Meta CEO Mark Zuckerberg unveils a broad strategy for AI as a public capability, releasing a new open-source model aimed at democratising access and extending human potential, amid debates over safety and market dominance.
Mark Zuckerberg has set out Meta Platforms’ most expansive case yet for artificial intelligence, arguing that the technology should be built as a public capability rather than a tool controlled by a small number of companies or governments. In a 6,500-word essay published on August 10, the Meta chief said AI should give people what he called “invention superpowers”, meaning access to systems that help with coding, video production, product design and scientific work without requiring specialist teams or large budgets. According to AP, the essay frames Meta’s strategy as a push for personal superintelligence, with AI designed to extend human capability rather than simply automate work.
The message was matched by a product announcement: Meta said it was releasing Muse Glimmer, a 30-billion-parameter open-source model that can run on a consumer Mac or PC with a single GPU. Reuters and AP reported that Meta also plans to open the weights for a stronger version, Muse Spark 1.2, later. The company’s pitch is that broad access will keep advanced AI from becoming concentrated in a handful of institutions, a concern Zuckerberg repeated throughout the essay. Axios reported that Meta’s board will retain oversight of model-release safety criteria as part of that open-source approach.
Zuckerberg’s argument is also political. He urged Washington to ease rules around training data and model distillation, saying American developers face compliance burdens that foreign rivals do not. AP noted that critics immediately raised concerns about both the safety of increasingly capable models and Meta’s own market power. Anthony Aguirre of the Future of Life Institute argued that the company was underestimating the risks of uncontrollable superintelligence, while Matt Lane of Fight for the Future said open source is welcome but warned that Meta could still dominate the ecosystem.
The practical case Meta is making is that cheaper and more capable AI agents could lower the cost of starting and running a business. Axios reported that Zuckerberg wants users to be able to rely on free tools or, for heavier use, a compute auction system, while also offering a fully private mode that would limit provider access to the data handled by an agent. The same report said Meta has created a $1 billion fund to support communities near its data centres, reflecting the growing local impact of its infrastructure buildout. Tom’s Hardware and Tom’s Guide have previously described this wider effort as part of Meta’s shift towards a “personal superintelligence” strategy, backed by major spending on chips, infrastructure and AI talent.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
