OpenAI CEO Altman States: China Will Excel in the AI Field
-
In November, San Francisco, in preparation for the 30th APEC Economic Leaders' Meeting (APEC 2023), the city once riddled with homeless people on its streets suddenly became much cleaner. The New York Times jokingly remarked that the air in San Francisco was filled with a 'teenager frantically cleaning up after a party before their parents come home' vibe.
In the Mission District of San Francisco, a famous hippie enclave, stands an inconspicuous gray building with no visible signage. Passersby would never guess that a technological revolution is unfolding inside this building.
This is the headquarters of OpenAI. The night before the interview, I was required to fill out a non-disclosure agreement online and upload my photo. On the day of the interview, I arrived at the building's entrance on time. First, I had to find a button on the gray wall beside the entrance, press it to connect with the front desk, confirm the name of the person receiving me, and then a black iron gate slowly opened.
Passing through the iron gate, the interior of the building reveals a completely different world. The reception area is spacious, designed in the industrial style popular in Silicon Valley. Directly opposite the gate is a massive mural with cool gray tones and intricate lines, resembling a bouquet formed by integrated circuits, with the central flower almost identical to OpenAI's logo. Next to the mural, a wooden wall is covered with vibrant climbing plants. On the sofa facing the reception desk sits a row of people clearly waiting for interviews, casting envious glances at OpenAI employees passing by.
After a brief wait at the reception, 38-year-old OpenAI CEO Sam Altman appears before me, holding a silver Apple laptop. After a short exchange of pleasantries, he skillfully swipes his access card to open the door to a small meeting room nearby.
This marks Altman's first interview with a Chinese journalist since becoming OpenAI's CEO in 2019. Over the next 45 minutes, Altman remains highly focused, with a notepad and pen at hand, occasionally jotting down notes and even asking a few counter-questions. He doesn’t check his phone once, promptly hanging up any incoming calls and apologizing.
Although Altman sidesteps some sensitive questions with his usual diplomatic evasiveness, his expression remains consistently sincere and humble.
"I believe China will excel in the field of AI and will be an important part of humanity's AI exploration journey," Altman stated directly. "I'm very much looking forward to seeing what Chinese entrepreneurs will create."
However, reality appears different. Just a week ago, OpenAI held its first developer day, which thrilled its 2 million global developers. Chinese media even dubbed it the "AI Spring Festival Gala," yet developers in China seemingly didn't receive invitations.
Interestingly, weeks before the developer event, Altman had asked me to provide a recommended list of Chinese developers. Addressing this, he sincerely apologized again: "This year we're just taking our first steps, and it's been very rushed. Our venue space was limited, but next year we certainly hope to invite more participants."
This is not my first interaction with Sam Altman. Before becoming the CEO of OpenAI, Altman was the president of Y Combinator (YC), Silicon Valley's most renowned incubator. During his tenure, he established YC China and invited Qi Lu, then Microsoft's Executive Vice President, to serve as its CEO. This quickly established YC's influence in China's entrepreneurial circle.
At the time, I was a journalist based in Silicon Valley. In September 2014, I led a group of Chinese entrepreneurs to visit Y Combinator's headquarters, where Sam Altman hosted us. Among the entrepreneurs was Zhang Yiming, the then-emerging founder of ByteDance. After returning to China, Zhang wrote a travelogue specifically mentioning that while YC doesn't provide physical space, it connects startup teams with renowned industry entrepreneurs as mentors. YC's only requirement for incubated companies is to focus on building a great product and rapidly scaling their user base, advising them not to worry about other matters—a valuable lesson for domestic incubators.
After that visit, I established contact with Altman and occasionally exchanged insights about China or Silicon Valley. He carried an aura of sincerity and humility. In Silicon Valley, early success is not uncommon, but Altman's consistent modesty and restraint were particularly striking.
A few months after that visit, around the same time OpenAI was founded, Altman wrote an article titled 'China' on his personal blog. In it, he addressed issues that made him seem more like a politician. He began by noting that if measured by purchasing power parity, China's economy had already surpassed the US in 2014.
