The editor of Downcodes learned that Lu Cheng, the developer of Rabbit R1, recently shared the latest progress of the AI agent. In the past six months, Rabbit R1 has undergone 16 update iterations, covering bug fixes, response speed optimization, and new feature additions. More importantly, Lu Cheng announced that Rabbit R1 will launch a new large-scale action model (LAM) on October 1, bringing users a wider range of functional expansion.
Rabbit R1 CEO Lu Cheng recently revealed that in the past six months, Rabbit R1 has undergone 16 updates, including bug fixes, shortened response time, and addition of small features. At the same time, Lu Cheng announced that Rabbit R1 will launch a new model - Large Action Model on October 1.
The R1 has gone through multiple over-the-air updates over the past six months and currently only works with seven specific services, like Uber and Spotify. Lyu admitted that early high expectations may be overly optimistic, but he emphasized that the new LAM will provide a wider range of functions, allowing users to perform tasks such as purchasing concert tickets, registering for websites, and playing online games.
The working principle of LAM is structured as steps, which interact by analyzing the content of web pages. Lyu demonstrated how it can automatically search for domain names and complete registrations, demonstrating its potential for performing simple online operations. Although LAM does not currently support user credentials, Rabbit plans to introduce hooking mechanisms in the future to protect user information at login.
Although the agent has yet to launch a "playground version", Lyu said that more improvements will be made in the future, including optimizing the user interaction experience. Currently, Rabbit has no obvious “killer app” yet, but its potential as a universal proxy opens up many possible application scenarios for users.
All in all, the continued updates of Rabbit R1 and the upcoming LAM model show its active exploration and development in the field of AI agents. Although the current functions are still in the stage of gradual improvement, its future potential cannot be underestimated. We look forward to the official release of the LAM model, which will bring users a more convenient and efficient online operation experience.