• by Colin Kirkland , Yesterday

As Meta invests billions in massive datacenters, automated ad tools, algorithms, AI chatbots, AI smart glasses and more, the tech giant is forming a team of “elite AI researchers” led by Yang Song, a former TikTok executive now in charge of making Meta’s social apps even more engaging and profitable.

According to a report from Business Insider, the “MRS Research” team exists within the company's Recommendation Systems division, which develops home feed algorithms for Meta’s social-media platforms, working with the Ads division on which users see what content when.

Originally formed in October, MRS Research, its purpose, or its members have been kept secret from the public until recently, when job listings on LinkedIn mentioned the team, including a Research Engineer position that described MRS AI as a team dedicated to “content understanding, user understanding, retrieval, and ranking” across Facebook, Instagram, Threads, and more.

In staffing the team, Meta has pulled in people skilled in recommendation systems, such as previous Amazon AI research Lihong Li, former OpenAI researcher Xiaolong Wang, former Google researcher Fei Sha, and Yang Song, the previous head of TikTok's powerful recommendation algorithms and head of MRS AI.

Within Meta's three-year plan to spend over $600 billion in the U.S. on the development of AI superintelligence, the company has already spent $14.3 billion to acquire a 49% stake in AI startup Scale AI, hiring its CEO Alexandr Wang as the new chief of Meta Superintelligence Labs (MSL).

Around the same time the MRS AI was formed, Meta also hired Andrew Tulloch, the co-founder of AI research and product company Thinking Machines Lab, of which Meta hired two additional AI researchers as well.

More recently, Meta has announced updates to its algorithms and ads systems. This week, for example, the company rolled out an enhanced ad serving process for Instagram powered by a new Adaptive Ranking Model designed to deliver more relevant ads.