Meta Platforms spends a fortune to gather the brightest minds in the field of artificial intelligence. Managing Director Mark Zuckerberg may have to note it: Research shows that excessive genius in a group can turn boomerang.

So far, more than 12 engineers from Openai have … self -sufficient in Meta, along with remarkable Google and Deepmind experts. Zuckerberg’s bet is that by gathering top talents and giving them unlimited resources, he can gain ground against his opponents and accelerate the development of artificial intelligence systems so advanced that they will approach her ‘General Artificial Intelligence’, The hypothetical point at which the model goes beyond human skills.

As it turns out, overloading a team with a lot of superstars is not always a victorious strategy, something that any frustrated sports fan could tell you. Without specialized management, excess talent in a group can lead to reduced performance – Or even in complete failure – if the “I” collides and the chemistry is pretty bad.

“There is this belief in Wall Street and Silicon Valley that you are simply bringing the most capable people, you are gathering them and the magic is just happening,” said Boris Groysberg, a professor at Harvard Business School, who is studying the dynamics of teams for over two decades. “No, no magic is created. What is created in many cases is a lot of jealousy, back stabs, sabotage. “

At Meta, the removal of superstars from these traps is a task that falls under Alexander Wang, 28 -year -old former Scale AI chief executive, and former GitHub Managing Director Nat Friedman, 48 -year -old Meta.

Decades of academic studies indicate the challenges they will face.

In the 1970s, Academic Management Meredith Belbin noted that groups that consist entirely of high IQs were prone to prolonged disagreements, showed little consistency and difficult to make decisions, with their members more interested in discussion.

In 2011, Gruysburg and others published a study that concluded that on Wall Street, after a certain point, the addition of more all-star analysts to high-level research teams actually harm the performance. There seems to be a turning point, which is usually achieved when the All-Stars with overlapping sectors of specialization roughly represented half of the research team, where analysts’ selfishness took the lead role and began to limit information instead of working together.

Other surveys, meanwhile, show that the performance of a group depends to a large extent on how well its members communicate and cooperate. Groups that allow more discussions among their members tend to have higher ‘collective intelligence’, regardless of the ‘raw’ spiritual power in the room.

‘Hyperoi Management Tips’

Some principles of management become particularly important during the supervision of strong teams. One is that each person’s duties must be clearly defined.

“If everyone has clear ” driving lanes ” (responsibilities), they are not going to see each other as a threat,” said Lindred Green, a professor at the Ross Business Administration School at the University of Michigan. If there are people with a similar background and talent, the grir said that this type of coating is ok – it is enough for the responsibilities to be kept separate.

Another trick is to openly determine from the beginning who will have the decision -making rights on key issues, otherwise competition for power can destroy the team.

“Sometimes the hierarchy has a bad name – you have a leader who tells everyone what to do,” said Anita Williams Wley, a professor at the Tepper School of Business Administration at Carnegie Mellon University. “But in fact, there are some ways in which the existence of a hierarchy helps the teams coordinate.” The hierarchy may change according to the problem, said Wiley, but clarity is vital in an environment where everyone wants to be at the top.

And then there is the team’s chemistry building, which means the development of confidence among the team members, open communication and creation of a sense of common purpose. While there is a lot of research on these aspects of the creation of great teams – from Richard Huckman’s 2002 Leading Teams to Google’s “Project Aristotle” – many leaders are not willing to spend the time needed, Groysberg said.

“We just don’t have many executives and chief executives who have the rigor and patience of implementation,” he said. “I always say, if you need a high performance team on Friday, Thursday is not the right day To start creating it. “

Although warning worldwide, it is particularly important for META, which brings together its oven -up team to prevent Google and Openai.

And then there is money, which almost never fails to complicate things.

With all the widely published attempts by Zuckerberg to attract talents in every way, which include wage offers exceeding $ 200 million, the details of remuneration for many of the newcomers have become common knowledge. This could burden the team building efforts and influence the team’s dynamics.

“For many groups in this kind of environment, the remuneration that people receive is almost like wearing your grade in your army influence – you enter and you have two stars and they have three stars,” said Wiley. “This perfectly determines what the hierarchy is. It will be important for leaders to be quite clear if this is the case here too. “

Michael Del, Managing Director of Dell Technologies, said in a recent interview that excessive earnings packages for new hires in the field of artificial intelligence they could bother Meta veteran employees. “People generally have a sense of justice, right? They want to be treated fairly in relation to others and in relation to the opportunities they have on the market, “he said.

If the dissatisfaction is not immediately apparent, Gruysburg said, it does not mean that it will not be manifested in the next wage cycle.

When stars manage stars

When he was called upon to comment on possible management challenges for the Meta Superintelligence Lab (MSL), a company spokesman told Bloomberg: “We know that there is a great deal of interest in MSL and, it seems, everyone has a view, but we only focus on the development of personal supernatural.”

Zuckerberg challenges some of the press reports on the specific packages it offers to artificial intelligence experts, but defends his strategy for hiring an All-Star team, saying in a recent interview with The Information that artificial intelligence will be something that will be the most important technology. It will support how we develop everything in the company and will greatly influence society. Therefore, we simply want to make sure that we will have the best people to work in it, from entrepreneurs to researchers and engineers who work in the data and infrastructure. “

Zuckerberg undoubtedly bets more on Wang. Meta invested $ 14.3 billion in Scale AI, without taking full control of the company. Scale AI has generated revenue of $ 870 million last year offering data for the training of artificial intelligence systems. Some of its most prominent customers, including Google and Openai, allegedly interrupted their links with the company after Meta’s investment, triggering the debate on whether Meta’s real goal was to acquire Wang and not 49% of the shares he now owns in the business.

Although Wang has a huge talent in mathematics and sciences and now a successful course as a founder – the MIT gave up after a year to co -establish Scale AI in 2016 – neither he nor his newly formed business has produced innovative research on artificial intelligence. At Scale AI, the Wang team worked with a team of external collaborators dealing with data labeling, who provided what it described as “axes and shovels” for the “golden fever” that is today artificial intelligence. At Meta, he now has to win the respect for a team of artificial intelligence scientists worldwide.

Wang has the rumor that he is relentless, and set a demanding pace in Scale AI, stating that “much is the right one”, as he wrote in 2024. He also became known for his attempt to keep the company as “rich in talent”, looking for people who could respond.

According to Gruysburg, who recently wrote a study on Scale AI, Wang participated in weekly recruitment meetings, personally examining each candidate’s details and forcing directors to defend their choices strictly. “Alex is looking at every person with great care in every position,” one manager told Gruysburg. “He is very interested in hiring impressive people. And the way we measure this often is through the balance between IQ and EQ (emotional intelligence). And then you combine it with an extremely large dose of determination. “

It remains to be seen if what worked in Scale will also work in Meta.

“I think Meta has acquired the smartest man I know”, Groysberg noted. “In this area, they acquired Star (Wang). And after the question is: Can they create Star Team? “