Zuckerberg publishes manifesto on the future of AI: monopoly is the main threat

Mark Zuckerberg, founder of Meta, published a manifesto on the future of artificial intelligence, as reported by The Wall Street Journal. He stated that he considers the main threat to AI development to be not the technology itself, but its monopolization by one company or state. The publication of the manifesto coincided with the presentation of a new open-source AI model, Muse Glimmer. In the manifesto, Zuckerberg noted that many fears about AI are exaggerated and do not account for its potential benefits. He believes developers should engage with governments early in model creation, but strict regulation just before product release could slow technological progress. Zuckerberg also reiterated support for distillation, a method of training smaller models using more powerful systems, and announced plans for a $1 billion fund to support regions near large data centers. On the same day, Meta presented Muse Glimmer, an open-source model whose parameters are available to developers and can run directly on user devices. According to The Wall Street Journal, Muse Glimmer is less capable than the most powerful modern models, but Zuckerberg considers the open approach important to keep access to the technology as broad as possible.