Shulou(Shulou.com)11/24 Report--

According to the news on April 3, the ultimate goal of OpenAI is to develop secure general artificial intelligence. Sam Altman, the company's chief executive, needs to use technology gains to advance artificial intelligence as well as non-profit to ensure the security of artificial intelligence technology. His close friends call it the art of balance.

Altman, 37, chief executive of OpenAI, has always dreamed that computers can communicate and learn from each other like humans. He is at the forefront of the artificial intelligence craze because of ChatGPT. Altman says what he remembers most is staying up late in his bedroom playing with the Macintosh LC II, a gift he got on his eighth birthday. At that time Altman suddenly realized: "one day, computers will learn to think."

In recent months, Altman has not only been building the future of artificial intelligence, but also trying to commercialize it. Last November, his company OpenAI released ChatGPT, a chat robot that can write articles like humans and has become one of the most viral products in the history of technology. In the course of ChatGPT's popularity, OpenAI has grown from a small non-profit to a multibillion-dollar company at a near-record pace, according to investor filings. This is partly due to the fact that OpenAI set up a for-profit unit and received $13 billion in financing from Microsoft.

This success is also part of a delicate balancing act. Altman says he is also worried about what will happen if artificial intelligence is put into society regardless of the consequences. He and others co-founded the non-profit research organization OpenAI eight years ago, when Altman thought it dangerous to use profits to drive the development of powerful models of artificial intelligence.

Altman is wary of using profits to encourage artificial intelligence development. He says he doesn't have a direct stake in OpenAI, which is rare in Silicon Valley. You know, the founders of successful Silicon Valley startups usually get rich through their equity holdings.

"like most people, I like to see wealth numbers rise, but I just don't want it to be the whole factor," Altman said of the economic benefits. " OpenAI said Altman's pay was "moderate", but did not give a specific figure. Mr Altman revealed that he had a small stake in a venture fund that invests in OpenAI, but that it "doesn't matter".

Altman made a lot of money by investing in start-ups in his early years and "made more than I needed". He owns three properties, including a mansion in a San Francisco neighborhood and a vacation home in Napa Valley. He hired dozens of employees to manage real estate and family investments.

Altman's goal is to establish a new order and liberate people with machines so that people are free to do more creative work. In Altman's vision, everyone would receive a basic income to make up for the work that was replaced by artificial intelligence. He even believes that human beings will like artificial intelligence very much, and advanced chatbots can become "an extension of people's will."

In the long run, Altman says, he wants to establish a global governance structure, monitor decisions about the future of artificial intelligence, and gradually reduce the influence of OpenAI's executive team on artificial intelligence technology.

Supporters say Altman is the ideal person to lead OpenAI. But there are also many people who say that Altman's business acumen is too developed to be immersed in Silicon Valley thinking and cannot afford the technological revolution that is reshaping business and social life.

In January, OpenAI signed a $10 billion deal with Microsoft to take a 49 per cent stake in OpenAI's for-profit division, according to investor filings. This corporate partnership, coupled with a series of initiatives by Altman to actively promote the commercialization of artificial intelligence technology, disappointed the early leaders of OpenAI. They believe that these decisions go against the original promise to develop artificial intelligence beyond the influence of shareholders.

Elon Musk, chief executive of Tesla, also criticised OpenAI. Musk co-founded OpenAI in 2015, but left in 2018 because of a dispute over control and direction. In February, Musk said on Twitter that OpenAI was founded as an open source non-profit organization "to check and balance Google, but now it has become a company with closed resources, pursuing maximum profits, and is actually controlled by Microsoft. I didn't mean it at all."

When asked about Musk's criticism, Altman paused. "I like Musk," he finally replied. "I pay attention to what he says."

In an open letter released last week, Silicon Valley celebrities, including Mr Musk and Apple co-founder Steve Wozniak, called on the industry to suspend the development of artificial intelligence that is more advanced than GPT-4, OpenAI's latest technology, for six months to avoid a technology race that could spiral out of control.

