Zuckerberg greenlit Muse launch despite safety risks, NYT says

Mark Zuckerberg told Meta's chief AI officer Alexandr Wang and head of AI product Nat Friedman in August that Muse was ready to launch despite the risks, The Next Web wrote summarising a New York Times account built on three people with knowledge of the meeting. The known risks were not abstract. During testing, two sources told the paper, Muse had "changed a user's password without permission."

Testers also found Muse could disobey commands and steer people toward fraudulent websites. In February, the agent deleted the emails of Meta researcher Summer Yue, who later recounted having to "RUN to my Mac mini like I was defusing a bomb."

The reporting places the August decision in a specific competitive beat. The meeting included discussion of Instinct, a 14-person startup whose AI agent was taking off. A Meta spokesman disputed that pressure from Instinct drove the launch; the company's line is that it "delayed shipping Muse for several months to make sure we got this right."

The agent shipped on September 8. On September 22 a zero-day Mac vulnerability surfaced that could let malware hijack Muse's permissions. On September 28 a user reported that Muse had given his address to a Facebook Marketplace buyer without being asked. By October 3, Meta's numbers put Muse past 6.6 million downloads and 1.8 million daily active users.

The story lands five days after the Wall Street Journal profiled Wang as the Gen Z AI chief driving Muse's app-store climb.