"As a hopeful American citizen, the most critical question is whether a country can remain as powerful as another with a population four times smaller. The US has never been the most populous country, but it has always been the largest economy. How did the US achieve this? One important method is to maintain excellence in innovation and the development of new technologies. Last century, a significant number of major technological advancements (far exceeding our proportion of the world's population) came from the US."
This new technology was soon revealed by Altman through his actions.
In 2015, Altman and Musk decided to establish a non-profit AI research lab, naming it "OpenAI." The name was Musk's idea, and the lab's goal was to counter Google's growing dominance in the field and "ensure AI does not harm humanity." However, the two had a falling out in 2018. It was reported that Musk withdrew after his proposal to lead OpenAI was rejected, also abandoning his commitment to continue funding OpenAI.
When OpenAI lost its main funding source and was in trouble, Sam Altman successfully persuaded Microsoft CEO Satya Nadella and CTO Kevin Scott at the annual tech leaders conference in Sun Valley, Idaho, to reach a cooperation agreement—Microsoft invested $1 billion in OpenAI, and the two parties also deepened their collaboration on Microsoft's Azure cloud platform.
After that, Altman resigned from all his positions at Y Combinator (YC) to become the CEO of OpenAI, devoting himself fully to the startup. Here, he was not involved in technical development or AI research; his role was more about setting strategic goals and agendas, bringing the team together.
Elon Musk was furious about the alliance between OpenAI and Microsoft and began relentlessly attacking OpenAI and Altman. He tweeted, 'OpenAI was created as an open-source, non-profit company to counterbalance Google, but now it has effectively become a profit-maximizing company controlled by Microsoft.'
In July this year, Musk announced the establishment of his artificial intelligence company x.AI, stating its goal was to 'understand the true nature of the universe.' However, this seems more like another instance of Musk's penchant for dramatic flair. In contrast, Altman has demonstrated his characteristic restraint.
Although Altman claimed in this interview that he is not an expert in politics, he was clearly being modest. Politics has long been an interest of Altman's. As early as 2017, rumors circulated that Altman was considering running for governor of California. He neither confirmed nor denied these rumors but expressed his hope to see someone from the tech community enter the gubernatorial race.
After becoming the CEO of OpenAI, he began continuously advocating for AI safety and regulation, which has provided him with a larger platform.
In May of this year, Altman attended a U.S. Senate hearing on AI regulation. During the hearing, he stated, 'If this technology goes wrong, it could go quite wrong.'
The hearing proceeded quite smoothly, with Altman successfully drawing U.S. political attention to artificial intelligence. Compared to the tense confrontations faced by Meta CEO Mark Zuckerberg and TikTok CEO Shou Zi Chew during their hearings, the atmosphere of this hearing could be described as harmonious.
Using his eloquence and communication skills, Altman answered each senator's questions with remarkable humility and sincerity. Media noted that after the hearing, Altman even approached the committee chairman like a college student, seeking "humble advice." The senators seemed to accept his warnings that AI could "cause significant harm to the world" and called for some regulatory measures for this emerging technology.
Following the hearing, Altman launched his global tour, aiming to share with governments and the public the advantages of AI and the importance of appropriate regulation. He proposed establishing an organization similar to the UN's atomic energy agency to jointly oversee AI.
His first stop was Europe, where he met with numerous leaders, including the British Prime Minister, French President, Spanish Prime Minister, and Polish Prime Minister. Beyond formal meetings with heads of state in suits, he also participated in numerous public events, resembling a political candidate on the campaign trail.
When he arrived to speak at University College London, people queued from the auditorium entrance all the way down the street, stretching across an entire city block. Inside the auditorium, Altman was greeted with the fervor typically reserved for rock stars.
"I am incredibly excited about this technology, which can restore the productivity lost over the past few decades—not just catching up," he reiterated his core belief that the world's two major "limiting factors," the cost of intelligence and the cost of energy, if significantly reduced, should benefit the poor more than the rich. "AI technology will elevate the entire world."
