Meta has released Glimmer, an open-weight AI model that people can download and run on their own hardware, alongside Mark Zuckerberg’s argument that AI should be available to everyone rather than concentrated among a small number of laboratories. The article centres on a TechCrunch Equity podcast discussion questioning how far that claim holds, given Meta’s more capable Muse Spark model remains accessible only through its APIs.

The podcast examines Zuckerberg’s 6,500-word manifesto and the differing access models for Meta’s AI systems. It also covers wider AI-industry issues, including the energy demands of data centres and a failed $250 million acquisition, placing Meta’s release within broader concerns about control, cost and infrastructure.