Sam Altman, OpenAI: The superintelligence era has begun
OpenAI CEO Sam Altman heralded a bold message that reverberates far beyond Silicon Valley: humanity has crossed the “event horizon” of artificial superintelligence. In other words, the point of no return has arrived and, therefore, the rise of AI is inevitable.
“We are past the event horizon; the takeoff has started,” Altman said. “Humanity is getting close to building digital superintelligence, and at least so far it’s much less weird than it seems like it should be.”
It may sound like science fiction, but Altman is adamant the revolution has started, not in humanoid robots or flying cars, but in invisible digital systems—like ChatGPT—that are presently beating human cognition in many domains. The world has changed, he argues. We just have not caught up yet.
Superintelligence Without the Sci-Fi Look
Though we’re not yet seeing robots roaming the streets or disease eradicated overnight, Altman believes the real action is happening behind the scenes. In AI labs, models are emerging that rival or surpass general human intellect.
“In some big sense, ChatGPT is already more powerful than any human who has ever lived,” Altman said, highlighting its widespread use for “increasingly important tasks” across industries.
This startling observation points to a quiet revolution. AI systems aren’t just assisting with emails or summarizing text anymore—they’re influencing decisions, automating cognitive labor, and gradually replacing knowledge work. And unlike humans, they operate 24/7 and scale instantly across millions.
But this also introduces risk. A minor flaw in such a system can be magnified a million times over, creating real-world consequences—something Altman acknowledges but offers little detail on how to address beyond calls for alignment (more on that later).
A Timeline That Might Change Humanity
Altman shares a timeline that is an almost surreal pace:
- By 2026, Agents that will be able to engage in *real cognitive work* will radically disrupt software development and other knowledge-based industries.
- By 2027, Gall and research that creates *truly* new insights will begin to emerge from AI, unprecedented in the history of these systems; whole papers could be generated, or discoveries revealed.
- By 2028, Robots will perhaps begin to perform *real* physical task in the world. This, once again, will reshape the economics of labor.
“We don’t know how far beyond human-level intelligence we might go, but we are about to find out,” says Altman ominously.
This accelerated roadmap suggests that OpenAI may even already be sitting on these AI capabilities that have been unseen by the public. His forecasts suggest change not just in what machines can do, but in what humanity is, and what will work, creativity, and intelligence even be in an AI-pervasive world?
Recursive Self-Improvement: The AI That Makes Better AI
One of Altman’s most important and scary points is that AI is already creating successors that are more capable than itself.
“Advanced AI is interesting for so many reasons, but probably none more than because we can now do AI research and development faster.”
This is what experts refer to as recursive self-improvement, and we are already seeing a “larval version” of this. Think of it as an accelerating loop: AI creates new models, which build new models, which then create more powerful ones.
Where we once measured technological advancement on the order of decades, we may start measuring it in weeks. Infrastructure scales with demand, economic return flows into R&D, and hardware manufacturing begins to self-automate. Throw into the formula robots that build more robots, and we might eventually be lulled by the rate of exponential growth.
“I can hardly even fathom what we will have discovered in 2035,” Altman contemplates, “Maybe we will go from solving high-energy physics one year and space colonization the next year.”
It sounds outrageous—until you remember that four years ago, ChatGPT was a fantasy. Now, it’s part of daily life for hundreds of millions.
Life in the Age of Superintelligence
Altman thinks that, even amid all these sweeping changes, many of life’s great pleasures will continue. People will still find love, create art, and pick up interesting and beautiful things when they go for a walk.
But society will end up looking something like a completely different reality.
“Whole classes of jobs” will disappear, he adds, and maybe faster than the economy can absorb the resulting unemployed.
The upside, however, is a world so productive and rich that it will allow for the consideration of ambitious policy proposals like universal basic income, four-day workweeks, or public-run AI utilities.
“A subsistence farmer from a thousand years ago would look at most of what we do and say we are doing fake jobs,” says Altman. “Our descendants may well have a similar reaction to what are now our most prestigious jobs.”
Although it may feel messy, Altman believes in the world’s potential for abundance – if we can handle it.
The Alignment Problem: Humanity’s Biggest Challenge
But there is still the enormous challenge of alignment.
“The alignment problem is making sure that AI learns and acts in ways that we collectively really want for the long term,” Altman says.
It is not enough for AI to simply be intelligent. It needs to be aligned with human values—values which we have not even fully articulated. And given the power these systems will wield, any misalignment, even a small one, could have catastrophic outcomes. This is no longer a theoretical consideration: we’ve already seen the impact of misaligned, engagement-driven social media algorithms on information ecosystems. We can only speculate on the results of a misaligned superintelligent system.
Altman advocates for a global conversation about what collective alignment means and who defines it, but such a conversation is not easy in an increasingly fractured world divided along political, ideological, cultural, and power-based lines.
OpenAI’s Mission: A “Brain for the World”
Altman usually talks about OpenAI’s mission in such a way that it directly parallels building “a brain for the world.” He is not being metaphorical.
“Intelligence too cheap to meter is well within grasp,” he points out—like all the other old promises of cheap, unlimited nuclear energy.
These cognitive systems we’ll employ don’t just work with us—they’ll be part of our homes, industrial and institutional lives, and possibly our very own minds. They’ll shape culture, facilitate science, and possibly help delineate law and governance. If OpenAI is successful, its systems may provide the substrate for modern civilization.
This is why transparency, regulation, and distributed control are now more important than ever. A superintelligent system owned and controlled by a few private firms is not just a commercial issue; it is a geopolitical and civilizational issue.
“If we told you in 2020 we would be where we are now, it probably would have sounded crazier than our current predictions for 2030,” Altman explains.
In other words, don’t dismiss today’s “crazy” predictions—they are becoming tomorrow’s reality.
Final Reflections: Is this a Prediction or a Prayer?
Altman concludes with a rather chilling benediction:
“May we smoothly, exponentially, and in a non-eventful manner scale through the superintelligence?”
In that plain sentence is both hope and fear.
The hope is that we can surf this wave without crashing.
The fear is of an evolved species when we may not know what tide we are riding.
Whether you view Sam Altman as a visionary, a pragmatist, an optimist with ambition, or all of the above, there’s one clear thing.
The race to superintelligence is not a future race — the race to superintelligence has already begun.
And the rest of us must now decide on how we wish to live in a world that may soon be ruled by something with far greater intelligence than our own.