CES 2024: Rabbit r1 AI Assistant Wants to Do Tasks for You

[ApplesAreSweet&Sour]

I like the idea. Why hardware? He's pitching the idea that AI can essentially replace apps. So, creating another app would defeat the idea. A cool concept. And Apple should have been down this road already. They've had Siri long enough to have gone down this road. But, as we all know, Apple doesn't usually producer first to market, just best to market.

Any of the major tech giants will run laps around rabbit as they can provide mobile device products with far better specs at much lower price than rabbit can.

Additionally, the level of integration into Android and iOS will be far stronger and make AI assisted voice assistants many times more useful than simply using a hub device, like rabbit r1, where you pull app data stored in the cloud or another device.

rabbit has the first mover advantage. But it'll be obsolete no sooner than when the next flagship Samsung, Google or Apple smartphones drop.

 

[Unregistered 4U]

Speak for yourself. This is what the future I am looking for where I (and others) don't have to sit with phones and scroll aimlessly to do simple things such as looking up a phone number, scrolling instagram/tiktok and all that other stuff. This gets the job done properly.

There is a significantly sized group of folks that have outgrown wanting to be seen and wanting to see other people being seen that deal with the current situation because “that’s what exists”. For those not entrenched in social media and games such that they need to see a screen at all times, this could be a winning product.

I’m not certain if I would use it, but ever since the iPhone, I’m not one to pooh pooh anything that’s different and that a large number of folks in the world just don’t understand how they’d ever use it.

 

[Unregistered 4U]

Be aware of your own tech bias. The things you want automated won't make this consumer logical for the mass market they are trying to appeal to.

I remember reading a story somewhere where they compared all the major digital assistants. And, they had tests that included what 90% of folks would ask a digital assistant to do (define words, provide directions, tell the time or weather) and other tests that were more complex in nature.

They found that, for what most people are asking them to do, there was little functional difference in how they responded. It was only in the more complex tests where Siri fell behind. So, one wonders, why would Apple be more focused on providing a feature that 90% of folks that pick up a phone would be delighted by? It’s a question that confounds many daily

 

[Unregistered 4U]

If the concept catches on, watch Apple steamroller it as usual - with AI Shortcuts in three years max.

I don’t know if they could. If this thing is really pulling data from the services you provide it access to and using that to build relationships that it would then use to respond to more successfully, I really don’t think Apple’s ever going to do that type of data collection analysis. They’re already avoiding a lot of features (or have shut down responses) that give them more access to users’ data than they want.

They may provide hooks for users to build shortcuts, but that’s more work for the average user than this product appears to be.

 

[tomtad]

Any of the major tech giants will run laps around rabbit as they can provide mobile device products with far better specs at much lower price than rabbit can.

Additionally, the level of integration into Android and iOS will be far stronger and make AI assisted voice assistants many times more useful than simply using a hub device, like rabbit r1, where you pull app data stored in the cloud or another device.

rabbit has the first mover advantage. But it'll be obsolete no sooner than when the next flagship Samsung, Google or Apple smartphones drop.

I agree. Then again I like rabbit, and it's nice to see some interesting hardware for a change.

 

[gjpinc]

Tech for dummies. So instead of teaching people how to do something, now they have to learn how to use this thing, so it can watch them not use their phone properly, and replicate it? If something like an iPhone is that difficult to use, then maybe it needs to have a user manual in each box. Remember user manuals? I wonder if rabbit comes with one.

It goes a bit beyond that. My parents (In their 80s) are not and have never been tech savvy. They barely make calls on their cell phones. This would be a great thing for them so they dont have to interact with separate apps for everything. If this thing is able to call an uber, order food, order instacart and make a reservation at a restaurant using natural language, its worth the $200 and the monthly cell bill because it provides some freedom for them and me...

 

[gjpinc]

How do you authenticate on the Rabbit hardware itself? Or can others (your kids, strangers, thieves, spouse, etc) just pick up someone else's Rabbit and purchase items for them?

Also, what is unique about the Rabbit hardware? I doubt the hardware is their advancement, as hardware is extremely hard. Why not use the amazing processors and components of the latest Android or iPhone and add a Rabbit app to do the fancy taking action on your behalf part?

Also, how is $200 sustainable for the company? AI is pretty expensive to run, hardware isn't cheap to build... my guess is it will go subscription based within a year or two of coming out.

Watch the keynote...it explains all of it...

 

[seggy]

I don’t know if they could. If this thing is really pulling data from the services you provide it access to and using that to build relationships that it would then use to respond to more successfully, I really don’t think Apple’s ever going to do that type of data collection analysis. They’re already avoiding a lot of features (or have shut down responses) that give them more access to users’ data than they want.

