Industry Newsllmrecommendation systemsmetacompute infrastructure
Meta Admits LLMs Lag In Ranking
7.2
Relevance Score
Meta CFO Susan Li said at the Morgan Stanley TMT conference in San Francisco on March 4 that the company is not yet broadly using large language model architectures for core ranking and recommendations. She cited compute and data-center capacity shortfalls that push wider LLM deployment toward 2027 or later, while limited LLM use exists for content understanding and Threads ranking.
