Why does OpenAI's 700 employees have to come back with Sam Altman?

Why does OpenAI's 700 employees have to come back with Sam Altman?

(Author: Wang Yuquan, Founding Partner of Hywin Capital)

After 5 consecutive days of multiple reversals, OpenAI's palace fighting skit finally has a temporary ending.

On November 22, OpenAI officially tweeted that Sam Altman would return as CEO and formed a new initial board of directors. The members are Bret Taylor, former CEO of Salesforce, Larry Summers, former U.S. Secretary of the Treasury, and Adam D'Angelo, former board member.

This is a happy ending. Sam Altman said he was looking forward to returning to OpenAI, and Microsoft CEO Satya Nadella was also pleased.

Altman's good brother, Greg Brockman, has hurried back to the office to take photos with his colleagues, and he is still holding his laptop and starting to write code. Emmett Shear, who was temporarily brought in by the original board of directors to save the scene, was also very happy and thought it would be good for everyone.

Why does OpenAI's 700 employees have to come back with Sam Altman?

This popular palace fight drama on the whole network has come to an end. According to The Verge, they will add six more directors in the future, and Altman and Microsoft may each have a board seat.

Although there will be some procedural work in the future, OpenAI can basically live a stable life for a while.

The question is, in addition to eating melons and watching dramas, what kind of revelations can we see?

In fact, the hasty dismissal of the CEO by the former board of directors of OpenAI did not damage Altman's image, but highlighted the immaturity of several directors.

It is important to know that in most companies, the function of the board of directors is to supervise and constrain the CEO and prevent the CEO from acting impulsively and aggressively, so the board of directors is usually more mature and makes well-thought-out decisions.

The so-called "coups" in Silicon Valley are not new, and many well-known companies have fired their founders and CEOs.

For example, Apple fired Steve Jobs in 1985, Twitter stepped down from founder and CEO Jack Dorsey in 2008, and Yahoo co-founder Yang Zhiyuan was forced to step down as CEO in 2009.

However, their abdications were precedent, and the board of directors and the founders deliberated and confronted each other before making a prudent decision. And there are clear reasons behind major decisions. For example, Steve Jobs and then-CEO John Scully disagreed on the idea of a new product, while Yang Zhiyuan opposed Microsoft's acquisition of Yahoo.

However, OpenAI's board of directors is different, not only is there no warning in advance, there is no room for negotiation and discussion, and after the dismissal of Altman, there is no explanation of his major mistakes or flaws. Their public claim that the CEO is hiding something from the board of directors and cannot be trusted is a serious accusation that must be supported by facts. As a result, OpenAI has no explanation for this, only to call into question their true motives.

What's even worse is that they also rushed to the doctor, changed three CEOs in three days, and even Emmett Shear, who took office temporarily, did not know why he fired Altman, as if the board of directors had become a mysterious black box.

In fact, OpenAI originally had more board members, such as Musk, LinkedIn co-founder Reid Hoffman, etc., and the rich board members could also check and balance each other, but many people left one after another for various reasons.

The few remaining board members tend to emphasize AI safety. For example, Ilya Sutskever has set up a "Superalignment" or Super Alignment team within OpenAI to solve AI security problems.

Tasha McCauley and Helen Toner are both advisory board members of the International Center for AI Governance. Helen Toner is also an employee of Georgetown University's Center for Security and Emerging Technologies.

Therefore, their views on the need to be vigilant about the development of AI are easily unified, and no matter how big OpenAI has made, it may not be cherished, or even become a fault. OpenAI's board of directors has the highest decision-making power, and after several people discussed it, it was decided to let Altman out.

However, the explanation they gave was unconvincing to the company's employees, which was equivalent to giving a group of immature children huge decision-making power, resulting in a global sensation and a clear statement that employees "do not trust the board".

Why does OpenAI's 700 employees have to come back with Sam Altman?

In contrast, Altman was very mature in this "palace fight" and handled it very decently.

Although he was wronged, and his connections and resources also prevailed, he did not really set off a counterattack like a "cool drama", but calmly accepted the decision of the board of directors. He also tried to negotiate amicably, and even if the negotiation failed, he announced that he would join Microsoft, and he had always considered the interests of OpenAI and did not publicly criticize the board of directors.

In the end, it wasn't he who brought Altman back, but more than 700 employees.

After seeing that Altman was about to join Microsoft, they signed a joint letter overnight demanding that the board resign and reinstate Altman, or they would go to Microsoft together.

It stands to reason that it is not uncommon for companies to change CEOs, and others may also have more management skills than Altman, so why should these more than 700 employees follow Altman?

The other day, I analyzed the structure of OpenAI's company, which combines non-profit and partial profits, and in fact, not only the management has the ideal of developing artificial general intelligence, but I believe that their employees also have the same ideals and beliefs.

But, as Dr. Yu Kai, the founder of Horizon, said, extreme idealism requires extreme realism as the foundation. Finding a balance between realism and idealism is very difficult and sometimes requires great courage and courage.

You know, before Altman became the CEO, OpenAI was still a completely non-profit organization, and it is not impossible to continue to maintain it, anyway, the money spent is fundraised, and the CEO does not need to shoulder the burden of profit.

However, Altman still designed a partial profit structure, and insisted on spending huge costs on the research and development of large models when the future of GPT was uncertain.

Many people think that GPT's success was a miracle and a lot of money, but the point is that Altman didn't need to make heavy bets, and there are very few managers like him who dare to make heavy bets on innovation even when they encounter difficulties, and Google gave up halfway.

Therefore, employees know the true value of Altman better than the board of directors, and believe that only he can lead them to realize the ideal of general artificial intelligence in a realistic way, which is why they would rather sacrifice their own career development than advance and retreat with Altman.

I have always advocated that tech entrepreneurs are the core force driving technological progress, and OpenAI's "Gong Dou" story just confirms this.

This gives us a revelation, when a new round of technological change comes, we should not only pay attention to the cultivation of scientific and technological research and development and technical talents, but also pay attention to the power of scientific and technological entrepreneurs.

(This article represents the author's personal views only)

Read on