They may provide hooks for users to build shortcuts, but that’s more work for the average user than this product appears to be.

The moment it suits them, marketing will reframe the argument. There's no way Apple Vision and its associated apps for example will even work well / be able to be developed to be more transparent in use without a ton of telemetry going back to or thru Apple, let alone LLMs. And you can bet Apple will not be left behind in that market especially once it is pervasive(ly invasive?).

 

[5232152]

Do you like the R1 for the simple interface and form factor? Would you use this instead of a smartphone?

Instead of a smartphone? Not yet. But I feel personally quite confident that we are getting closer to an phone interface that works like the movie Her, rather than an iPhone/Android. The idea that you have to use apps, keyboard and opinion your search to get the most basic things done seems illogical with what we have seem ChatGPT do the last year.
I would go so far to bed that whatever Apple is doing next (Maybe not iOS 18, but probably 19) will be less manual work and more R1 because it goes faster and is easier. Hell, seeing how much parents and grandparents struggle with basic smartphone use it would be much better for them to have a digital assistance to communicate with.

 

[swm]

oh, yet another solution looking for a problem.

 

[Boeingfan]

It goes a bit beyond that. My parents (In their 80s) are not and have never been tech savvy. They barely make calls on their cell phones. This would be a great thing for them so they dont have to interact with separate apps for everything. If this thing is able to call an uber, order food, order instacart and make a reservation at a restaurant using natural language, its worth the $200 and the monthly cell bill because it provides some freedom for them and me...

If a simple telephone is complicated then another device won't necessarily help. Think of the situations where they voice a request and it gets it wrong, how will they understand how to stop the action taking place? Throw it into a bucket of water? My point was that some people like to be shown things, taught, to feel comfortable, build confidence, not have them try to work out yet another machine which is going to try to work out what they are asking for. Perhaps the tech is not the issue, perhaps it's that they simply don't want to use tech. If a company doesn't allow them to dial a call to book a table or a car, then they don't use that service. Companies that don't have this as an option are knowingly making life difficult for some of their clientele, and they are happy to accept that risk, and some elderly people are happy to accept that those companies are not for them.

 

[TracesOfArsenic]

Why do companies always give same lame tasks as examples? Order a pizza, hail a rideshare.. yawn.....

That's like 0.001% of things I would like automated.

Examples then? What would you like it do for you? I love the idea of this device!

 

[Unregistered 4U]

There's no way Apple Vision and its associated apps for example will even work well / be able to be developed to be more transparent in use without a ton of telemetry going back to or thru Apple, let alone LLMs. And you can bet Apple will not be left behind in that market especially once it is pervasive(ly invasive?).

One of the apps that will be native on release is “What the Golf” and other Apple Arcade games. I see nothing that indicates those would not work well without a ton of telemetry. Watching videos or looking at pictures won’t require a ton of telemetry either. And, speaking of LLM, Apple last year released an open source project that will make LLM possible on client devices like the VP, iPad and iPhone.

Apple quietly released an open source multimodal LLM in October

Apple is belatedly being celebrated on social media for releasing Ferret, an open source multimodal LLM, in October 2023.

venturebeat.com

What’s most likely is that they will continue to try coming up with ways to do things locally like this.

 

[Stuartglossop]

AI startup Rabbit today announced the launch of the Rabbit r1, a mobile device that's designed to simplify the experience of using a smartphone. The r1 runs rabbit OS, an operating system powered by a Large Action Model (LAM).

The Large Action Model is able to learn how humans use computers, or more specifically, how r1 users take advantage of apps. Rabbit says that the r1 can understand complex user intentions, operate user interfaces, and perform actions on behalf of the user. The OS was trained on data from people interacting with apps, allowing the r1 to do tasks like ordering pizza, booking ride services, playing music, communicating with people, and more, through voice commands.

https://twitter.com/x/status/1744781083831574824

The Rabbit r1 can be interacted with using a push-to-talk button on the side, and it understands natural language requests. There are no apps on the r1, and tasks are performed over the internet.

Designed in partnership with Teenage Engineering, the r1 has a 2.88-inch touch screen, a camera that rotates for capturing photos and videos and taking video calls, a scroll wheel for navigation purposes, and a USB-C port for charging. There is a 1,000mAh battery, and battery life depends on usage. While there is no subscription associated with the Rabbit r1, it does need a SIM card for cellular data.

Users can teach the r1 to perform tasks, including multi-step tasks such as researching a location, booking a hotel, securing restaurant recommendations, and finding activities. Once the r1 has learned a task, it can repeat the action independently. It supposedly can even interact with apps like Photoshop, performing multi-step actions within the software.

