谁在“夺舍”你的Mac？

Core Insights - OpenClaw, a newly launched AI project, has gained significant attention and endorsements within a week, raising concerns about security and privacy as it requires deep system access to function effectively [1][2][5] Group 1: Project Overview - OpenClaw is marketed as a "digital assistant" that can manage tasks through various communication platforms, but it poses serious security risks by requiring extensive permissions [3][5] - The project was founded by Peter Steinberger, who previously sold a company for $119 million, indicating a blend of elite background and potential security oversight [5] Group 2: Security Concerns - A significant percentage (22%) of employees have granted OpenClaw access without IT audits, exposing sensitive data to potential threats [5] - OpenClaw's underlying technology, the MCP (Model Context Protocol), has a flawed trust model that could allow hackers to take control through seemingly benign interactions, such as reading emails [7][8] Group 3: Economic Implications - OpenClaw, despite being marketed as open-source and free, incurs high operational costs due to API usage, leading to what is termed a "computational tax" [9] - The cost of running OpenClaw can exceed $30 per month, raising questions about its value proposition as a productivity tool [9] Group 4: Industry Context - The rapid rise and fall of OpenClaw highlight vulnerabilities in the AI ecosystem, where projects can be easily disrupted by larger corporations through legal means [10][11] - Gartner predicts that by the end of 2026, 40% of enterprise applications will integrate AI, emphasizing the need for robust security measures in the face of rapid technological adoption [11][12]