In London, Altman maintained his stance—regulation must strike the right balance. He expressed hope for a regulatory model "somewhere between the traditional European approach and the traditional American approach." He warned that excessive regulation could harm small companies and the open-source movement.
During those weeks in May, he astonishingly visited 22 countries. However, many noticed that Altman's global tour did not include China. Yet, Altman is well aware of China's significance. At the end of his 2015 article China, he wrote with remarkable foresight:
"...What we need to do is find a way to coexist with China. If nothing else, China and the US will remain the world's superpowers for some time. The world is now so interconnected that completely separate governments operating under different rules simply won't work. Perhaps we can find a way to focus on what we each truly excel at while fostering at least partial intergovernmental cooperation, rather than repeating the historical pattern where mutual hostility escalates until conflict erupts."
When I learned that Singapore was one of Altman's stops on his global tour, I texted him asking if he'd be willing to meet with Chinese AI entrepreneurs there, as he had done previously. Altman quickly replied that he was "very interested" and emailed his assistant to arrange a time. However, his assistant expressed difficulty since Altman would only be in Singapore for half a day, with his schedule limited to a university speech.
A few weeks later, after returning to Silicon Valley, Altman participated remotely via video in Beijing's Zhiyuan AI Conference, delivering a speech calling for collaboration on managing AI development.
"China has the best AI talent in the world." He emphasized, "With the emergence of increasingly powerful AI systems, the importance of global collaboration has never been higher."
In July, OpenAI announced the formation of the "Frontier Model Forum" in collaboration with Google and Microsoft, an industry body aimed at ensuring the safe and responsible development of cutting-edge AI models. Altman disclosed to me that he's considering inviting Chinese companies to participate: "We're currently in discussions with Tencent and ByteDance about this matter, and other (Chinese) companies may also be involved."
Recently, I contacted Altman to inquire if he would be willing to grant an official interview to China Entrepreneur during the 2023 APEC meetings, to which he agreed. Thus, this interview came to fruition.
The following is the interview transcript, partially edited:
China Entrepreneur/Pandaily: What do you think about the role China will play in the AI era?
Sam Altman: I believe China will excel in the field of AI and will be an important part of humanity's exploration of AI.
China Entrepreneur/Pandaily: Have you thought about how to collaborate with China and do you have specific plans?
Altman: I think it's still too early. We haven't figured out how to collaborate with other countries yet. But this is definitely something we want to do.
China Entrepreneur/Pandaily: Are you familiar with some Chinese AI companies?
Altman: I'm familiar with some companies, but not others. This year has been extremely busy for me, and I haven't had much time to learn about (other companies).
China Entrepreneur/Pandaily: Before the developer conference, you asked me to send you a list of recommended Chinese companies to attend. But I later learned that none of the companies on the list were invited, and to my knowledge, no Chinese developers were invited this year.
Altman: This year was just our first step, and it was done in a very rushed manner. Our venue space was limited, but next year we certainly hope to invite more people.
China Entrepreneur/Pandaily: Last time you also mentioned wanting to invite Chinese companies to join the Frontier Model Forum and were in discussions with Tencent and ByteDance. Could you share any updates on that progress?
Altman: I am not personally involved in this matter, and the Frontier Model Forum is not operated by our company alone. Overall, I believe the need for collaboration between the U.S., Europe, and China in AI research is very clear. I strongly support cooperation in safety and responsibility—it is clearly an important endeavor worth pursuing.
China Entrepreneur/Pandaily: The movie Oppenheimer was particularly popular in China recently. You once mentioned sharing the same birthday with him?
Altman: I didn’t mention it; others discovered we share the same birthday, and it’s true. (Note: New York Times reporter Cade Metz recorded that he revealed it himself.)
China Entrepreneur/Pandaily: You often mention Oppenheimer in various contexts. What aspect of him do you admire the most?
