SANTA MONICA — AI startup rabbit Inc. has announced the beta release of its teach mode agent system, now available to all users of its r1 device. Teach mode allows users, regardless of coding experience, to train AI agents to automate tasks across digital interfaces, starting with websites. This milestone is part of rabbit’s mission to simplify human-computer interaction through its AI-native operating system, LAM.
Teach Mode: Empowering Custom AI Agents
Teach mode enables users to demonstrate tasks to the AI agent, which can then replicate and adapt those actions autonomously. By observing user input, the agent learns to perform tasks and adjust for variations, making it a powerful tool for automating repetitive or complex workflows. Rabbit encourages experimentation during this beta phase, leveraging user feedback to refine functionality and ensure teach mode becomes increasingly intuitive and reliable.
Redefining Human-Computer Interaction
Rabbit’s innovative approach aims to eliminate the complexity of managing multiple apps and interfaces. LAM operates as an AI-native system that overlays existing software, allowing users to state their intentions while the AI executes the actions. Jesse Lyu, rabbit’s CEO and founder, likened this evolution to the shift from command-line terminals to graphical user interfaces, saying, “Teach mode makes apps invisible, turning them into a seamless, intuitive experience for users.”
Rapid Development Through User Collaboration
Following a successful alpha testing phase that generated over 400 lessons from 20 testers, rabbit accelerated the beta