For a long time, people have imagined that general artificial intelligence is not only good at generating text or images, but also can understand and learn like human beings, and even perform better than human beings. OpenAI doesn't have to be the first team to develop general artificial intelligence, but their ultimate task is to develop secure universal artificial intelligence, Altman said.

According to investor documents, OpenAI sets a profit cap for investors, and any income that exceeds a certain level will flow to its non-profit organizations. The specific value of the profit ceiling ranges from 7 to 100 times the amount of money invested, depending on the investment time. OpenAI and Microsoft have also formed a joint security committee that includes Altman and Kevin Scott, Microsoft's chief technology officer. If the committee thinks that the products released by Microsoft and OpenAI are too dangerous, it has the right to withdraw them.

Peter Thiel, a billionaire venture capitalist and early donor of OpenAI, is a close friend of Altman. The possibility of general artificial intelligence, he says, gave Altman the idea that some similar technology created the universe. Thiel has long supported the idea that humans and machines will one day merge with each other.

After the release of ChatGPT, a series of competing artificial intelligence systems have emerged one after another. In February, Google announced that it was testing its chat robot Bard and opened it to the public in March.

"they are racing to issue press releases," Mr. Altman said of competitors. "obviously, they are lagging behind."

OpenAI promised in its founding charter that OpenAI would abandon existing research if another project was closer to developing general-purpose artificial intelligence. The company says this is to put human security first and avoid competing to build dangerous artificial intelligence systems.

Headquartered in San Francisco's ecclesiastical district, OpenAI's offices are littered with stone fountains, succulent plants and ferns. Altman goes on and on about the central staircase he designed so that the company's 400 employees have a chance to meet up close every day. Employees also have to work face-to-face from Monday to Wednesday. The building also has a university-style cafeteria, cafeteria and library. The design of the library is a combination of Altman's favorite Paris bookstore and Bender Room, the quiet study space on the top floor of the Stanford University Library.

Altman wears a typical tech company CEO outfit: grey sweaters, jeans and shiny white sneakers. He said he grew up in a very simple way.

Altman grew up on the outskirts of St. Louis and was the eldest of four children. His mother is a dermatologist and his father has worked as a lawyer. Family members say Altman's father's real career was running comfortable housing's non-profit organization, which spent years trying to revive downtown St. Louis.

Altman said that one of the lessons his father taught him was, "you should try to help others." even if you don't have time, you have to find a way. "

Altman's mother said her son started using the family's VCR at the age of 2 and changed his ticket home from summer camp at the age of 13. She said that when Altman was in third grade, she began to help local public school teachers solve computer problems. In high school, he transferred to the private John Burroughs School.

"the teachers liked him because he was really smart, hardworking and very sociable," said Andy Abbott, the then principal. "he was funny and had a lot of personality."

Altman later went to Stanford University to do research in an artificial intelligence laboratory. As a sophomore, he co-founded Loopt, a location-based social networking service, which later became part of the first course of Y Combinator, an accelerator for entrepreneurship. Altman then left school.

Loopt has never been a big hit, selling for $434 million in 2012, close to the amount invested by investors such as Sequoia Capital.

Mr Altman then started a venture fund that made friends such as Mr Thiel and Paul Graham, co-founder of Y Combinator. These people helped turn Altman's company into an influential broker in Silicon Valley. Altman has invested in dozens of companies with his own money, including cloud software company Asana and message board site Reddit.

While dealing with Y Combinator, Altman became increasingly concerned that large research labs like DeepMind were developing potentially dangerous artificial intelligence technologies out of the public eye. Musk is equally worried that powerful artificial intelligence machines will dominate human society.

Both Altman and Musk thought it was time to set up their own laboratories. In the end, a team of the two men pledged $1 billion in start-up funding to the non-profit organization OpenAI.

OpenAI stumbled in its infancy and tried some small projects, such as teaching robots how to spell Rubik's cube. In 2017, OpenAI laid off a large number of employees.

OpenAI researchers quickly concluded that the most likely way to achieve general artificial intelligence lies in large language models, computer programs that mimic the way humans read and write. Altman says this model requires a lot of training based on a lot of text content and a lot of computing power. But as a non-profit organization, OpenAI does not have the conditions for large-scale financing.

