Artificial intelligence specialists approach the job market as if it were an NBA star, like LeBron James. They seek advice from their entourage and play hardball with those offering more to recruit them.
During the summer, Matt Deitke received a call from Mark Zuckerberg, the CEO of Meta.
Zuckerberg wanted Deitke, a 24-year-old artificial intelligence researcher who had recently contributed to a startup, to join Meta's research efforts focused on "superintelligence," a technology that could hypothetically surpass the processing capacity of the human brain.
Zuckerberg promised him about $125 million in cash and stocks over a four-year period if he joined the company. The offer was not enough to lure Deitke, who wanted to stay at his startup, according to two people familiar with the discussions, as reported by the New York Times. He turned down Zuckerberg.
So Zuckerberg met with Deitke in person. Then, Meta came back with an updated offer of about $250 million over a four-year period, with up to $100 million potentially payable in the first year. The increase in compensation was so unexpected that Deitke consulted his colleagues on what to do. After many discussions, some of them encouraged him to accept the deal - which he did.
The Battle for Bright Minds in the US Intensifies
The competition for talent in the artificial intelligence field in Silicon Valley has become so fierce and extravagant that it increasingly resembles the high-flying salary market of the NBA.
Youthful researchers in artificial intelligence are recruited as if they were LeBron James, with nine-figure compensation packages structured to be paid over several years. To navigate this complex reality, many twenty-somethings have turned to unofficial agents and entourages to devise strategies. And they play hardball with companies to secure large sums, much like basketball players seeking the best deals from teams.
The difference is that, unlike NBA teams, major artificial intelligence companies like Meta, OpenAI, and Google, have no salary cap. This has made the battles for talent in the artificial intelligence field even fiercer, as reported by the American newspaper.
In recent weeks, recruiting free agents in the artificial intelligence field has become a spectacle on social media, much like the period leading up to the transfer deadline in basketball or soccer.
As Meta, Microsoft, Google, and OpenAI poach employees from each other, job postings have been shared online with graphics resembling blockbuster sports transfers, hosted by the online streaming platform TBPN, which airs a show similar to ESPN about the world of technology and business.
"BREAKING: Microsoft Has Recruited Over 20 Employees from DeepMind in the Last Six Months," read a recent TBPN post about Microsoft's hirings from Google's DeepMind lab.
Jordi Hays, a host at TBPN, stated that as technology and artificial intelligence have become mainstream, more people are following the recruitment process "just as our college friends are obsessed with sports and follow the stars, players, and leagues."
Zuckerberg says Meta will continue to invest money in artificial intelligence talent, "because we believe superintelligence will improve every aspect of what we do." Super artificial intelligence will not only enhance the company's business but will also become a personal tool that "has the potential to start a new exciting era of individual empowerment."
Big Money at Stake, Few People in the Market
The job market for artificial intelligence researchers has long had common elements with professional sports.
In 2012, after three academics from the University of Toronto published a research paper describing a fundamental artificial intelligence system that could recognize objects like flowers and cars, they became the subject of a bidding war initiated by the top corporate bidder, Google, which offered $44 million to hire them.
This marked the start of a talent race across the entire tech industry. By 2014, Peter Lee, head of research at Microsoft, likened the market to that of rising professional soccer players, many of whom were earning about a million dollars a year.
"Last year, the cost of a world-class deep learning expert was roughly the same as that of an NFL defender. The cost of this talent is quite remarkable," Lee said at the time to Bloomberg BusinessWeek.
The advantage researchers in artificial intelligence have in negotiating employment terms has increased since OpenAI launched the ChatGPT chatbot in 2022, sparking a frenzy.
They have been aided by the labor shortage: only a small number of people have the technical knowledge and experience required to work on advanced artificial intelligence systems.
And that's because artificial intelligence is built differently from traditional software. These systems learn by analyzing massive amounts of digital data. Few researchers have experience with the most advanced systems, which require massive computing power available only to a few companies. The result has been a new war for talent recruitment, with benefits escalating from millions of dollars per year to hundreds of millions of dollars.
In April, Zuckerberg - whose company was struggling to advance its artificial intelligence research - intervened by sending personal messages to potential recruits, offering them increasingly larger sums.
His approach was similar to that of sports club owners, according to two Meta employees. Even though the offers seemed absurd, if the new hires could contribute to increasing revenue by even half a percent - especially for a company approaching a market capitalization of two trillion dollars - the investment would be worth it, the sources said.
"If I'm Zuck and I'm spending $80 billion in a year just on capital expenses, is it worth investing another $5 billion or more to acquire a truly world-class team that will take the company to the next level? The answer is obviously yes," said Hays.
Meta's initial offers to engineers varied but were around tens of millions of dollars, according to three industry sources.
The company also offered recruits something that was likely more attractive than money: computing power.
Potential hires were told they would be allocated 30,000 graphics processing units, or GPUs, for their artificial intelligence research. GPUs, powerful chips ideal for running calculations that power artificial intelligence, are highly coveted.
Zuckerberg Has a List
Seeking to attract the best talents on the market, Zuckerberg has worked professionally: he recruited with the help of The List, a document containing the names of top minds in the artificial intelligence field, according to two people familiar with the process. Many on the list have three main qualifications:
- a Ph.D. in a field related to artificial intelligence;
- experience in a top laboratory;
- contributions to artificial intelligence research discoveries.
However, artificial intelligence is a narrow field where people usually know each other. They exchange information about the companies to approach for another offer so they can raise their price, the individuals said.
Some researchers on the list have created Slack and Discord chat groups to discuss offers: when someone receives an offer, they can leave the details in the group chats and ask colleagues for their input. By collaborating instead of keeping their received offers secret, they can push salaries higher.
Moreover, working alongside friends can be as important as money. After a researcher joins a new lab, the first thing they do is try to recruit friends, according to two sources.
Who Wins the Offer War
The talent war has started to cause problems. OpenAI has adjusted its internal system to take into account the market shift and is asking those approached by competitors to consult with directors before immediately accepting offers, company employees said.
"Are we fighting back? Yes," said Mark Chen, OpenAI's research director, at a company meeting earlier this month, according to a recording reviewed by the American newspaper. He added that OpenAI did not match Meta's offers because "personally, I believe that to work here, you have to believe in OpenAI's potential."
However, not all offers made by Meta have been successful. The company was turned down by some researchers, said two people, partly because Zuckerberg's vision of artificial intelligence was unclear compared to that of other companies.
Nevertheless, the frenzy has even allowed some lesser-known researchers, like Deitke, to change their destiny.
Deitke, who recently dropped out of a computer science doctoral program at the University of Washington, had been working simultaneously at an artificial intelligence lab in Seattle, the Allen Institute for Artificial Intelligence. There, he led the development of a project called Molmo, an AI-based chatbot that juggles images, sounds, and text - the kind of system Meta is trying to build.
In November, Deitke and several colleagues from the Allen Institute founded Vercept, a start-up trying to build AI-based agents that can autonomously use other software programs from the internet to perform tasks. With around 10 employees, Vercept has raised $16.5 million from investors like former Google CEO Eric Schmidt.
Then came the exchanges between Deitke and Zuckerberg. After Deitke accepted the approximately $250 million offer over four years from Meta, Vercept's CEO posted a suggestive message on social media: "We look forward to joining Matt on his private island next year."