Rabbit’s r1, the AI assistant gadget that has seen its initial excitement wane since its CES debut, has some updates to announce — though they might not be enough to sway its critics just yet.
The new “beta rabbit” mode enhances the device’s conversational AI capabilities, especially for handling more complex or multi-step instructions. It should also improve its ability to ask follow-up questions when it encounters uncertainty.
As an example, they offer something like the following:
“beta rabbit, can you suggest three books similar to ‘the power of now’, include page length, year of release, and ratings, and save that as a note titled ‘reading list’. also include pictures of the authors”
[followed by:] “beta rabbit, can you also get me summaries for those three books?”
Creating a travel itinerary and finding deals or product recommendations are other suggested uses. As occasional chatbot users, we’ve found these tasks impressive in demos but rarely practical in reality.
For instance, chatbot-generated itineraries can be odd and unpredictable, and comparing specs and prices demonstrates web-scraping skills but is ultimately inconvenient on such a small device. And who would trust book recommendations sourced so haphazardly?
There are also some improved alarm and timer features (you can see all the new updates here), which are welcome but sometimes fall into the “wait a second…” category. For example, “Set a timer for baking chocolate chip cookies”: At what temperature? How many? What kind of cookies? This could lead to a culinary disaster. However, it would be reasonable to ask, “How long should two dozen chocolate chip cookies bake at 300 degrees?”
Of course, everyone is eagerly awaiting the much-hyped yet elusive “large action model” that the company promoted back in January. The pitch — which seemed aspirational but not unrealistic at the time — was that the model was trained on phone and web app interfaces and could navigate them autonomously to complete user-chosen tasks. So far, this capability has not been demonstrated outside of demos, or if it has, it’s indistinguishable from what an API or standard action scripting could achieve.
I remain optimistic about the eventual utility of this quirky little gadget, which is why, despite having had almost no reason to use it since we received it for review, I haven’t relegated it to a drawer — yet.
I’ve asked Rabbit about when we can expect news on the LAM and will update this post if I hear back.