After Meta recently released its personal AI assistant Muse, Andrew Tulloch, a core developer, announced his departure and will join Anthropic's reasoning and performance team next week. Tulloch had an 11-year career at Meta before moving to OpenAI, where he participated in training models such as GPT-4o and co-founded Thinking Machines Lab. Last year, Meta attempted to lure him back with a compensation package worth $1 billion to $1.5 billion. He initially declined but returned to Meta in October 2025, only to serve for approximately 11 months this time. Previously, Yu Jiahui, another highly compensated hire by Meta, also left shortly after the release of Muse Spark 1.2. Currently, the frequent movement of top AI talent is not unique to Meta, as such individuals prioritize the future direction of AI when choosing platforms. It is expected that this talent mobility will continue until top talent is no longer scarce and companies possess stable and replicable engineering capabilities.