In an interview at the Aspen Ideas Festival on Tuesday, Mustafa SuleymanCEO of Microsoft AI, made it very clear that he admires Open AI CEO Sam Altman.
CNBC's Andrew Ross Sorkin He asked what the plan will be when Microsoft's huge AI future is not so dependent on OpenAI, using the metaphor of winning a bicycle race. But Suleyman dodged it.
“I don't believe in the metaphor that there is a finish line. This is another false frame,” she said. “We have to stop framing everything as a cut-throat race.”
He then proceeded to follow Microsoft's corporate line on his company's deal with OpenAI, in in which he invested about 10,000 million dollars through some combination of cash and cloud credits. The deal gives Microsoft a large stake in OpenAI's for-profit business and allows it to integrate its AI models into Microsoft products and sell its technology to Microsoft cloud customers. Some reports indicate that Microsoft may You will also be entitled to some payments from OpenAI.
“It is true that we have fierce competition with them,” Suleyman said of OpenAI. “They are an independent company. We do not own or control them. We don't even have members on the board of directors. So they do what they want. But we have a deep relationship. I am very good friends with Sam, I have great respect and confidence in what they have done. And that's how it's going to be for many, many years to come,” Suleyman said.
For Suleyman it is important to express this close/distant relationship. Microsoft's investors and enterprise customers appreciate the close relationship, but regulators became curious and in April, The EU agreed that its investment was not a genuine acquisition. If that were to change, in all likelihood so would regulatory involvement.
Suleyman says he trusts Altman on AI safety
In a sense, Suleyman was the Sam Altman of AI before OpenAI. He has spent most of his career competing with OpenAI and is known for his own ego.
Suleyman was the founder of artificial intelligence pioneer DeepMind and sold it to Google in 2014. He was reportedly placed on administrative leave following allegations of bullying employees. as Bloomberg reported in 2019, then moved on to other roles at Google before leaving the company in 2022 to join Greylock Partners as a venture partner. A few months later, he and Greylock's Reid Hoffman, a Microsoft board member, launched Inflection AI to build their own LLM chatbot, among other goals.
Satya Nadella, CEO of Microsoft I tried but failed to hire Sam Altman last fall, when OpenAI fired him and then quickly reinstated him.. After that, Microsoft hired Suleyman and much of Inflexion in March, leaving behind a small business and a big check. In his new role at Microsoft, Suleyman has been auditing the OpenAI code, Semafor reported earlier this month. As one of OpenAI's previous big rivals, it can now take a deep dive into the crown jewel's enemy competitor.
There is yet another problem in all this. OpenAI was founded on the premise of conducting research on AI safety, to prevent a one-day evil AI from destroying humanity. In 2023, while still an OpenAI competitor, Suleyman published a book titled “The Coming Wave: Technology, Energy, and the Biggest Dilemma of the 21st Century” with researcher Michael Bhaskar. The book discusses the dangers of AI and how to prevent them.
A group of former OpenAI employees signed a letter Earlier this month they raised fears that OpenAI and other AI companies are not taking security seriously enough.
Asked about that, Suleyman also proclaimed his love and confidence for Altman, but also that he wants to see both regulation and a slower pace.
“Maybe it's because I'm a Brit with European leanings, but I don't fear regulation in the way everyone seems to do by default,” he said, describing all this finger-pointing by former employees as a “healthy dialogue.” . .” He added: “I think it's fantastic that technologists, entrepreneurs and company CEOs like Sam and I, who I love very much and think is incredible,” are talking about regulation. “He is not cynical, he is sincere. “He genuinely believes it.”
But he also said: “Friction will be our friend here. “These technologies are becoming so powerful, they will be so intimate, they will be so ever-present, that this is a time when it is okay to take stock.” If all this dialogue slows down AI development by six to 18 months or more, “it's time well spent.”
It's all very cozy between these players.
Suleyman wants cooperation with China and AI in classrooms
Suleyman also made some interesting comments on other topics. On the AI race with China:
“With all due respect to my good friends in DC and the military industrial complex, if the default framework is that it can only be a new Cold War, then that’s exactly what it’s going to be because it’s going to become a self-fulfilling prophecy. They’re going to fear that we’re going to fear that we’re adversaries, so they have to be adversaries and this is only going to escalate,” he said. “We have to find ways to cooperate, be respectful of them and at the same time recognize that we have a different set of values.”
On the other hand, he also said that China is “building its own technological ecosystem and they are spreading it around the world. “We should really pay close attention.”
Asked for his opinion on children using AI for schoolwork, Suleyman, who said he has no children, played it down. “I think we need to be a little bit careful about fearing the downsides of every tool, you know, just like when calculators came out, there was kind of a gut reaction, oh no, everyone will be able to solve all the equations instantly. And it will make us dumber because we couldn’t do mental math.”
He also imagines a time, very soon, when AI is like a teacher's assistant, perhaps chatting live in the classroom, as AI's verbal skills improve. “What would it be like for a great teacher or educator to have a deep conversation with a live AI in front of their audience?”
The bottom line is that if we want the people who build and profit from AI to govern and protect humanity from its worst effects, we may be setting unrealistic expectations.
Leave feedback about this