TECHTechnology

Meta’s Muse and the Shadow of OpenClaw

As Meta’s new AI agent, Muse, climbs the App Store charts with 600,000 daily users, the tech community is debating its origins. Critics allege the platform is little more than a polished wrapper for OpenClaw, citing identical file structures and nearly identical personality guidelines governing the agents' tone and behavior.

September 24, 2026638 reads0

Nat Friedman, head of product for Meta’s Superintelligence Labs, denies the allegations, asserting that Muse was built from scratch. He admits, however, that the product was heavily inspired by OpenClaw, noting that his team adopted specific file names and formatting because they believed creator Peter Steinberger had perfected them. While Steinberger was hired by OpenAI earlier this year, the influence of his project remains visible in the current wave of agentic software, including the high-valuation platform Instinct.

The central tension lies in whether these corporate iterations actually improve upon the open-source pioneer. OpenClaw gained massive traction for its ability to automate daily tasks through messaging apps, yet it struggled with significant security flaws, with research suggesting that 15 percent of its skill repository contained malicious instructions. Meta claims Muse addresses these risks through an isolated virtual machine, though current reports of zero-day vulnerabilities and data-sharing defaults suggest the "ground-up" security remains a work in progress. Ultimately, the shift from grassroots projects to corporate products like Muse and Instinct prioritizes mass accessibility, trading the decentralized ethos of OpenClaw for the integration and ease of use offered by Big Tech.

Comments (0)

Leave a comment

No comments yet. Be the first!