On May 16, according to the WSJ, sources said that Meta Platforms (META.O) will delay the release of its flagship AI model, raising internal concerns about its multi-billion-dollar AI investment direction.
Sources stated that engineers are working hard to improve the performance of the large language model called "Behemoth", with employees questioning whether its improvements are sufficient to support a public release. The model was originally scheduled to debut at Meta's first AI Developer Conference in April, then postponed to June, and now further delayed until fall or later.
Despite Meta's previous praise for quickly catching up with competitors like OpenAI and spending $72 billion to realize Zuckerberg's AI ambitions, the "Behemoth" training has been hindered, resulting in performance below expectations. The management team is dissatisfied with the Llama 4 development group and is considering a management restructuring of the AI product division. (Jintian)
