Zuckerberg’s Bold Bid for AI Supremacy Faces Early Setbacks as Key Researchers Jump Ship
- Rapid Departures Signal Instability: Less than two months after launching Meta Superintelligence Labs (MSL), at least eight key employees, including veterans and recent high-profile hires, have exited, highlighting tensions in Meta’s aggressive push to rival OpenAI and Google.
- Talent Wars and Internal Friction: Lavish compensation packages for newcomers have sparked resentment among longtime staff, exacerbated by frequent reorganizations, leading to a wave of resignations to competitors like Anthropic and OpenAI.
- Broader Industry Implications: This exodus underscores the volatile nature of the AI talent market, where Meta’s billions in investments face challenges in retaining expertise amid a high-stakes race for “personal superintelligence.”
In the fast-paced world of artificial intelligence, where companies are locked in a fierce battle for dominance, Meta’s ambitious foray into superintelligence is already showing signs of strain. CEO Mark Zuckerberg‘s vision for Meta Superintelligence Labs (MSL)—a new division announced just months ago to deliver “personal superintelligence” to billions—has been hit by a surprising wave of departures. At least eight employees, ranging from seasoned researchers and engineers to a senior product leader, have left the company, many heading to direct competitors like OpenAI and Anthropic. This talent drain comes amid Meta’s massive investments, pouring billions into closing the gap with industry leaders, and raises questions about the sustainability of its high-stakes bet on advanced AI.
The exits began surfacing shortly after Zuckerberg’s recruiting blitz, where he personally lured top talent with eye-watering compensation packages worth hundreds of millions. These departures aren’t just from fresh faces; many are Meta veterans who built the company’s core AI infrastructure. For instance, Bert Maher, a 12-year stalwart at Meta, played a pivotal role in developing PyTorch, the open-source framework that’s become a cornerstone for AI training worldwide, and Triton, a tool for optimizing AI model efficiency. Maher recently jumped to Anthropic, expressing excitement on X about accelerating their chatbot Claude. Similarly, Tony Liu, after over eight years managing PyTorch GPU systems crucial for Meta’s large-scale AI models, announced his departure on LinkedIn to launch a newsletter sharing insights on scaling AI systems.
Adding to the narrative of instability, Chi-Hao Wu, an AI specialist with more than five years at Meta, left to become chief AI officer at startup Memories.ai, which focuses on AI-driven video analysis. Wu candidly shared with Business Insider that constant reorganizations created a sense of unease among AI teams. “A lot of people in the AI team maybe feel things are too dynamic,” he noted, pointing to multiple manager changes and structural shake-ups. Meta has indeed restructured its AI division several times this year, including dissolving one team and forming four new ones, as detailed in an internal memo. While Meta’s communications chief Andy Stone dismissed media scrutiny as “navel gazing,” these changes have evidently fueled dissatisfaction.
The talent flow isn’t one-way—some are heading straight to Meta’s arch-rivals. Chaya Nayak, a nearly nine-year Meta veteran who directed product management for generative AI, including oversight of the Llama family of large language models and election transparency initiatives, joined OpenAI for special projects. Afroz Mohiuddin, a senior staff engineer poached from Google just last year after 14 years there, also defected to OpenAI’s technical staff. Even research scientist Aram Markosyan, who spent over four years at Meta focusing on safety and fairness in large AI models and features for smart glasses, has departed, though his next move remains unclear.
Notably, the exodus includes some of Zuckerberg’s prized recent recruits, underscoring the cutthroat nature of AI talent poaching. Avi Verma and Ethan Knight, both enticed from OpenAI to join MSL, lasted less than a month before returning to their former employer—Verma didn’t even make it to his start date. Rishabh Agarwal, who arrived from Google DeepMind in April, praised Meta’s “talent and compute density” in an X post but left after five months to join Periodic Labs, a startup using AI for materials discovery founded by ex-OpenAI and DeepMind researchers. Meta downplays these moves, with a spokesperson stating that some attrition is “normal” for a company of its size, especially among longtime employees, and that intense recruiting can lead to last-minute decisions to stay put.
From a broader perspective, this talent churn reflects the intense pressures of the global AI arms race. Meta’s push to compete with OpenAI, Anthropic, and Google involves not just financial muscle but also cultural shifts that can alienate existing staff. The influx of lavishly compensated newcomers has reportedly sparked tensions and threats of further desertions within Meta’s AI operations. As Zuckerberg bets big on superintelligence—envisioning AI that could transform everyday life—these early cracks suggest that building a stable, innovative team is as challenging as the technology itself. In an industry where expertise is the ultimate currency, Meta’s ability to retain and integrate top talent will be crucial to avoiding a brain drain that could derail its grand ambitions.
Meta remains undeterred, emphasizing that most departures are from veterans who’ve contributed immensely and wishing them well. As the AI landscape evolves, with constant innovations and ethical debates, this episode serves as a reminder that even tech giants aren’t immune to the human elements of ambition, loyalty, and burnout. Will Meta stabilize its superintelligence efforts, or will more exits follow? Only time—and perhaps the next recruiting wave—will tell.