"We didn't really realize how expensive the project would be," Mr. Altman said. "We didn't know it at the time."

Altman says he explored ways to raise more money for OpenAI that year, such as seeking government funding or launching a new cryptocurrency. "No one wanted to fund this research," he said. "it was a really difficult time."

Tensions between Altman and Musk are rising, according to people familiar with the matter. Musk is frustrated by the slow progress of the research and wants to gain more control over OpenAI.

OpenAI executives eventually came up with a completely different idea to create a for-profit division, OpenAI LP, and report to the non-profit parent organization.

Reid Hoffman, co-founder of recruitment platform LinkedIn, was a consultant to OpenAI at the time and later became a member of the company's board of directors. He said the idea is to attract investors who are eager to benefit from the commercialization of OpenAI technology and accelerate the development of OpenAI.

"you want to be the first person to achieve your goals, you want to set standards," Hoffman said. "that's why the speed of development is part of the moral and ethical issue."

The decision further alienated Mr Musk's operating philosophy from OpenAI, according to people familiar with the matter. In February 2018, he parted ways with OpenAI.

Former employees who attended the meeting said Mr. Musk announced his departure at a staff meeting. Musk explained at the time that he thought he would have a better chance to develop general artificial intelligence through Tesla, because he could get more resources in Tesla.

A young researcher questioned whether Musk had considered safety, former employees said. Musk became obviously depressed, they said, and stunned employees by calling the intern "stupid." This was the last time many of them saw Musk himself.

Soon after, a senior OpenAI executive commissioned a "fool" trophy for the young researcher and later gave it to him on a pillow. "you have to have some fun," Mr. Altman said. "that's part of the company's culture."

Musk's departure was a turning point for OpenAI. Later that year, OpenAI's leadership informed employees that Altman would lead the company. He officially became CEO of OpenAI and helped create a for-profit division in early 2019.

Of the $1 billion initially pledged, OpenAI said it received about $130 million. After the establishment of the for-profit department, OpenAI no longer needs any more donations. Musk said on Twitter that he had donated about $100 million to OpenAI.

Musk announced his resignation at a full meeting, according to former employees who attended the meeting. Musk explained at the time that he thought he could get more resources through Tesla, which would give him a better opportunity to develop general artificial intelligence.

A young researcher asked Musk if he had considered safety, former employees said. They said Musk was clearly frustrated and surprised his employees by calling the intern a "fool". This was the last time many of them saw Musk himself.

Soon after, OpenAI executives commissioned a "fool" trophy and put it on the researcher's pillow as a gift. "you have to have some fun," Altman said. "it's part of the company's culture."

Musk's departure was a turning point for OpenAI. Later that year, OpenAI's leadership informed employees that Altman would lead the company. He officially became CEO of OpenAI and began to create a for-profit division in 2019.

Of the $1 billion donation initially announced, OpenAI said it had received about $130 million. After creating a for-profit department, OpenAI no longer needs any more donations. Musk said on Twitter that he had donated about $100m to OpenAI.

At the same time, Altman began looking for investors. In the summer of 2018, he attended Allen's annual meeting in Sun Valley, Idaho, and happened to meet Microsoft CEO Satya Satya Nadella in the stairwell. Altman introduced OpenAI to Nadella and got his interest. That winter, the two sides began a formal conversation.

"I remember when I came back later, I said it was the only suitable partner," Altman recalled. "they got something safe, they got artificial intelligence. And they have capital and a lot of math."

Former OpenAI employees said Altman published the contract in a timely manner during the negotiations and arranged for all staff to go to work to allay concerns and prevent the partnership from being seen as contradictory to OpenAI's initial commitment to develop artificial intelligence outside the corporate sector.

But some employees still think it's a deal with the devil. OpenAI chief security researcher Dario Amodei (Dario Amodei) and his deputies worry that the deal will allow Microsoft to sell products before OpenAI technology passes adequate security testing. They believe that OpenAI technology is far from ready for large-scale release, let alone working with one of the world's largest software companies. They worry that the technology may malfunction or be abused, causing unpredictable harm.

