Meta Platforms Inc. has hired four former OpenAI researchers as part of its initiative to strengthen its generative-AI division. The new hires will join the FAIR (Facebook AI Research) team to develop large language and multimodal models for creative applications. Insiders say Meta plans to scale LLaMA training to multi-trillion-parameter levels by mid-2025. The move reflects growing competition among AI labs for talent and computational resources amid rapid commercialization of foundational models.