This new $199 AI gadget is no-subscription pocket companion

Why not make a better AI assistant and put it on my phone? If it needs to control apps you let it. Another device doesn’t help anyone.

If this runs the apps on virtual phones in the cloud for its actions, what about apps that use an in-app advertising revenue model? Those ads will be hidden from the user.

i think apps which work by well voice are things that generate audio responses ( playing songs, podcasts, weather reports, news, calling or texting humans ) and advertising already exists in those output streams anyway… so no loss there for advertisers i think

i don’t really get the utility of the device though. general computing seems to make specific devices somewhat useless except if they’re very powerful or very niche use cases… neither which seems to apply here

The “lifetime” license scheme worked just fine - until the company offering it was bought by a competitor who then discontinued the product line (because this was why they had bought it after all) and the AI running day-to-day operations concluded, logically, that the way to honour the contracts was to kill all the subscribers.


As I understand, on a voice command, Rabbit will open up the non-voice phone app that you normally use for whatever in the background, and perform a complex “AI”-powered macro to login, menu select, fill in the boxes, click, and read the result off the virtual screen, and give you a voice/screen summary.

Unless RabbitOS can run the phone apps directly, I’m guessing that it runs them on a virtual cloud phone.

If the person had used the app normally, they would have seen the ad, but in this case, the impression went to the void. Advertisers and app companies wont be happy about that.

I assume that they’ve worked out how to install apps on their Rabbit cloud, especially paid apps, that might want to lock to a specific device id. I wouldn’t like Google continually complaining that a new device was accessing my account.


Swiped from @vermes82:


Unless there’s someone involved who still hasn’t come down from the high of snorting zero-interest VC cash; I can only assume that the stable business model involves opening with an impulse purchase price and hoping that the device is sticky enough that they’ll endure a subscription added retroactively.

Or some really sleazy shenanigans with the data you’d need to expose in order to make something like this useful; though even that might struggle to pay the Jensen Huang tithe.

1 Like

It would be very interesting to know what they are up to in that regard. Just throwing together an android VM is pretty trivial(might even benefit from the increasing number of cloud outfits who are offering ARM instances cheap; vs. hoping that everything behaves as expected with the less-loved x86 Android or incurring the overhead of emulation), and the amount of state you’d need to store to at least have it only show up as a new device once isn’t huge; but for applications that care life gets more difficult: you aren’t going to satisfy something that is asking for hardware backed key storage without a cert derived from Google’s root; and more than a few applications that do chores people would be interested in demand the “Play Integrity API” checks come in clean; which is probably less impossible to fake than it looks(since, like Authenticode or WHQL signing, it is ultimately intended to be widely deployed to bottom-feeder gear produced by the lazy, incompetent, or badly rushed; so even if the design is solid there is almost certain to be at least one badly broken, but official, handset that you can make clones of).

Obviously things like video service DRM don’t matter too much, just because of what the device is and isn’t for(audio service might upset a few people; but probably still not critical); but a device intended to be able to stitch together arbitrary chore interactions for you could have a real problem if certain steps are unexpectedly impossible because of (user invisible; just try to find a list of applications that hit the Integrity API, and what they demand, without deep grubbing in XDAforums or the like…) demands of certain applications.

Makes me wonder if they either do, or hope to if they get enough of an installed base to get the time of day, plan to move as much as possible to direct collaboration with certain common use case companies(uber/lyft/etc.) and get some sort of API access from them: should be more efficient for both sides, no ad-fraud problems; but not going to happen unless they get enough traction for vendors to care; such support probably wouldn’t imply heroic engineering effort; but it absolutely would cost money to set up and support.


This topic was automatically closed after 5 days. New replies are no longer allowed.