world models
Meta AI chief Yann LeCun is betting $3.5 billion that world models—not language models—will achieve true machine intelligence. This architectural pivot could reshape AI video generation and physical simulation.
AI training data
The Wikimedia Foundation announces paid agreements with Amazon, Meta, and Perplexity for AI training data access, marking a shift in how tech giants source information for their models.
Meta AI
Meta's V-JEPA 2 challenges the assumption that generating photorealistic video means understanding the world. The architecture reveals why predicting latent representations may outperform pixel-level synthesis.
JEPA
Meta's Chief AI Scientist argues current generative models are fundamentally flawed. His Joint Embedding Predictive Architecture offers an alternative that could reshape how AI understands video and reality.