Meta’s superintelligence push fuels internal rifts, sparking employee discontent and talent flight

Meta’s aggressive push to build “personal superintelligence” has triggered internal unrest among its AI researchers, sparking tensions, talent departures, and growing frustration within its existing AI teams. The company’s high-stakes recruitment campaign—offering top researchers from rivals like OpenAI, Google, and Anthropic salaries up to 10 to 50 times higher than their current roles—has intensified Silicon Valley’s talent wars but also created a sense of division among long-time employees. At the heart of the friction is the formation of Meta Superintelligence Labs (MSL), a newly centralized team overseeing the company’s most advanced AI efforts. While MSL has attracted elite talent, including key hires from OpenAI such as Shengjia Zhao, the perceived prioritization of this new unit has left other AI teams, particularly the generative AI group known as GenAI, feeling sidelined. GenAI, responsible for developing the Llama 4 model, has faced criticism for its lukewarm reception. Former team members say the shift in focus to MSL made them feel like their work had been devalued. “Meta has really good researchers. Their treatment of researchers who are not part of superintelligence has been sub-par,” wrote Rohan Anil, a former Llama 4 contributor, in a now-deleted X post. Anil left Meta in June to join Anthropic, a move echoed by other researchers. Some current employees say the new structure feels like a “giant social experiment.” The perception that outsiders are being rewarded with massive compensation and prestige while internal teams are left behind has fueled resentment. One former AI researcher noted that the situation made it feel like Meta was signaling that the GenAI team had failed, even though it had delivered major models like Llama 3. The fallout has led to threats of resignation and internal chaos. Some employees are reportedly quitting to negotiate positions in MSL, while others are considering leaving entirely. Meta has stated it does not counter offers from employees who threaten to leave, a policy that has drawn criticism. The issue extends beyond pay. Access to cutting-edge computing resources and the prestige of being part of MSL’s core “TBD Lab” are major factors. While MSL now oversees GenAI, the original team has seen its autonomy diminished. Meanwhile, Meta’s long-standing AI research arm, FAIR, remains relatively insulated. Yann LeCun, FAIR’s chief scientist, continues to lead his small team focused on I-JEPA, a model he believes could be key to artificial general intelligence. He remains a key advisor to Mark Zuckerberg, though his lab operates with more independence. Other AI labs are capitalizing on the unrest. Elon Musk’s xAI has hired over a dozen Meta researchers in recent weeks, while Microsoft is actively recruiting talent from Meta as well. Despite the tensions, some insiders acknowledge that the reorganization may be necessary. One said the GenAI team had grown too quickly after the ChatGPT boom, making management difficult. Others believe that if Zuckerberg wants to start fresh, some departures could be beneficial. The internal rift underscores the high stakes of Meta’s superintelligence ambitions—and the human cost of racing to lead the next wave of AI innovation.