Meta and Alibaba Shake Up AI Leadership
Major tech leadership changes are underway as the AI race intensifies. Mark Zuckerberg has reportedly cut ties with Meta's AI chief as part of a push for flatter, faster orgs. Meanwhile, Alibaba is forming a new task force after the head of its Qwen AI division departed.
Meta's restructuring creates a new applied AI engineering organization led by long-time executive Maher Saba, who will report directly to the CTO, Andrew Bosworth. This new group will feature an unusually flat structure with as many as 50 individual contributors for every manager, a design intended to accelerate decision-making and reduce bureaucracy. The move strategically splits Meta's AI efforts, distributing responsibility between Alexandr Wang's research-focused Superintelligence Labs and Saba's applied engineering team. This organizational redundancy is designed to mitigate risk; if one team encounters obstacles, others can continue to advance on key projects like the AI models codenamed Avocado and Mango. At Alibaba, the departure of Qwen's technical lead, Lin Junyang, follows the exits of at least two other senior AI executives this year: the head of post-training and a staff research scientist focused on coding. Lin, born in 1993, was considered a core architect behind Qwen's rapid progress and a key reason for its competitiveness with fewer resources than rivals. Lin's resignation is reportedly tied to a planned reorganization of the Qwen team. The proposal would shift the division from a vertically integrated system to horizontal teams focused on functions like pre-training and multimodal development, a move that would have reduced Lin's direct management role. Analysts point to an underlying tension between Qwen's open-source strategy, which has cultivated a large global community, and the commercial pressure to generate revenue directly through APIs. This conflict between community-building and monetization represents a core strategic challenge for leaders in the open-source AI space. These leadership shuffles highlight the immense difficulty of scaling AI engineering teams. Unlike traditional software development, AI projects involve non-linear processes and unpredictable edge cases that challenge established organizational structures. As a result, only 22% of organizations believe their current architecture can support AI workloads without significant modification. The broader trend is a fundamental redesign of the tech organization itself, moving beyond treating AI as a simple productivity tool. Companies are restructuring workflows to integrate human-AI collaboration, shifting the role of engineers from executors to orchestrators of increasingly autonomous systems.
