Sam Altman, CEO of OpenAI, has just laid the groundwork for the immediate future of AI and its impact on society. In an extensive (very extensive) communiqué filled with interesting reflections, he declared what seems like a point of no return: “We have already crossed the event horizon; the takeoff has begun.”
In his letter, which comes shortly after an interview in which the CEO of Google, Sundar Pichai, also reflected on the future of this technology, he sketches a picture of the present and near future where humanity coexists, grows, and redefines itself through a new era: that of digital superintelligence. The horizon: the 2030s.
Although we are not yet walking among humanoid robots and AI does not yet govern the day-to-day lives of most people, the advancement is undeniable. According to Altman, there are already systems—such as GPT-4 or the recently launched o3-pro—that surpass human beings in multiple cognitive tasks and have begun to transform work, science, and productivity. For him, the most difficult moments of technical development are behind; what lies ahead is expansion, application, and acceleration. As he explains in his publication:
“AI will contribute to the world in many ways, but the improvements in quality of life thanks to the boost in scientific progress and the increase in productivity will be immense; the future can be much better than today. Scientific progress is the greatest driver of overall progress; it is tremendously exciting to think about how much more we could achieve.
In an important sense, ChatGPT is already more powerful than any human who has ever lived. Hundreds of millions of people rely on it daily for increasingly important tasks; a small new capability can have an enormously positive impact; a small misalignment multiplied by hundreds of millions of people can cause a large negative impact.”
A central theme of the text is the impact that advanced artificial intelligence will have on scientific progress. Altman argues that tools like ChatGPT have significantly enhanced the productivity of millions of people and that the pace of discovery could radically multiply.
“In the 2030s, intelligence and energy—the ideas, and the capacity to realize them—will be extremely abundant. These two factors have been the main constraints of human progress for a long time; with abundant intelligence and energy (and good governance), theoretically, we could have anything else.
We already live with incredible digital intelligence, and after the initial impact, most of us have already gotten used to it. We quickly moved from being amazed that AI can generate a beautifully written paragraph to wondering when it will be able to generate a whole novel; from being amazed that it can make life-saving medical diagnoses to wondering when it will develop cures; from being amazed that it can create a small computer program to wondering when it will be able to create an entire company. This is how the singularity advances: amazement becomes routine, and then basic requirements.
We know of scientists who are now two or three times more productive than before thanks to AI. Advanced AI is interesting for many reasons, but perhaps none as significant as the fact that we can use it to research AI faster. We could discover new computational substrates, better algorithms, and who knows what else. If we can do ten years of research in a year, or in a month, then the pace of progress will clearly be different.”
Altman suggests that there are already positive feedback loops where AI not only helps build better AI but also creates economic value that is reinvested in infrastructure—like data centers and robotics—to support new generations of more powerful systems. Concepts such as robots manufacturing other robots and data centers that self-construct are looming on the horizon, which could further accelerate the technological curve.
Moreover, in this scenario, Altman suggests that the cost of artificial intelligence will progressively reduce to equate with that of electricity, making intelligence abundant and almost free.
“The creation of economic value has initiated a cycle of infrastructure building that compounds and accelerates the execution of these increasingly powerful AI systems. And robots that can build other robots (and, in a sense, data centers that can build other data centers) are not that far away.
If we have to build the first million humanoid robots the traditional way, but then they can operate the entire supply chain—extracting and refining minerals, driving trucks, operating factories, etc.—to build more robots, which in turn can build more chip factories, data centers, etc., then the pace of progress will be evidently different.
As data center production becomes automated, the cost of intelligence should eventually converge with the cost of electricity.
One robot creating another: a scene that illustrates the future Sam Altman envisions for 2030.A ChatGPT query consumes what an energy-saving light bulb consumes in a couple of minutes.
In any case, Altman has taken the opportunity to contextualize the cost that the use of AI already implies and to calm critical voices over its ecological impact: “People often ask how much energy a ChatGPT query consumes: the average is about 0.34 watt-hours, what an oven consumes in just over a second, or what a low-energy light bulb consumes in a couple of minutes. It also uses approximately 0.000085 gallons of water, like a fifteenth of a teaspoon.”
