Categories

Reaction to r1 from Rabbit

The launch video of r1 from Rabbit is interesting, it also gives a somewhat ominous feeling that AI can now operate all or most software (equally ominous as all the other things that have already been said about AI).

In the pizza example, it mentions ‘the most chosen pizza,’ then only ‘confirm’ is shown (no choice, no cancel?).

Will they have to do that integration on a per-app basis? On the other hand, they claim to let the AI handle the interaction, but it’s not entirely clear.

The hardware looks funny, a bit thick, and has a large edge with a scroll wheel. If there’s a touch screen, I wonder what the scroll wheel is for.

In the example with the camera and Rick Astley, you can see that the screen is not very good, and the camera app doesn’t show Rick in the correct proportions.

The analysis of the spreadsheet with the camera is impressive, communicating via email is interesting 🙂

Teach mode is cool, something others will likely implement.

They are not app-based; instead, they are web-services based (although apps are, of course, also web-service based).

For some free web services, you might wonder how they survive without ads or donations. But, of course, it’s better if they are directly paid for.

$200 is a very low price and explains something about the hardware.

It’s not a standalone device because you need another device to train and log in.

I don’t know if people need a device to perform a subset of tasks; that model is useful for existing smartphones. If the device can easily store pieces of information or chain of thought, it could work as a personal note-taker and assistant. Now, it can be frustrating when an AI understands you to a certain extent but then goes off track.

Very interesting what they have conceived and developed, in any case.