Sam Altman: I think his ability to effectively combine scientific progress with engineering progress is something I greatly admire. This is very important but also extremely difficult.
China Entrepreneur/Pandaily: Do you see any similarities between Oppenheimer's era and the current AI age?
Altman: I think they are quite different. People like to draw analogies because they do have some similarities. But I believe AI technology and nuclear technology are actually very different, and it is dangerous to overdraw analogies between the two.
China Entrepreneur/Pandaily: You recently attended the AI Safety Summit in the UK but didn’t make any public remarks. Could you share what happened during the closed-door meeting on the second day?
Altman: I participated in some panel discussions but didn’t give an independent speech. I agreed with many of the publicly stated views. However, the key now is to translate words into action. It’s easy for people to say the right things at these summits. But now we need to see whether they will turn their words into action.
China Entrepreneur/Pandaily: How do you usually communicate with Musk?
Altman: We mainly use Twitter DMs now.
China Entrepreneur/Pandaily: I remember you mentioned visiting China many times during your YC days?
Altman: Not many times, just four or five times in total, mainly to Beijing and Shanghai.
China Entrepreneur/Pandaily: You used to interact quite frequently with Chinese entrepreneurs during your YC days, but this stopped after joining OpenAI. What do you think is the biggest difference between entrepreneurs from China and the US?
Altman: People always ask me this question. And I always say, even though they come from different regions and industries worldwide, great entrepreneurs are remarkably similar. They might focus on different fields or other matters, but the spirit required to create something from nothing and truly excel at it is fundamentally similar.
China Entrepreneur/Pandaily: Do you remember that meeting with Zhang Yiming in 2014?
Altman: Of course, he was quite impressive.
China Entrepreneur/Pandaily: But he was very shy and quiet back then.
Altman: He was indeed quite shy back then, but he was brilliant. We've met a few times since then. I remember one occasion was in China around 2017 or 2018, and we met once more after that.
China Entrepreneur/Pandaily: What's your perspective on Chinese competitors working on large models? Do you think they still have opportunities given the chip shortages and geopolitical tensions?
Altman: Absolutely! China has many excellent researchers working in the field of large models. I believe this represents a fundamentally important and valuable scientific advancement. We'll see people worldwide developing AI and utilizing it in various ways. This is completely new territory, and I consider it a positive development.
China Entrepreneur/Pandaily: Do you have any message for Chinese entrepreneurs?
Altman: For a long time, I have been interested in and impressed by the capabilities, spirit, and the amazing tech companies built by Chinese entrepreneurs. I believe we are stepping into a new technological platform and an AI revolution.
The opportunity to build incredible new products and services will redefine how we all use computers and accomplish tasks efficiently. This is an extraordinary time, and I am very much looking forward to seeing what Chinese entrepreneurs will create.
China Entrepreneur/Pandaily: Do you remember last time when I almost brought a group of Chinese AI entrepreneurs to meet you in Singapore, but it didn't happen due to your tight schedule? Will there be another opportunity?
Sam Altman: Absolutely. I've been thinking about doing another similar global tour after some time, though I'm not sure about the schedule yet. These kinds of trips are really exhausting.
China Entrepreneur/Pandaily: Do you have any plans to visit China recently?
Altman: Not recently, but I hope to visit someday. The last time was before the pandemic.
China Entrepreneur/Pandaily: What do you think is OpenAI's core competitive advantage?
Altman: I hope we can be the best in the world in research and innovation, while also having the capability to manufacture and deliver excellent products. I believe having all these capabilities concentrated within one company is a huge competitive advantage for us, and I expect we will continue to maintain this position.
China Entrepreneur/Pandaily: Will you launch your own killer AI applications, or focus on building a solid underlying platform?
Sam Altman: ChatGPT is such an application, and I think it's doing very well.
China Entrepreneur/Pandaily: What about applications like character.AI? It seems you could easily create a similar one.