The Rabbit r1 needs to be provided with a user's logins for various services in order to complete tasks involving those services.

The Rabbit r1 can be pre-ordered from the Rabbit website for $200. It is expected to ship to customers in March or April of 2024

Article Link: CES 2024: Rabbit r1 AI Assistant Wants to Do Tasks for You

This is a bargain! Great hardware, no subscription. Put a cheap SIM in it and off you go!

 

[beanbaguk]

After watching through the insanely boring presentation, I've come to one conclusion. This will be an abject failure.

Why you might ask?

• Who wants to spend their time talking to a device to do everything? I get it at home, but there is no privacy in this when out and about.
• Noise. He's dictating in a quiet room with controlled scenarios. Try real-world scenarios and watch it fail.
• That fridge is unlike any fridge I've ever seen. That wouldn't work in the real world, ever!
• I don't want a machine controlling my choices. That's why we like to do things ourselves. I don't want it choosing my pizza, deciding on the cab I'm going to get, (it's ordering an Uber, but where is the collection point, for example?

There are so many caveats and issues with this device, I could go on and on. It's a gimmick. Nothing more, and what will kill it right out of the box will be mature AI embedded into Google, Alexa and Siri.

 

[beanbaguk]

Speak for yourself. This is what the future I am looking for where I (and others) don't have to sit with phones and scroll aimlessly to do simple things such as looking up a phone number, scrolling instagram/tiktok and all that other stuff. This gets the job done properly.

Here's the caveat. You will most likely still need a smartphone to set it up. And a smartphone will, I believe, this year do everything this device does, and do it better.

Moreover, Apple will never charge for Siri. Even with advanced AI. This is how they keep people in their ecosystem. Rabbit simply won't be able to afford to offer you AI services for free. It's not sustainable. So they will sell your data onto 3rd parties, or charge a subscription in a year or two (if it still exists).

It also has what appears to be a shoddy camera, a tiny screen, and activities like scrolling through mail, accessing calendars, messaging and more will be an absolute pain on a device like this. So most people will end up carrying both, which again is pointless.

This will fail.

 

[5232152]

Here's the caveat. You will most likely still need a smartphone to set it up. And a smartphone will, I believe, this year do everything this device does, and do it better.

Moreover, Apple will never charge for Siri. Even with advanced AI. This is how they keep people in their ecosystem. Rabbit simply won't be able to afford to offer you AI services for free. It's not sustainable. So they will sell your data onto 3rd parties, or charge a subscription in a year or two (if it still exists).

It also has what appears to be a shoddy camera, a tiny screen, and activities like scrolling through mail, accessing calendars, messaging and more will be an absolute pain on a device like this. So most people will end up carrying both, which again is pointless.

This will fail.

I have worked in tech long enough to not trust critics that does not take the time to verify their own claims.

You need a smart phone to set it up
I cannot find a single piece of evidence to this claim. I doubt you can either. Feel free to do so.

Rabbit simply won't be able to afford to offer you AI services for free
Where is the evidence? I make this task simpler for you. Show ANY kind of evidence that Moore's Law does not apply for LLM/LAM specific hardware.

So they will sell your data onto 3rd parties, or charge a subscription in a year or two (if it still exists)
Again, where is the evidence? What will stop them for charging for more advanced feature later? What will make that unfair? You pay for premium, even in the Apple sphere. Or what about charging money for affiliate when people sign up for some of the pre-configured services? Your assumptions are borderline lying to yourself and others.

It also has what appears to be a shoddy camera, a tiny screen, and activities like scrolling through mail, accessing calendars, messaging and more will be an absolute pain on a device like this.
You are mad a $199 first generation device? You cannot know a lot of about hardware development if this somehow surprises you.

So most people will end up carrying both, which again is pointless.
As my original post: Speak for yourself.
This is a device that does not have any piers and will find its own place over time. If you think nobody will want in on the less-device-interface journey, I think you will be surprsied.

This will fail.
Your fear of the future is funny to me. You will not have a great time the next couple of years as the AI era is upon us.

 

[canyonblue737]

AI startup Rabbit today announced the launch of the Rabbit r1, a mobile device that's designed to simplify the experience of using a smartphone. The r1 runs rabbit OS, an operating system powered by a Large Action Model (LAM).

The Large Action Model is able to learn how humans use computers, or more specifically, how r1 users take advantage of apps. Rabbit says that the r1 can understand complex user intentions, operate user interfaces, and perform actions on behalf of the user. The OS was trained on data from people interacting with apps, allowing the r1 to do tasks like ordering pizza, booking ride services, playing music, communicating with people, and more, through voice commands.

https://twitter.com/x/status/1744781083831574824

The Rabbit r1 can be interacted with using a push-to-talk button on the side, and it understands natural language requests. There are no apps on the r1, and tasks are performed over the internet.

