With abundant intelligence and energy and good governance, we can theoretically have anything else.
And here's where it gets even crazier. Altman points out that we're already kind of used to this. What used to blow our minds—like GPT-3 writing a paragraph—now feels kind of basic. We go from "wow, it can write a poem" to "okay, but can it write a book?" From "it can help diagnose cancer" to "cool, but can it invent the cure." That's how this gentle singularity works: wonders become routine, then expected, then required just to keep up. And the wildest part is AI isn't just accelerating science; it's accelerating AI itself. He says scientists are already reporting they're two to three times more productive using AI. And now they're using AI to build better AI. He admits it's not full-on recursive self-improvement yet, but we're in the very early stages of it. Essentially, the tools are starting to build better tools. The takeoff has begun, and OpenAI actually just released 03 Pro, which might be one of the first real glimpses of that next step; a model that's not just smarter, but feels like it's starting to improve how it improves. I'll be breaking this release down more in detail in my weekly AI recap later this week once there's more information on it, so keep an eye out for that.
Now, Altman's not the only one saying this. Sundar Pichai, the CEO of Google, recently said something similar on the Lex Fridman podcast, that what makes AI fundamentally different from past breakthroughs is its ability to literally improve itself.
Check this out: "Great question. Look, many years ago, I think it might have been 2017 or 2018. Um, you know, I I said at the time like, you know, AI is the most profound technology humanity will ever work on. It'll be more profound than fire or electricity. So, I have to back myself. I, you know, I still think uh that's the case. You know, when you asked this question, I was thinking, well, do we have a recency bias, right? You know, like in sports, it's very tempting to call the current person you're seeing the greatest player, right? and and so is there a recency bias and you know I do think uh from first principles I would argue AI will be bigger than all of those I didn't live through those moments you know two years ago I had to go through a surgery and then I processed that there was a point in time people didn't have anesthesia when they went through these procedures at that moment I was like that has got to be the greatest invention humanity has ever ever done right so look We we don't know what it is to have uh lived through those times but you know and many of what you're talking about were kind of this general things which pretty much affected everything you know electricity or internet etc. But I don't think we have ever dealt with a technology both which is progressing so fast, becoming so capable. It's not clear what the ceiling is and the main unique it's recursively self-improving, right? It's capable of that. And so the fact it is going it's the first technology will kind of dramatically accelerate creation itself like creating things building new things can can improve and achieve things on its own right I think like puts it in a different league right and so uh different league and so I think the impact it'll end up having uh will far surpass everything we've seen before uh obviously with that comes a lot of uh important things to think and wrestle with. But I definitely think that'll end up being the case."
So yeah, I think this is what either excites people, terrifies people, or both; the idea that we're on the verge of creating self-improving AI. And while that could take us to new heights, especially in scientific discovery, it also means we won't be fully in control of how fast or how far this thing goes.
Altman writes, "From here on, the tools we have already built will help us find further scientific insights and aid us in creating better AI systems." Of course, this isn't the same thing as an AI system completely autonomously updating its own code, but nevertheless, this is a larval version of recursive self-improvement. So, yeah, not Skynet just yet, but we're on our way.
Now, on top of all that, Altman points out that there are other self-reinforcing loops kicking in. For one, the economic value creation has started a flywheel of compounding infrastructure buildout to run these increasingly powerful AI systems. And then he drops this: Robots that can build other robots and, in some sense, data centers that can build other data centers aren't that far off. If we have to make the first million humanoid robots the old-fashioned way, but then they can operate the entire supply chain, digging and refining minerals, driving trucks, running factories, etc., to build more robots, which can build more chip fabrication facilities, data centers, etc., then the rate of progress will obviously be quite different.
So this is the part I think a lot of people overlook. When we talk about AI, we usually focus on the software, the models, the code, the outputs. But to actually create AI at scale, you need two things: software and hardware; intelligence and the machinery to run it. And if the hardware starts building itself, well, that's not just automation anymore; that's compounding intelligence infrastructure.
