Meta’s aggressive pursuit of artificial superintelligence is creating significant internal friction as the company poaches top AI researchers from rival labs with compensation packages reportedly 10 to 50 times higher than existing employees receive. The tensions are particularly acute within Meta’s generative AI team, known as GenAI, which developed the Llama 4 model that received a lukewarm reception earlier this year.
Rohan Anil, who worked on Llama 4 before leaving Meta for Anthropic in June, captured the sentiment in a now-deleted X post stating that Meta’s “treatment of researchers who are not part of superintelligence has been sub-par” and described it as feeling “like a giant social experiment.” The exodus comes as CEO Mark Zuckerberg publicly launched an intense phase of AI talent wars, establishing Meta Superintelligence Labs (MSL) to pursue what he calls “personal superintelligence.”
Meta has successfully recruited well over a dozen leading AI researchers, primarily from OpenAI, including ChatGPT co-creator Shengjia Zhao, as well as talent from other competitors. However, this aggressive external hiring strategy has left some existing AI staff feeling undervalued and questioning their contributions. According to former Meta AI researchers, the massive compensation disparity made it feel like Zuckerberg was signaling that GenAI employees had failed.
The internal tensions extend beyond compensation to include access to computing resources and prestige associated with being part of the elite MSL team. Some AI employees are reportedly considering leaving for rival labs, while others are threatening to quit to negotiate positions within MSL itself, creating organizational chaos. Rival companies are capitalizing on this discontent—Elon Musk’s xAI has hired more than a dozen Meta researchers, with some joining as recently as the past few weeks, while Microsoft is also actively recruiting Meta AI talent.
Meta’s FAIR (Fundamental AI Research) lab, led by chief scientist Yann LeCun, has remained relatively insulated from these tensions and maintained its independence. LeCun focuses on advising Zuckerberg and leading a small team of approximately 10 researchers working on I-JEPA, a model he believes is key to achieving artificial general intelligence. Despite the internal turmoil, Meta maintains it has the best retention rates among the Magnificent 7 tech companies and is growing engineering teams two to three times faster than losing them, according to venture capital firm SignalFire.
Key Quotes
Meta has really good researchers. Their treatment of researchers who are not part of superintelligence has been sub-par. It feels like a giant social experiment.
This quote from Rohan Anil, a former Meta researcher who worked on Llama 4 before joining Anthropic, captures the frustration among existing AI staff who feel marginalized by Meta’s superintelligence push and the preferential treatment given to newly recruited external talent.
People join Meta because they want to work with the best in the industry, have access to industry-leading levels of compute, and develop superintelligence. Those pushing certain narratives typically have ulterior motives.
A Meta spokesperson’s response dismisses internal concerns and suggests critics have hidden agendas, reflecting the company’s defensive posture amid reports of internal discord and researcher departures to rival AI labs.
Every action has a reaction; the unintended side effects of creating a SI team.
Erik Meijer, a senior engineering director at Meta until 2024, wrote this on X in response to distinguished research scientist Laurens van der Maaten’s departure to Anthropic, highlighting how Meta’s superintelligence team creation has produced unexpected negative consequences for employee retention and morale.
Our Take
Meta’s internal turmoil reveals a fundamental tension in the AI industry: how do you build the future while maintaining the present? Zuckerberg’s superintelligence gambit appears strategically sound—acquiring top talent is essential in AI’s winner-take-all dynamics—but the execution has created a two-tier system that undermines existing teams. The comparison to the monkey experiment (where unequal rewards trigger anger) is apt: perceived unfairness can be more demotivating than absolute compensation levels. What’s particularly striking is how this internal friction benefits competitors—Anthropic, xAI, and Microsoft are essentially getting pre-vetted Meta talent without paying premium recruitment costs. The situation suggests that in AI’s talent wars, cultural cohesion and organizational trust may be as valuable as computing resources. Meta’s challenge now is whether MSL can deliver breakthroughs that justify the organizational disruption, or whether this becomes a cautionary tale about the limits of talent acquisition as a strategy.
Why This Matters
This story reveals critical fault lines in the race toward artificial superintelligence and highlights how the AI talent wars are reshaping Silicon Valley’s competitive landscape. Meta’s internal tensions demonstrate that aggressive external recruitment strategies can backfire, creating morale issues and potentially driving away existing talent to competitors like Anthropic, xAI, and Microsoft.
The situation underscores the extreme value placed on top AI researchers in an industry where breakthrough capabilities could determine market dominance. With compensation packages reaching 10-50 times existing salaries, the AI talent market has become increasingly distorted, raising questions about sustainability and organizational culture.
For the broader AI industry, this represents a zero-sum talent competition where companies are essentially redistributing the same pool of elite researchers rather than expanding it. The internal discord at one of the world’s leading AI companies suggests that the path to superintelligence may be as much about organizational management as technical innovation. As companies race toward AGI and superintelligence, maintaining cohesive teams while pursuing aggressive growth strategies will prove increasingly challenging, potentially affecting the pace and direction of AI development itself.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- Tech Workers Are the Real Winners in the AI Talent War, With Pay Set to Soar by 2024
- Sam Altman Addresses OpenAI Executive Exodus at Italian Tech Week
- OpenAI’s $157B Valuation: Can It Win the Brutal AI Race?
- Microsoft Pay Data Reveals Significant Salary Premiums for AI Workers
- Amazon Invests Additional $4 Billion in AI Startup Anthropic
Source: https://www.businessinsider.com/meta-ai-talent-war-superintelligence-push-tension-desertion-2025-8