← Back to stories A scientist interacts with a robot helper, demonstrating modern technological innovation.
Photo by Pavel Danilyuk on Pexels
凤凰科技 2026-04-09

Meta’s First Work After Spending 14.3 Billion to Poach Talent: Alexandr Wang Launches Closed-Source Model, Yang Likun Likes It

Report

It has been reported by ifeng (凤凰网) that Meta — after a hiring spree reportedly costing 14.3 billion (currency not specified in the report) to attract AI talent — has produced a first public deliverable: a closed‑source AI model linked to Alexandr Wang. The article says the model has drawn public praise from Yang Likun. Details on the model’s capabilities, deployment plans or the precise nature of Wang’s involvement remain sparse and unverified; the claims should be treated as preliminary.

Alexandr Wang is best known as the founder and CEO of Scale AI, a U.S. data‑labeling and AI services firm. According to the report, his name is now associated with a proprietary model that Meta is positioning as part of its renewed push into foundation models. Yang Likun’s comments were described as positive, though the piece did not provide verbatim quotes or clear context for her endorsement.

Why it matters

Why would Meta move to a closed model after an enormous talent spend? Because closed‑source models can accelerate product rollouts, protect intellectual property and reduce exposure to adversarial or regulatory risks — at least in theory. But closed systems also raise questions about transparency, auditability and the geopolitical tensions around AI: export controls on high‑end chips, U.S.-China technology competition and regulatory scrutiny over large models all shape where and how these systems can be deployed.

It has been reported that this episode underscores a broader industry debate: does chasing top talent and locking models behind proprietary walls deliver practical advantage, or does it invite greater public and regulatory pushback? For now, verification is needed. Meta, Alexandr Wang and Yang Likun have not publicly confirmed the details presented in the ifeng report, and observers will be watching for technical papers, product announcements or regulatory filings that clarify what exactly has been built — and for whom.

AISpace
View original source →