Meta Disbands Protein Folding Team, Pivots to Commercial AI

August 8, 2023 – In a surprising turn of events, Meta, the social media giant formerly known as Facebook, has recently disbanded a team of scientists known as ESMFold, who had been utilizing artificial intelligence to construct an extensive database containing over 600 million protein structures. This decision indicates Meta’s shift from pure scientific endeavors towards the development of revenue-generating artificial intelligence products.

ESMFold, a team once praised by experts in pharmaceutical development and novel therapies, had developed a large-scale language model capable of processing massive biological datasets to predict protein structures. However, as part of a broader company-wide restructuring, three insiders disclosed that ESMFold was dissolved earlier this spring.

Earlier reports had highlighted Meta CEO Mark Zuckerberg’s emphasis on this year being the “Year of Efficiency” for the company, marked by restructuring and workforce reductions aimed at boosting profitability and growth. Despite being an early investor in the AI field, Meta has lagged behind competitors like OpenAI, Microsoft, and Google in the realm of generative AI development.

As a part of its new direction, Meta has established a generative AI team led by their Product Head, Chris Cox, which includes former members of the company’s foundational AI research lab, Fair. Meta plans to launch a series of chatbots resembling distinct personas under Cox’s leadership, scheduled for release in September. While Meta remains committed to exploratory research through Fair, the focus has now shifted to integrating research findings into tangible products.

The ESMFold project had previously created an open-source database that facilitated scientists’ access to specific protein structure data pertinent to their work. However, concerns have arisen regarding whether Meta will continue to bear the costs of maintaining the database and running the ESM algorithm service for new protein sequences. This uncertainty has led to questions about the research community’s long-term access to these resources, reflecting the importance of sustainability and consistency of resources for ongoing scientific progress.

Leave a Reply