Generative AI is stepping into a fresh stage of “automated execution,” and Google is speeding up the integration of its in-house technology stack. On May 4, despite the discontinuation of Google’s experimental web automation project, Project Mariner, its fundamental technologies have been woven into product matrices like Gemini Agent and AI Mode. Since its debut in late 2024, Project Mariner has stood out for its capability to “act on behalf of users,” comprehending cross-web instructions and carrying out multi-step tasks. Following technological enhancements, it can now proficiently manage 10 tasks simultaneously. The cessation of the project is, in essence, a technological metamorphosis. Google has already embedded automation technologies into its ecosystem, and the Chrome browser has also demonstrated an “auto-browse” feature. Industry analysts posit that Google’s action is geared towards capitalizing on its vast user base and browser ecosystem to gain a competitive edge in the AI agent race.