Altman: Developers can utilize our GPTs to build such applications, but we won't. We aim to be a platform that empowers other developers. I think it's fantastic that people are now using customized GPTs to create these amazing chat experiences.
China Entrepreneur/Pandaily: November 30 marks the one-year anniversary of ChatGPT's release. Looking back, what do you think was the best decision made during this journey?
Altman: You reminded me—we should celebrate the anniversary. As for the best decision, it's interesting because I always think about the bad ones. But we always have the opportunity to correct those bad decisions, which, in my view, is excellent.
China Entrepreneurs/Pandaily: By the way, let's also talk about those bad decisions.
Sam Altman: I think a good decision was our choice to release ChatGPT. It seems obvious now, but it wasn't at the time. We weren't sure if people would like it. So I believe the decision to release it was truly excellent.
Since then, we've improved and expanded it - all good decisions. Later, we decided to make it subscription-based rather than ad-supported, which I also think was a good idea. We added features enabling it to see, hear, and generate images, which proved very popular. I believe focusing on the code interpreter was another smart move.
But there have also been some bad decisions, such as frequently making mistakes in product decisions. We didn't consider the issue of continuous scaling (Note: One hour after the interview ended, OpenAI suspended ChatGPT Plus registrations); and there's also the fact that we can never build enough features to meet user demands.
China Entrepreneur/Pandaily: What is your ultimate vision for ChatGPT?
Altman: I think it's a very effective tool that can make people more efficient. We've already seen this in some industries. What it does to help programmers, for example, is truly amazing.
But I think we can apply it to many other industries. We can improve people's work efficiency and allow them to do more meaningful things, which is great. Moreover, it can also help them learn more knowledge, gain more fun, and so on.
China Entrepreneur/Pandaily: At this Developer Day, you launched GPTs, hoping to help developers create their own ChatGPT. Do you expect a winner-takes-all effect to form? What percentage of total GPTs revenue do you estimate the top 100 GPTs will account for?
Altman: Good question. Honestly I don't know, ask me again in a few months, I really don't know right now.
China Entrepreneur/Pandaily: You often mention scaling laws - achieving better output by continuously increasing computing power. But will this approach eventually hit a bottleneck?
Sam Altman: Continuous scaling is good. But as we've always said, it won't last forever. We must continue to innovate and invent new knowledge.
China Entrepreneur/Pandaily: What do you think native AI applications should look like?
Altman: I think some of the best examples right now are Copilot products like GitHub Copilot. That's one direction—it just plugs into your workflow; ChatGPT is another great example, which essentially serves as your interactive interface.
China Entrepreneur/Pandaily: What principles should developers follow when building their own native AI applications?
Altman: I think it's really worth reconsidering how we use computers at a fundamental level. Computers can now understand quite nuanced things, which is exactly what science fiction has long predicted. But in reality, we haven't deeply thought about how we'll use computers in a future world beyond scenarios like Star Trek or Her.
《中国企业家》/ Pandaily: Can you offer some advice to developers?
Sam Altman: I believe that opportunity knocks but once. Let's look back at an example, and then I'll say what I originally intended to.
After the introduction of smartphones and the launch of the Apple App Store, a wave of companies emerged that couldn't have existed before, such as Uber and Snapchat. These opportunities were very special and only appeared rapidly following the platform shift. Those who seek out new things, new types of companies, and pursue them will greatly benefit. I think this is always very exciting.
China Entrepreneur/Pandaily: What about native AI hardware products? I see you've also invested in some hardware companies like 1X, Humane, Rewind, etc. Are there certain criteria you follow when investing in these companies?
Altman: We're very interested in figuring this out, but to be honest, we don't know yet. I mean, it's still in the exploratory phase - this is a new fundamental capability, and all these new possibilities will emerge. I think it's very important to keep an open mind and continue exploring. Many things only become clear in hindsight.
China Entrepreneur/Pandaily: Do you think native AI hardware products will be created by traditional phone companies like Apple, or by completely new companies?
Altman: I believe both new and traditional companies have opportunities.