Meta Platforms CEO Mark Zuckerberg has advocated for reduced U.S. restrictions on open-source AI to enhance competition with Chinese counterparts. The company unveiled Muse Glimmer, a smaller and customizable AI model suitable for running tasks on standard computers with a single graphics card. Zuckerberg hinted at the imminent release of even larger models in a video post accompanying his essay promoting the democratization of AI.
The endorsement of open-weight AI by Zuckerberg follows concerns over escalating AI costs and recent cybersecurity incidents involving models from various companies. Open-weight models, unlike closed models, allow users to customize core components and are more cost-effective. Notably, Chinese open-weight models like Kimi K3 have gained attention for their performance.
Meta, alongside other tech firms, expressed support for open-weight AI in an open letter, as Chinese startups lead in this area. In a bid to regain ground in the AI race, Meta plans to release Muse Spark 1.2, its advanced model developed by a dedicated superintelligence team. Zuckerberg also announced a $1 billion fund to address community concerns stemming from data-center expansions, a contentious issue in the U.S.
The CEO emphasized the need for U.S. policy adjustments to empower domestic firms in open-weight AI, highlighting the importance of data utilization and distillation methods. Meta intends to establish a governance structure enabling independent directors to approve model releases based on safety criteria. Despite recent discussions indicating that the U.S. will not subject open-weight AI models to voluntary safety tests, Zuckerberg stressed the necessity of reducing constraints on American labs to foster innovation.