Later in his communiqué, he is even more forceful in his predictions, speaking of “an intelligence so cheap that it does not even need to be measured being within reach.” An eye-catching phrase that parallels a famous quote from the energy sector. In the 1950s, it was said that nuclear energy would be “too cheap to meter”—so cheap that there would be no need to charge for consumption or measure it individually. Although this did not turn out to be true for nuclear energy, the expression remained as a metaphor for practically free or extremely cheap access to a resource that was previously expensive or scarce.
Although he foresees radical changes, Altman also emphasizes that many essential aspects of human life—love, creativity, play—will persist. “In the most important aspects, the 2030s might not be so different,” he writes. But in others—such as the way we work, discover, or relate to knowledge—the 2030s “will be radically different from any previous era.”
He also warns about the social challenges that will come with the acceleration of AI: the disappearance of traditional jobs, the risk of power concentration, and the need for a new social contract. Against this, he proposes a broad and equitable distribution of access to superintelligence, as well as a global regulatory framework that collectively defines the boundaries and objectives of these systems. In his text, he mentions “new political ideas that were previously unthinkable,” without specifying further but hinting that they could lean toward wealth redistribution, something currently anathema in Trump’s America:
“There will be difficult parts, like entire classes of jobs that will disappear, but on the other hand, the world will become much richer so quickly that we will be able to seriously consider new political ideas that were previously unthinkable. We probably will not adopt a new social contract all at once, but when we look back in a few decades, the gradual changes will have been significant.
If history is a guide, we will discover new things to do and new desires, and we will quickly integrate new tools (the job shift after the industrial revolution is a good recent example). Expectations will rise, but so will capabilities, and we will all have better things. We will continue building increasingly wonderful things for each other. Humans have a crucial and enduring advantage over AI: we are programmed to care about other people, what they think and do, and we do not care much about machines.
A subsistence farmer from a thousand years ago would see what many of us do today and think we have fake jobs, that we simply play to entertain ourselves since we have unimaginable food and luxuries. I hope we will look at the jobs a thousand years from now and think they are also fake, and I have no doubt they will feel incredibly important and satisfying to those who do them.”
Coexistence of robots and humans in the future of AI
Another of the most urgent topics for Altman is resolving the alignment problem: ensuring that artificial intelligences act in accordance with long-term human interests. To exemplify this, he takes a small jab at the social media sector, speaking of social media algorithms that exploit cognitive biases as cases of misaligned AI with tangible negative effects.
Thus, Altman advocates for global governance that combines freedom of use with clear limits defined by society. The conversation about these boundaries, he insists, must begin as soon as possible.
Altman concludes with a powerful image: superintelligence as a “brain for the world,” highly personalized and accessible. He believes we are about to experience a paradigm shift, where anyone with an idea—once ridiculed for lacking technical means—will be able to carry it out. “The limitations will be the good ideas,” he declares.
An unimaginable, yet close future.
As you can see, Altman’s vision is as bold as it is unsettling. He maintains that in just a decade, we could see advancements that today sound like science fiction: from space colonization to advanced brain-computer interfaces.
The rate of new achievements will be immense. Today, it is hard to even imagine what we will have discovered by 2035; we may go from solving high-energy physics one year to beginning space colonization the next; or from a major breakthrough in material science to high-speed brain-computer interfaces. Many people will choose to live their lives as they are now, but at least some will probably decide to “connect.”
And yet, he asserts that living in that future will be, like all human progress, a gradual process, almost imperceptible: “the singularity happens slowly, and the fusion occurs gradually.”
“We (the whole industry, not just OpenAI) are building a brain for the world. It will be extremely personalized and easy for everyone to use; we will be limited by the good ideas. For a long time, tech-industry startups have mocked ‘idea people’; those who had an idea and sought a team to develop it. Now it seems to me they are about to reach their full potential.
OpenAI is many things now, but above all, we are a superintelligence research company. We have a lot of work ahead, but most of the path is already illuminated, and the dark areas are quickly clearing. We feel extraordinarily grateful to do what we do.”
In a mix of excitement, warning, and hope, Altman closes with a promise: that the ascent toward superintelligence will be smooth, exponential, and without surprises.
Image: ChatGPT
Your email address will not be published. Required fields are marked *
Δ