Meta’s AI Recruitment Campaign Finds New Targets

Mark Zuckerberg is At Warpath, we are recruiting top talent in the AI field for the newly formed Meta Superintelligence Labs. After trying to destroy Openai (and successfully poached it Some top researchers), it appears he has set his sights on the next target.

Over 50 people from Thinking Machines Lab, a 50 startup at Mira Murati, have received offers from Tech Giant that they approached or received. (Murati was like that for those who don’t remember. Previously, he was Chief Technology Officer at Openai.. ) One of those offers was over $1 billion over a multi-year span, so a source with knowledge of negotiations told Wired. The rest was between $200 million and $500 million over a four-year span, multiple sources confirmed. In the first year alone, some staff were guaranteed to make between $50 million and $100 million, sources say (lab spokesman declined to comment).

So far, in Machines Lab thinking, no one person has received an offer.

Meta Communications Director Andy Stone disputed the report in a statement to Wired. “We only made offers to a small number of people in TML, and there was one pretty good offer, but the details are off,” he said. “At the end of the day, I ask all this, who spins this story and why.”

According to a message viewed by Wired, Zuckerberg’s first outreach is modest. In some cases, he sent directly messages to employers asking them to talk to whatsapp. From there, the interviews move rapidly. A lengthy call with the CEO himself follows a conversation between Chief Technology Officer Andrew “Bos” Bosworth and other meta-executives.

This is a premeta super intelligence lab that employs messages Zuckerberg sent to potential recruits (tone hasn’t changed much today):

“We have followed your work on technological advancements and the benefits of AI for everyone over the years. We have made some important investments across research, products and infrastructure to build the most valuable AI products and services for people. We are optimistic that everyone who uses our services has world-class AI assistants to help them do things. Developers build cutting-edge open source models.

In these conversations, Boz exists ahead of time about his vision of how Meta competes with Openai. The tech giant lags behind small competitors in building cutting-edge models, but sources say they are willing to use open source strategies to undercut Openai. The idea is that Meta can commoditize technology by releasing an open source model that competes directly with the ChatGPT maker.

“The pressure has always been there since the beginning of the year. I think we saw Lama 4 rushing out the door,” a Meta source tells me. That’s how Meta’s latest family model was developed. delay There were a lot because we struggled to improve that performance and when it was released Drama About companies that appear to play benchmarks to make other models look better than they actually are.