Sam Altman's latest article: Thousands of days from now, humanity may enter the era of superintelligence
TechFlow Selected TechFlow Selected
Sam Altman's latest article: Thousands of days from now, humanity may enter the era of superintelligence
Deep learning has worked, and as it scales up, its effectiveness is expected to improve further.
Navigating Web3 tides with focused insightsAuthor: Sam Altman
Compiled by: Academic Post
[Editor's Note] Currently, Artificial General Intelligence (AGI) is a goal that many large model companies both domestically and internationally are racing to achieve. AGI is a hypothetical technology capable of matching human intelligence across numerous tasks without requiring specialized training. In contrast, superintelligence goes beyond AGI—it can be seen as a hypothetical level of machine intelligence that vastly outperforms humans in any intellectual task.
In the early hours of today, OpenAI CEO Sam Altman outlined his vision for AI-driven technological progress and global prosperity in a personal blog post titled "The Intelligence Age."
The article paints a picture of AI accelerating human advancement, with Altman suggesting that "superintelligence" might emerge within the next few thousand days.
"We may have superintelligence (ASI) in just a few thousand days (!); it might take longer, but I believe we will get there," he wrote.
He believes AI models will soon become autonomous personal assistants, carrying out specific tasks on our behalf—such as coordinating medical care. At some point in the future, AI systems will become so advanced that they can help us build better next-generation systems and drive comprehensive scientific breakthroughs.
Altman attributes this next leap in societal prosperity to deep learning, writing, "Deep learning works—and as it scales, it predictably gets even better. We are investing more and more resources into it."
Altman argues that if we want as many people as possible to access AI, we must reduce computing costs and make computation abundant (which requires vast amounts of energy and chips). If we fail to build sufficient infrastructure, AI will become a highly limited resource, sparking competition among people and primarily serving as a tool for the wealthy.
Of course, no one can truly predict the future of AI. Despite criticism, as CEO of OpenAI, Altman may have already seen AI advancements not yet widely known to the public. Thus, even though his timeline is broad, his statement has drawn significant attention.
In fact, not everyone shares Altman’s optimism. Computer scientist Grady Booch wrote on X: "I'm deeply tired of all the AI hype: it has no basis in reality—it's merely about inflating valuations, exciting the public, grabbing headlines, and distracting from the real work happening in computing."
"Humans inherently desire to create and empower one another, and AI will allow us to amplify these abilities like never before," Altman wrote in the article. "Much of the work we do today would seem like trivial time-wasting to people centuries ago, but no one looking back wishes they had been a lamplighter."
Academic Post has compiled the following text with minor edits while preserving the original meaning.
Content:
The Intelligence Age
In the coming decades, we will accomplish things that would seem like magic to our ancestors.
This phenomenon isn't new—but it will accelerate. Over time, human capabilities have dramatically increased; we now routinely achieve what previous generations deemed impossible.
We are more capable not because of genetic changes, but because we benefit from societal infrastructure far smarter and more powerful than any individual; in an important sense, society itself is a form of high-level intelligence. Our ancestors—and theirs—built great things, contributed to human progress, and all of us today benefit from their achievements. AI will give humanity tools to solve difficult problems, adding new pillars to the scaffolding we cannot build alone. The story of progress continues, and our descendants will accomplish things beyond our reach.
This won’t happen overnight, but soon we’ll collaborate with AI to do far more than we could without it; eventually, each of us will have a personal AI team composed of virtual experts across domains, co-creating almost anything we can imagine. Our children will have virtual tutors offering personalized instruction in any subject, at any speed, in any language. We can envision similar advances improving healthcare, creating any imaginable software, and much more.
With these new capabilities, we can achieve shared prosperity unimaginable today; in the future, every person’s life will be better than it is now. Prosperity doesn’t guarantee happiness—there are many miserable rich people—but it will tangibly improve lives worldwide.
We can view human history narrowly: after thousands of years of scientific discovery and technological advancement, we’ve learned how to melt sand, add impurities, and arrange it with astonishing precision and microscopic scale into computer chips, run energy through them, and ultimately create increasingly powerful AI systems.
This may be the most important fact in all of history so far. We may have superintelligence in just a few thousand days (!); it might take longer, but I believe we will reach this destination.
How will we achieve the next leap in prosperity?
Three words: Deep learning works.
Fifteen words: Deep learning works, and as it scales, it predictably gets even better—we’re investing more and more resources into it.
That’s it. Humanity has discovered an algorithm that can truly learn any data distribution (or uncover the underlying “rules” generating any data). The more computing power and data available, the more precisely it helps solve hard problems. No matter how long I reflect on this, I still can’t fully grasp its significance.
We still face technical challenges, but fixating on any single obstacle is a mistake. Deep learning works, and we will solve the remaining issues. Much can be said about what comes next, but the key point is clear: AI will improve with scale, leading to meaningful improvements in people’s lives worldwide.
AI models will soon become autonomous personal assistants, performing specific tasks for us—like arranging medical appointments. One day, AI systems will become so capable that they help us design better next-generation systems and achieve sweeping scientific progress.
Technology brought us from the Stone Age to the Agricultural Age, then to the Industrial Age. From here, the path to the Intelligence Age is paved by computing, energy, and human will.
If we want to put AI in as many hands as possible, we must lower computing costs and make computation abundant (requiring massive energy and chips). If we don’t build enough infrastructure, AI will become a very scarce resource, fought over by people, and primarily serve as a tool for the rich.
We need to act wisely and decisively. The arrival of the Intelligence Age is a major development, facing extremely complex and severe challenges. This won’t be a purely positive story, but its potential benefits are so immense that we have a responsibility—to ourselves and future generations—to confront the risks ahead.
I believe the future will be so bright that no one today can fully describe it; a defining feature of the Intelligence Age will be widespread prosperity.
While this will unfold gradually, astonishing victories—restoring the climate, establishing space colonies, discovering all of physics—will eventually become commonplace. With near-infinite intelligence and abundant energy—the ability to generate great ideas and the power to realize them—we can achieve extraordinary things.
Like other technologies, AI will also bring downsides. We must start now to maximize its benefits and minimize its harms. For example, we expect this technology to significantly reshape labor markets in the coming years (with both positives and negatives), but most job transformations will happen more slowly than people anticipate, and I’m not worried about running out of meaningful work (even if such work doesn’t resemble “real jobs” by today’s standards). Humans inherently desire to create and empower one another, and AI will allow us to amplify these abilities like never before. As a society, we’ll return to an expanding world where we can once again focus on playing positive-sum games.
Most of the work we do today would seem like trivial busywork to people centuries ago, but no one reflecting on history wishes they had been a lamplighter. If a lamplighter could see today’s world, the surrounding prosperity would seem unimaginable. If we could fast-forward a century from now, the prosperity around us would also seem unimaginable.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
