201–210 of 309 posts
Followed by a demo of someone copying the APK and the whole thing more or less working... I think Lyu forgot that the statement was for androidauthority.com (where people who understand android hang out) and not for his 80-year-old-uncle...
I haven't heard anything about the Rabbit R1, what is it supposed to be? Based on the video it looks like this is just Google Assistant except not by Google, what makes it different from a normal phone?
google wont allow alternative voice assistants on android, so they had little choice.
"Assistant App" is a setting in the settings app where any assistant can be enabled. API is documented in the SDK docs
Running android is actually a good decision. We should celebrate it rather than scold Rabbit's founders. Creating OS is hard. Selling smart and intriguing devices for $200 is also hard. They had to pick one.
But now we can see that it is just a weird shaped Android phone hard-coded to only run one app. It shatters the illusion that the R1 hardware was worth the $200 price tag, as the software under-the-hood could be equally if not more functional if it were just distributed as an app instead of as a piece of hardware.
It would be different if the hardware had special sensors and processors that help maximize its functionality. There would be no issue for AOSP to be the base OS - in fact I would agree with your assessment that it was a good decision that should e celebrated. The hardware itself having novel capabilities would be the important aspect, not the OS they chose to build upon.
Absent any special hardware functionality, it throws the entire Rabbit business model into question. They aren't charging licensing fees, and the service itself seems to be a lifetime subscription, so they aren't planning on making money selling software or SaaS services. The profit-making part of this business is selling hardware - and the hardware is just a worse version of the phone you already have...
The software also seems barely useful. Pretty much just a bare-bones implementation of speech-to-text, GPT via REST API, and text-to-speech, with a handful of basic integrations.
So once they've sold you the hardware, do they have any reason at all to improve the software or the service's offerings? If the R1 hardware can't do anything your phone can't, can they really compete with an app (even a paid app) that does the same thing on your almost certainly higher-spec'd smartphone?
Given the founders' history of grifting, I am going to guess that the R1 hardware costs Rabbit <$50, and that the company will soon disappear, taking down any expensive cloud-based AI functionality with them, as soon as they feel they've sold enough units to line their pockets with the profits, and the R1 owners will be left with a brick.
Earlier quoted context omitted.
Their product doesn’t make any sense besides as an app. No one is going to carry a separate device around when your smartphone can do everything it can better. The chatGPT app on the iPhone already does everything the rabbit does better. They offer nothing.
The ChatGPT app requires a subscription, and offers no "Large Action Model" behavior (i.e. API tie-in to perform actions on other apps like Spotify/Doordash/Uber/Midjourney). The former is a nice bonus (10 months of ChatGPT $20 sub = one time purchase price of Rabbit R1), the latter has the potential in its fully realized form to make Rabbit infinitely more useful as an assistant.
Earlier quoted context omitted.
why are you so incredulous that android might have some annoying privacy restrictions that a custom AOSP can sidestep? I would google rabbit's reasoning for this but I don't care enough
> why are you so incredulous that android might have some annoying privacy restrictions that a custom AOSP can sidestep? I don't doubt that such restrictions exist. But I'm also curious as to what, specifically, they would be? Apps can access all sensors, cameras, microphone, network. So what's left?
There's a good number that might be useful for the R1.
Earlier quoted context omitted.
I doubt anyone outside of a handful of HN commenters knows Teenage Engineering.
Anyone interested in synthesizers knows TE.
I saw this news and my reaction was, "Of course it is, doing anything else would be stupid." But that on its own doesn't mean it could just be a phone app if the hardware adds something useful to the way the user interacts with it. For example, old Sony mirrorless cameras just ran Android under the hood but the specific hardware is what made those products, not the OS. As a hardware person, both the Rabbit R1 and the…
Probably. I think we've hit a plateau with user interfaces for most connected gear. Sticking a physical scroll button on what is essentially a phone buys me nothing but annoyance. Phones just kind of do what we want and have an acceptable enough interface that the alternatives get in our way.
The Humane Pin added the laser display but ... nobody wants it? This is starting to feel like sticking a spoiler on a phone and pretending it adds value. Maybe VCs are impressed by this stuff, I don't know.
If there were a magic version of this it would have the model and processing onboard. That's obviously extremely cost prohibitive right now, but that's what it's going to take to create a real AI-powered assistant:
* I need it to work without network latency.
* I need multiple forms of input/output depending on the context.
* I don't need Teenage Engineering's usual form-over-function design. I need the form to be out of my way most of the time.
It's great that people are experimenting in this space. It's less great that people are getting multiple millions in funding and selling a phone and web service as an "AI device." This will hurt future development.
Earlier quoted context omitted.
They already have. Apple calls theirs "siri". The dumb part here is a company getting attention for making a virtual assistant just because they added "AI".
Siri is no longer seen as an AI assistant, at least not with the post-chatgpt meaning of the word 'AI'.