Sam Altman on 'AI Now and Next' in India: A One-Hour Q&A on Math, Employment, Democratization, and AGI
At Express Ada held in New Delhi, India, OpenAI CEO Sam Altman spoke candidly about the current state and future of AI. We summarize the key points of the one-hour dialogue, which covered a wide range of topics: the shift in AI capabilities reaching the level of mathematical research, the impact on India's IT industry, the need for democratization to avoid the concentration of power, and the distance to AGI.
1. Changes in AI capabilities—What has changed in a year?
Altman first illustrated just how much AI capabilities have changed over the past year.
"A year ago, AI could handle high school math reasonably well. Last summer, it became capable of competing in the world's most difficult math contests, and last week, it solved 7 out of 10 research problems that mathematicians had recognized as unsolved."
While this is just about mathematics, he says a similar change is happening in physics. This is an example showing that AI is transitioning from the stage of "organizing known knowledge" to "generating new knowledge."
The same thing is happening in the field of programming. "A year ago, people were amazed by code autocompletion. Now, you can generate an entire application just by inputting an idea." Altman stated, "The profession of programmer has changed more in the last year than at any time since I became an adult."
2. India and AI—From consumers to builders
2-1. India's builder energy
Altman compared his previous visit to this one, describing it as "a transition from the era when India was an AI consumer to the era of builders." Codex (OpenAI's coding agent) is seeing its fastest growth in the world in India, and he praised the energy of the young entrepreneurs he felt at IIT Delhi, calling it "a place where India will surprise the world."
2-2. Impact on the IT and enterprise services industry
On the other hand, he frankly acknowledged the impact on the IT services industry, which accounts for about 8% of India's GDP. "I think the change will be significant. Pretending otherwise is not helpful."
However, he also provided a historical perspective. "Every technological revolution has feared the disappearance of jobs, and every technological revolution has found new jobs on the other side. I think that will happen this time too." What he emphasized was not to deny the change but to "stand on the right side"—that is, to keep adapting to the change. He also used the metaphor that people during the Industrial Revolution could not have predicted jobs like YouTube influencers or CEOs of AI companies.
3. Employment and education—What should we learn?
To the question of how the spread of AI will affect children's education and career choices, Altman said this: "Proficiency with AI tools, resilience, adaptability, the ability to read what people want, and the ability to cooperate with others—these are skills that will be effective no matter what happens."
He does not see employment disappearing entirely, but he also says, "Many professions will end in their current form." He cited his own background as a specific example. "I was trained as a software engineer, but writing code the way I learned is now effectively meaningless."
Regarding the relationship with education, he stated that there are both children who are happy to have "finished all their homework" using AI and children who are moved by "being able to do more than before," and that the latter are the majority. His view is that just as teachers lamented that "there is no point in memorizing" when Google appeared, but eventually adapted, education will also evolve.
4. Concentration of power and democratization—Altman's repeated claims
4-1. Superintelligence must not be monopolized
"No country, no company, and no individual should monopolize superintelligence—including the United States." This was the point Altman emphasized most throughout the entire dialogue.
Regarding competition with China, he evaluated it realistically, saying, "In some areas they are ahead, in some they are behind," and acknowledged that China has a clear advantage in electric motors, magnets, energy deployment, and manufacturing robotics. However, he did not waver in his stance that "it is important for democracy to lead in AI."
4-2. An era where one- or two-person companies can compete with large corporations
“In the last six months, especially the last three, the power that a one- to three-person company can wield is something that was impossible a few years ago,” says Altman. With the spread of agents like Codex, we are entering an era where small teams can produce results equivalent to those of large traditional organizations, which he positioned as the “decentralization of economic power.”
5. AGI and Superintelligence—How far have we come?
Altman stated that “ASI is a few years away,” and described AGI as “quite close.”
“If I had said six years ago that there would be a system capable of generating new knowledge on its own, it would have seemed quite general and intelligent. We are now approaching that stage.”
He cited the biggest mistake in corporate AI adoption as the attitude of companies planning to “formulate strategy in 2026, prepare internal infrastructure in 2027, and implement in 2028.” He noted, “That might work for a large-scale ERP migration, but doing that with AI would be a catastrophic failure.”
6. Altman's positions revealed in rapid-fire
Here are some points clarified in the Q&A format. Regarding data centers in orbit he dismissed it as “currently meaningless in terms of cost-effectiveness.” He stated that arguments criticizing the water usage of data centers are “completely false,” and regarding energy consumption he said, “While it is a challenge in terms of total volume, it may be more efficient than humans on a per-query basis.”
To the question of whether AI will make humans passive, he replied, “I don’t think so,” and predicted that as creators utilize AI, “faster iteration loops will be created, leading to better content.”
The comment, “I don’t ask ChatGPT. I ask smart people,” regarding how to seek happiness, well represents Altman’s own sense of distance from AI.
7. Investment and future outlook
Regarding investment opportunities, he said, “There has never been a time when the investment landscape was as open as it is now.” “Previously, we had to focus on a single frontier, but now there is a feeling that we can do anything.”
He did not reveal details about the new device project with Jony Ive, but stated, “We are trying to create a new family of devices designed for AI that participate in our lives,” hinting at some kind of announcement within the year.
Regarding the achievements of AI development, he maintained a humble stance from the beginning. “This is the story of researchers. The miracle of a small number of researchers who discovered that deep learning could learn anything is the starting point for everything.”