Designed in partnership with Teenage Engineering, the r1 has a 2.88-inch touch screen, a camera that rotates for capturing photos and videos and taking video calls, a scroll wheel for navigation purposes, and a USB-C port for charging. There is a 1,000mAh battery, and battery life depends on usage. While there is no subscription associated with the Rabbit r1, it does need a SIM card for cellular data.

Users can teach the r1 to perform tasks, including multi-step tasks such as researching a location, booking a hotel, securing restaurant recommendations, and finding activities. Once the r1 has learned a task, it can repeat the action independently. It supposedly can even interact with apps like Photoshop, performing multi-step actions within the software.

The Rabbit r1 needs to be provided with a user's logins for various services in order to complete tasks involving those services.

The Rabbit r1 can be pre-ordered from the Rabbit website for $200. It is expected to ship to customers in March or April of 2024

Article Link: CES 2024: Rabbit r1 AI Assistant Wants to Do Tasks for You

it's interesting for sure (and priced decently) but this tech is what is inevitably being built directly into iOS and other smartphone OS systems as soon as this year and will IMHO immediately replace the need for a device like this.

 

[beanbaguk]

I have worked in tech long enough to not trust critics that does not take the time to verify their own claims.

You need a smart phone to set it up
I cannot find a single piece of evidence to this claim. I doubt you can either. Feel free to do so.

Rabbit simply won't be able to afford to offer you AI services for free
Where is the evidence? I make this task simpler for you. Show ANY kind of evidence that Moore's Law does not apply for LLM/LAM specific hardware.

So they will sell your data onto 3rd parties, or charge a subscription in a year or two (if it still exists)
Again, where is the evidence? What will stop them for charging for more advanced feature later? What will make that unfair? You pay for premium, even in the Apple sphere. Or what about charging money for affiliate when people sign up for some of the pre-configured services? Your assumptions are borderline lying to yourself and others.

It also has what appears to be a shoddy camera, a tiny screen, and activities like scrolling through mail, accessing calendars, messaging and more will be an absolute pain on a device like this.
You are mad a $199 first generation device? You cannot know a lot of about hardware development if this somehow surprises you.

So most people will end up carrying both, which again is pointless.
As my original post: Speak for yourself.
This is a device that does not have any piers and will find its own place over time. If you think nobody will want in on the less-device-interface journey, I think you will be surprsied.

This will fail.
Your fear of the future is funny to me. You will not have a great time the next couple of years as the AI era is upon us.

OK, some answers:

How do you propose setting up all your services on the device through voice? Are you seriously thinking dictating a password is going to work? As an example, here is a typical password that I use:

k.Ch6!iYhtg2QfEL*XeW2_2sfE_FdGT.

Good luck dictating that correctly! The device will have to pair with a mobile device or another system to get it set-up. I find that quite ironic.

The second and third points you mention astounds me, (and not positively). Companies exist to make money. This isn't a charity, and both the computing power and cost of running an AI model is massive. They need to recoup costs, and a single, one-off hardware purchase of $199 won't offer them anything. Aside from recouping their development costs, there is the actual cost of the device, so you are probably looking at about, I guess, $50 to build each unit. Then add R&D (about $50 per unit, depending on how many they sell). Then add marketing costs, $25, and you are left with about $75 gross profits. You then need to pay employees, pay tax, etc, etc, etc....and then from profits, you need to subsidise your AI platform.

Simply put, this isn't going to pay for jack. So they will either sell your data, or eventually charge a subscription. It's not exactly rocket science.

On the fourth point; My argument is, why would I want to replace my iPhone camera with a device that takes grainy, low-resolution photos? You can see even on the keynote, the quality was dreadful! I don't care if it's 1st gen or not. It's bad quality.

Fifth point, I say "most people". Yes, there are always a small handful of hardcore techies who want to try all sorts of hardware out. I've been in this game long enough to know this, and at one point in my life, I was one of those. I come from an era where I connected online using BBS's, and I would have to balance my pulse dial phone over a receiver to connect at 2400 baud and wait for the menus and information to load up on my Amstrad 1640. I battled with my Palm Pilot, and even a Palm Top PDA running Windows CE. They all failed, even before devices like the iPhone eventually arrived. But I was one of those hardcore techies at one point.

And on the last point, you put words I never said into my mouth.

I said this device would fail. I didn't say AI would fail. AI is the future, and I am personally embracing that.