Mark Zuckerberg

CEO, Meta

2 minute read

A couple of years ago, it looked like Mark Zuckerberg’s influence may have peaked. Beset by safety scandals, he renamed Facebook to Meta and poured billions of dollars into creating the “metaverse.” The billionaire CEO hoped to foster a new era of technology, but it was a false start. Amid consumer ambivalence and heavy spending on research and development,

Meta’s market valuation slid from a high of $1.07 trillion in summer 2021 to $240 billion in early November 2022, the eve of OpenAI releasing ChatGPT.

Anyone counting Zuckerberg out at that point would have been wrong. The Meta founder was early to the world of AI, first hiring some of the world’s best researchers back in 2013. And after ChatGPT ignited the generative AI boom, he quickly reoriented his company around the technology. Meta’s Llama series of models are now some of the most powerful outside of OpenAI, Anthropic or Google. And in a significant break from those leading labs, Zuckerberg decided to publish the “weights”—essentially the underlying neural networks—of the Llama models online. By allowing the open-source community to build on top of Meta’s AI, Zuckerberg applied unwelcome pressure to his competitors, who profit from selling customers access to their proprietary models. He also aligned Meta with those who believe no single company should have a monopoly on the best technology, helping attract talented AI researchers in a highly competitive labor market. That position may surprise some, given that Meta has in the past been accused by the Federal Trade Commission of monopolistic practices in other areas of its business. And it has rankled some open-source purists, who argue that Llama models’ open source licenses come with restrictions that fall short of true openness. Nevertheless, it has worked wonders on the market, with Meta stock surging to new highs this year. Back in a position of power, Zuckerberg announced in January that his company was now, much like other leading labs, also working toward “artificial general intelligence.”

More Must-Reads from TIME

Write to Billy Perrigo at billy.perrigo@time.com