Inside Meta's attempts to play catch-up with AI

A year ago, Mark Zuckerberg appointed Alexandr Wang to accelerate Meta’s artificial intelligence initiatives, resulting in Muse Spark, the company’s most significant AI model to date. Zuckerberg’s decision to entrust a then-28-year-old startup founder with this responsibility, rather than a seasoned researcher, aimed to leverage an outsider’s drive and ambition to succeed where Meta’s established AI units had faltered.

Wang has successfully assembled a top-tier research group attracting talent with substantial salaries and has restructured parts of Meta’s AI operations. He has become one of the most influential executives within the company, notably attending a White House dinner with Silicon Valley leaders alongside Zuckerberg. The launch of Muse Spark, the first major model from Wang’s secretive TBD Lab, is viewed by proponents as a clear indication that Meta’s AI rebuilding is gaining momentum, with successor models expected to further narrow the gap with leading AI companies.

Despite these advancements, Wang’s leadership has faced criticism, with some internal voices describing his approach as frenetic and the progress as incremental. There is skepticism among current and former employees about Meta’s ability to achieve a leading position in frontier AI under Wang’s direction. However, Meta maintains that Wang’s achievements, including building a strong research team and establishing foundations for advanced models, speak for themselves.