Llama, FAIR, and Meta's Superintelligence push.
Meta ships the Llama family of open-weight foundation models and runs two of the field's most influential research groups: FAIR, led by Yann LeCun, and Meta Superintelligence Labs, led by Alexandr Wang. Its bet is that open models, distributed across billions of users on Facebook, Instagram, WhatsApp, and Ray-Ban Meta glasses, are the fastest path to scale.
Meta AI news — Llama model releases, FAIR research, and the Meta Superintelligence org — from Meta directly, plus Yann LeCun and Alexandr Wang.
Meta's V-JEPA explainer gives readers the primary-source version of LeCun's long-running critique of next-token-only systems.
Meta's V-JEPA research post explains the joint embedding predictive architecture behind LeCun's long-running world-model thesis.
Meta's official Llama 4 announcement is the source for Zuckerberg's open-model strategy and the company's multimodal AI push.
Meta's leadership page confirms Wang's role leading MSL after joining from Scale AI.