Facebook parent Meta (META) is spending billions of dollars to grab the lead in the AI race, building out the data centers needed to develop and power high-end large language models. And now, the company is pouring billions more into snatching up top talent and technologies to grab the lead in the AI wars.
In its latest hiring coup, Meta has poached three OpenAI (OPAI.PVT) researchers — Google DeepMind alums Lucas Beyer, Alexander Kolesnikov and Xiaohua Zhai — from its Zurich office, The Wall Street Journal reported late Wednesday. OpenAI CEO Sam Altman claims that Meta CEO Mark Zuckerberg has offered the company’s employees upward of $100 million to join his AI efforts.
Meanwhile, Meta is reportedly in talks to bag Safe Superintelligence CEO Daniel Gross and former Github CEO Nat Friedman for its planned superintelligence lab. Zuckerberg was initially eyeing all of Safe Superintelligence, but was shot down by co-founder Ilya Sutskever, according to CNBC.
The moves come after Meta invested $14.3 billion in AI startup Scale AI and hired its CEO and co-founder Alexandr Wang. That’s not all: Meta also wanted to buy Perplexity AI (PEAI.PVT) but couldn’t come to terms on a deal.
“Meta is doing this because they want to win the AI race, period,” Forrester analyst Mike Proulx told Yahoo Finance. “AI is everything right now.”
All of this follows Meta’s decision to postpone the debut of its massive Llama 4 Behemoth AI model. According to The Wall Street Journal, the company won’t launch the model until later this fall over concerns that it isn’t a big enough upgrade over prior models.
Scale AI founder and CEO Alexandr Wang poses for photos at the company’s office in San Francisco, May 15, 2023. (AP Photo/Jeff Chiu, File) · ASSOCIATED PRESS
“I think this is two things: No. 1 [is] confirmation that Llama is struggling,” Deepwater Asset Management managing partner Gene Munster told Yahoo Finance. “And second, is [it’s] also a sign that Zuckerberg is not OK with that.”
For Meta, the goal is clear: bring in as much fresh talent as possible to push its AI program forward and take the lead in the AI wars.
Meta’s effort to rule the AI world differs from chief rivals OpenAI, Google, xAI, and others. Rather than closing off its AI models, the company offers them as open-source software that developers and companies can use on their own.
Meta imposes some restrictions on how users can take advantage of its models. For instance, the company requires firms to request a license from Meta if their product has more than 700 million monthly active users.
Regardless, Meta’s ultimate goal is to get as many people as it’s comfortable with using and developing products via its AI models. Why not charge everyone who wants to access its software? Because Meta benefits every time a company alters its models, giving it greater insights into how it can improve them down the line.
Story Continues