Meta Poaches OpenAI Scientist to Help Lead AI Lab (www.wired.com)

Meta has tapped Yang Song, a senior OpenAI researcher, to be research principal at Meta Superintelligence Labs (MSL), reporting to Shengjia Zhao, the lab’s newly formalized chief scientist. Song, who led OpenAI’s strategic explorations team since 2022, started at Meta earlier this month as part of a broader hiring blitz that brought at least 11 top researchers from OpenAI, Google and Anthropic. The move signals Meta’s push to rapidly scale MSL with heavyweight talent and codifies an internal leadership axis built around Zhao and other OpenAI alumni. Technically, Song brings expertise in scaling models to handle large, complex, multimodal datasets—a background that includes a Stanford-era technique that helped inform DALL·E 2—making him well suited to influence Meta’s work on multimodal generation, dataset curation, and model training strategies. The hire also reflects the competitive, fluid talent market in foundation-model research: a few researchers have already left or returned to OpenAI, and internal roles across Meta’s AI org remain complex (Yann LeCun still leads FAIR). Song’s arrival could accelerate MSL’s capabilities on cross-modal understanding and generation, but underscores ongoing churn and rivalry as leading labs jockey for the engineers and scientists who will shape next‑generation AI.