“Science and Technology Innovation Board Daily”, November 21 (Researcher Tian Xiao) OpenAI personnel changes continue. In just two days, former CEO Sam Altman went from being forced to “sweep the floor and go out”, to turning over and “trusting” Microsoft, to being jointly retained by many employees. The ups and downs of wrestling are still unfolding in this AI company that has garnered worldwide attention.

On the afternoon of Monday this week, more than 700 OpenAI employees signed a letter saying “it is impossible to work for people who lack competence, judgment, and don't care about our mission and employees.” They urged all board members to resign and reinstate Altman, otherwise they would collectively move to Microsoft. Ilya Sutskever (Ilya Sutskever) also appears in the signature of this letter.

As OpenAI's chief scientist, it was Ilya who convinced the three independent directors to directly “fire” Altman without him present. However, at 5 a.m. local time on the 20th, Ilya posted on the X platform, “I am deeply sorry for my participation in the board of directors. I never thought I'd hurt OpenAI, I love everything we've built together, and I'll do everything I can to bring the company back together.”

Who is Ilya, and what actually happened between him and Altman in this drama from “backstabbing” to making peace?

Who is “Jobs”?

Being swept away by his own company, this experience made people want to compare Altman with Jobs. However, judging from its commitment to technology and its vision of innovation, Ilya is probably more like OpenAI's “Jobs.”

Ilya Sutskevi is an Israeli-Canadian. Born in Russia during the former Soviet Union, she grew up in Jerusalem since she was five years old. After that, speaking Russian, Hebrew, and English, he moved to Canada and studied under Geoffrey Hinton (Geoffrey Hinton), the father of artificial intelligence.

Compared to Altman's rich career in the business world, Ilya has devoted her entire life to artificial intelligence research. He once mentioned that since he was a child, he has considered using artificial intelligence to solve chess problems. He then chose the longest path in artificial intelligence — deep learning.

I chose this difficult path because Ilya is very interested in human “consciousness.” In 2003, the biggest achievement of artificial intelligence at the time was IBM's chess robot Deep Blue, but Ilya believed that Deep Blue was too “simple and crude” and did not give machines the ability to learn. Therefore, how to make machines have a “sense” of learning similar to that of humans is the problem he wants to solve the most.

“The question I'm facing is, can I really understand how intelligence (Intelligence) works and contribute to it? That was my initial motivation.” Ilya's statement is its biggest goal in the field of artificial intelligence.

In 2015, Irya, who has been sitting in a book for a long time, met Y Combinator President Sam Altman. One day, Altman sent him an email: “Hey, let's go have dinner with a bunch of cool people. ”

Just like that, Ilya had dinner with Elon Musk, Greg Brockman (Stripe co-founder), and other influential people. After that meeting, OpenAI was founded, and Ilya became the company's chief scientist, earning $1.9 million a year.

At the time, Ilya was making a big splash on Google. When planning to leave his job, Google offered to increase Ilya's treatment to 4 million to 5 million US dollars, as long as he was willing to stay. Ilya, on the other hand, believes that he has already earned enough money. Next, he wants to create a new platform where technology is not monopolized by private companies. As a result, he came to OpenAI as a non-profit organization full of ambition and spent the $130 million donated by Musk and other bosses within four years.

“I have a very clear goal, which is to make a potentially small but significant contribution to the development of artificial intelligence.” In 2023, once ChatGPT was launched, it brought a major whirlwind of technological revolution to the world. While Altman became famous all over the world overnight, the low-key Ilya continued to reaffirm her “tiny dream” with enthusiasm during the interview.

The concerns of “good people”

Musk once believed that Ilya Sutskevi was the key to OpenAI's success. He said bluntly, “Besides being smart, Sutskvi is also a 'good person' and has a good quality--heart.”

As a “good person,” over the years he has dealt with “consciousness,” what he is most concerned about is when artificial intelligence actually has personal awareness, how to handle the next moral dilemma.

In fact, before ChatGPT came to prominence, Ilya's teacher and father of deep learning, Hinton said bluntly, “I really regret my lifelong work.” He said, “We've almost taught computers how to improve themselves. This is dangerous; we have to seriously consider how to control it.”

As an apprentice and partner with whom she worked for many years, Ilya shared the same sense of crisis and urgency.

In July of this year, Ilya and her colleagues wrote an OpenAI announcement about the Super Smart Alignment or “Super Alignment” project. They warned that while superintelligence could help “solve many of the world's most important problems,” it could also be “extremely dangerous and could lead to human disempowerment and even human extinction.”

“Obviously, no one can build a superintelligence out of control, which is important.” In a few interviews, Ilya emphasized it over and over again.

With this in mind, his mind changed and began working to guide artificial intelligence systems to meet people's desired goals or moral principles rather than pursuing unexpected breakthroughs. This is the root cause of his disagreement with Altman, who aspires to commercialize and pursue technological breakthroughs.

Ilya seems to have a slightly conservative and pessimistic attitude, which is inseparable from his understanding of humans and machines. He once posted on the X platform, saying, “If you think intelligence is more important than all other human qualities, you're going to have a terrible life.” He sees artificial intelligence, which is developing rapidly, as a “child.” Compared to the development of “intelligence,” he wants to pay more attention to the child's moral maturity.

“Just as human parents usually want to help their kids, they want their kids to succeed. They are happy to help children, feed them, dress them, and take care of them. I am convinced that AGI will do the same. We can program AGI this way, design it so that it has a similar deep driving force, and we are happy to achieve that drive. This driving force is to help humanity prosper.” Ilya once said.

As an early supporter of OpenAI, Musk wrote on the morning of the 20th, “Ilya is not a power seeker. Unless he feels it is necessary, he will not take such drastic action.” Faced with Almante, who is suspected of “pulling seeds to fuel”, the “father” Ilya is desperate to protect his “son,” or the reason for the beginning of this farce.

So, why did Ilya, who fiercely opposed Almante, quickly shake hands and make peace with this old friend? We have no way of knowing.

Amidst the turmoil of just a few days, what happened between Ilya and Almante was not only a battle over commercial interests and technological routes, but also a turbulent world view, values, and feelings. Behind this seemingly dramatic story is an expression of the complex and multifaceted inner world of humans. This is a field that machines have yet to touch, and it is also the origin of human existence.