So now that we've got robots building robots, what happens next? Altman writes, "As data center production gets automated, the cost of intelligence should eventually converge to near the cost of electricity, which is a wild thing to think about. I mean, if that's true, then intelligence essentially becomes a utility like electricity or water, except it can write code, run your company, and make scientific discoveries." He even throws in a fun stat here: A single ChatGPT query uses about .34 W hours; roughly what your oven uses in 1 second or what a high-efficiency light bulb uses in 2 minutes. And in terms of water usage, about 1/15th of a teaspoon. So yeah, it's not nothing, but it's also getting infinitely more cost-effective. I mean, going back to 03, not only did they drop 03 Pro, but they also reduced the price of 03 by 80%. So this is happening extremely fast.
Now, from there, Altman shifts gears a bit. He says that even as progress accelerates, humans are surprisingly good at adapting. And sure, there will be hard parts like entire job categories disappearing, but the world will also be getting richer fast, and that opens the door to policy shifts we couldn't even consider before. He doesn't predict some sudden new social contract, but he does suggest that in hindsight these slow shifts will look massive. He even gives the example of how a subsistence farmer from a thousand years ago would look at what many of us do today and say we have fake jobs and think that we're just playing games to entertain ourselves since we have plenty of food and unimaginable luxuries.
Now, here's where Sam Altman goes full philosophical for a second. He writes, "Looking forward, this sounds hard to wrap our heads around, but probably living through it will feel impressive but manageable. From a relativistic perspective, the singularity happens bit by bit and the merge happens slowly. We are climbing the long arc of exponential technological progress. It always looks vertical looking forward and flat going backwards, but it's one smooth curve. Think back to 2020 and what it would have sounded like to have something close to AGI by 2025 versus what the last 5 years have actually been like." So yeah, even though, in the grand scheme of things, we're living through exponential technological progress or the singularity, it's not necessarily going to feel like it to us since we're actually living through it. This kind of goes back to the start of the blog post where he declares we've crossed the threshold and yet nothing has really changed.
Now, obviously, it's not all rainbows and exponential curves. Altman makes it clear there are serious challenges ahead. For one, the alignment problem: making sure AI systems actually do what we want them to do. He compares today's social media algorithms to misaligned AI. They're great at optimizing for clicks, but terrible at aligning with your actual long-term values. They essentially exploit you and hack your psychology, even though they're technically doing exactly what they were told. Now, imagine a misaligned super intelligence. How detrimental that can be. But on the flip side, imagine this power being used for something good. Altman believes the best path forward would be to first solve the misalignment problem and then make super intelligence cheap, widely available, and not too concentrated with any person, company, or country. I just hope we can do it in that order.
Altman then ends on a pretty poetic note. We're building a brain for the world; a brain that's personalized, always on, and capable of helping anyone with any idea. And he says the real bottleneck might not be compute or safety or data, but just the availability of good ideas. He even throws a little love to the idea guys; the people who've been memed into irrelevance in startup circles, but now might finally be getting their shot. And finally, he closes out with this: "Intelligence too cheap to meter is well within grasp. This may sound crazy to say, but if we told you back in 2020 we were going to be where we are today, it probably sounded more crazy than our current predictions about 2030. May we scale smoothly, exponentially, and uneventfully through super intelligence."
So yeah, I mean, he's not wrong. The last 5 years have been completely unreal. And if the next 5 years look anything like what he's describing in this post, then the gentle singularity might not feel so gentle for long.
Anyway, if you made it this far, I'd love to hear your thoughts. Are we really past the point of no return, or is this just Sam Altman saying some Altman things? I mean, I appreciate his optimism, and I actually enjoy reading these posts, but even for me, sometimes it all feels just a little too far ahead of where we are. But let me know what you think in the comments. Also, hit that subscribe button if you haven't already.