Curated from 30+ sources. Scored for relevance. Never algorithmic. Updated daily.
Meta is systematically investing in and refining its internal infrastructure and processes to enable scalable, responsible, and efficient development of advanced AI models.
Meta AI is enhancing the open-source ecosystem for computer vision and self-supervised learning by releasing three new tools that streamline research and development.
SAM-Audio represents a significant step towards a universal audio segmentation model, mirroring the impact of SAM in visual segmentation.
The collaboration leverages Meta AI's SAM to significantly improve the speed and accuracy of flood mapping from satellite data, directly aiding disaster relief efforts.
Forest signifies Meta AI's ongoing commitment to pushing the frontiers of self-supervised learning for more powerful and data-efficient computer vision systems.
Meta is significantly investing in custom AI chips (MTIA v2) to optimize its massive AI infrastructure for efficiency and reduce dependency on external hardware.
Meta continues to advance its foundational image segmentation capabilities with SAM 3, offering improved precision and efficiency for diverse computer vision tasks.
TRIBE v2 advances the concept of brain-inspired predictive world models, aiming to enable AI systems to better anticipate and interact with complex environments.
Meta is leveraging AI to automate routine seller interactions on Facebook Marketplace, aiming to reduce friction and improve efficiency for users.
This feature aims to streamline communication and reduce manual effort for sellers on Facebook Marketplace by automating routine buyer interactions.