Some former employees said Mr. Amodei was also concerned that the deal would tie OpenAI and Microsoft together, making it harder for OpenAI to live up to its founding charter promise that it would turn around to help another project if it implemented general artificial intelligence first.

Microsoft initially invested $1 billion in OpenAI. The deal brings OpenAI the money it needs, but it also brings some restrictions. OpenAI agreed to use Microsoft's giant servers only through Azure cloud services to train artificial intelligence models and license OpenAI technology to Microsoft products.

"when you are standing on the edge of a cliff, the only thing you can do is jump and expect to land steadily," Nadella said in a recent interview. " He called it a process of platform transformation.

Gary Marcus, professor emeritus of psychology and neuroscience at New York University and co-founder of a machine learning company, said: "this deal completely undermines the principles that OpenAI follows to gain status as a non-profit organization."

According to Marcus, Altman "made a U-turn, which, in my opinion, is only a verbal expression of concern for human security."

Altman disagrees. "what's unusual about Microsoft as a partner is that it allows us to retain all the principles that we think are important to the mission." These include a cap on the return on the investment and a commitment to turn around to help another project if general artificial intelligence is implemented first, he said.

This funding has undoubtedly accelerated the development of OpenAI and enabled researchers to acquire the key computing power needed to improve large language models. OpenAI quickly developed a more powerful language model, GPT-3, and sold the technology to developers through the application program interface (API) in June 2020.

According to former employees, Altman and Amodei clashed again over the API release. Amodei believes that product launches should be limited to a smaller user base and carried out in phases to reduce public attention and give security teams more time to test.

A few months later, Amodei and others left OpenAI to start a competing artificial intelligence laboratory, Anthropic. "they disagree with us on how to get secure universal artificial intelligence," Altman said. "

Anthropic has received more than $300 million in investment from Google and launched the artificial intelligence chat robot Claude in March this year, which developers can call through API.

According to a recent investor report, Anthropic says it is "committed to large-scale commercialization" to develop secure general-purpose artificial intelligence and plans to release a commercialization approach that "meets the promise" in September. Anthropic was founded as an artificial intelligence security and research company and plans to create business value from its products.

Microsoft invested a total of $3 billion in OpenAI in the three years after the initial deal, according to investor filings.

Last November, ChatGPT gained more than 1 million registered users in just five days after its launch, a speed that even surprised Altman. Before that, OpenAI introduced DALL-E 2, which can generate complex images based on text prompts.

After the spread of ChatGPT, people use the chat robot for a variety of interesting applications, such as writing poems or planning birthday peer-to-peer, causing a heated discussion on the Internet.

By February, ChatGPT had 100 million active users, the fastest spread of consumer apps in history, according to analysts at UBS.

Altman's close friends applaud his ability to balance OpenAI priorities. Mr Thiel says no one is better suited to strike a balance between "idealism" and "reality".

Altman said he delayed the release of the latest version of GPT-4 until March this year in order to conduct more security tests. Many users report some problems when using the GPT-4 model, such as GPT-4 making up wrong answers in Bing search integrated into Microsoft.

"the right way is to get people involved, explore these systems, study them and learn how to make them safer," says Mr Altman. "

After recovering the initial investment, Microsoft will make a profit of about 49 per cent before OpenAI reaches its revenue ceiling, up from 21 per cent previously arranged, according to the documents. OpenAI, the non-profit parent company, will receive residual income.

In addition to OpenAI, Altman has invested in other projects, including Worldcoin, which he co-founded, with the goal of providing cryptocurrency services for everyone.

In recent years, Altman has invested almost all his liquidity in two companies. Altman says he has invested $375 million in Helion Energy, which is studying how to create carbon-free energy through nuclear fusion, which is close to "creating a net energy gain in a real demonstration."

In addition, he has invested $180 million in Retro with the goal of extending human life by 10 years through innovative technologies such as "cell reprogramming, plasma therapy and autophagy".

Morally, these problems are much easier than artificial intelligence, says Altman. "if you are studying nuclear fusion, it must be advantageous. It's a good thing." "if you are developing artificial intelligence, it can be very good or very bad," he said. "

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.