The company also teased a "large action model" in January, which was trained on phone and web app interfaces and was supposed to navigate them autonomously to accomplish user-chosen tasks. However, this capability has not been demonstrated outside of demos, and any action claimed to be using it is indistinguishable from what an API or ordinary action scripting could accomplish. Despite the skepticism, the author remains optimistic about the potential utility of the gadget.
Key takeaways:
- Rabbit’s r1, an AI assistant gadget, has introduced a new 'beta rabbit' mode that enhances its conversational AI capabilities, particularly for complex or multi-step instructions.
- The device is now better at asking follow-up questions when it isn't sure about something, and can perform tasks like setting a travel itinerary, finding deals or recommendations for products, and sourcing book recommendations.
- Despite these updates, the article suggests that the utility of these features may be limited in real-world applications, and raises concerns about the device's ability to handle more nuanced tasks.
- The company had previously touted a 'large action model' that would allow the device to navigate phone and web app interfaces autonomously, but this capability has not yet been demonstrated outside of demos.