Meta may delay the release of its Behemoth model because they haven't been able to improve Llama enough to justify a public release. Is it surprising when Yann LeCun, Meta's Chief AI scientist, doesn't believe in LLMs and is more focused on JEPA? Two possibilities: Either Yann LeCun is right and LLMs are hitting a plateau, or Meta simply doesn't have the right talent to help them build the best LLMs. I guess we'll just have to see what OpenAI, Anthropic, Google, and DeepSeek release next.
Download the medial app to read full posts, comements and news.