Mark Zuckerberg says that when it comes to attracting top artificial intelligence researchers, the much-discussed hefty paycheck is only one factor. Speaking on The Information's TITV on Monday, the Meta CEO noted these highly sought-after specialists have two other priorities: minimal management responsibility and maximum access to computing power.
Zuckerberg sums up the ask like so: "I want the fewest number of people reporting to me and the most GPUs"—referring to the high-powered chips essential for training and operating AI models. "Having the most compute per researcher is a strategic advantage, not just for doing the work but for attracting the best people." Business Insider backs up Zuckerberg's take with a 2023 comment from Perplexity's Aravind Srinivas, who recounted trying to unsuccessfully woo a Meta AI researcher who told the CEO, "Come back to me when you have 10,000 H100 GPUs."
On the computing front, Zuckerberg wrote in a Monday Threads post that Meta is building a number of massive data centers to fuel its AI efforts. Bloomberg reports "most data centers today house only hundreds of megawatts of capacity apiece," but Zuck's plans include "Hyperion, which will be able to scale up to 5GW over several years," he wrote.