Mark Zuckerberg says “slowing the pace of model innovation” would hurt AI progress

Meta CEO Mark Zuckerberg published a lengthy essay alongside the release of Meta’s newest open-source AI model, Muse Glimmer, defending broad access to AI technology. He called for cooperation with U.S. officials, including giving them early, structured access to models still in training, while warning against regulatory choices that would set back development or keep new systems from reaching users. He also pushed back on dire warnings about AI, saying he could not understand why people expecting widespread job loss would race to build it. The essay enters a dispute with Anthropic CEO Dario Amodei, who has voiced concern that authoritarian governments could exploit open-source models, and follows Senator Bernie Sanders’ warning that Congress may act if companies do not voluntarily slow development.