Rabbit R1 Is Just An Android App..
Vložit
- čas přidán 2. 05. 2024
- Recorded live on twitch, GET IN
Article
www.theverge.com/2024/4/30/24...
By: Allison Johnson | www.threads.net/@allisonjo1
My Stream
/ theprimeagen
Best Way To Support Me
Become a backend engineer. Its my favorite site
boot.dev/?promo=PRIMEYT
This is also the best way to support me is to support yourself becoming a better backend engineer.
MY MAIN YT CHANNEL: Has well edited engineering videos
/ theprimeagen
Discord
/ discord
Have something for me to read or react to?: / theprimeagenreact
Kinesis Advantage 360: bit.ly/Prime-Kinesis
Hey I am sponsored by Turso, an edge database. I think they are pretty neet. Give them a try for free and if you want you can get a decent amount off (the free tier is the best (better than planetscale or any other))
turso.tech/deeznuts - Věda a technologie
It's so funny how these devices stopped working today, at the exact time that OpenAI had a server outage 🙃
Almost as like... No that cant be possible..
They use cloud AI to do stuff
@@sp123 Yeah, but it's funny that they don't even use their own solutions, instead opting for OpenAI. So you are nearly entirely are paying ridiculous prices for a really bad device.
@@raremc1620Using a pre-made model is going to be much better than to build their own on top of the rest of the things they claim to work on.
they had an outage on the chatgpt servers, not the api servers. so very haha, but irrelevant unless rabbit is scraping chatgpt for an api. ps no fan of the r1, just funny how everyone is an internet detective but doesn't know how to use the internet to investigate their beliefs 🙃
6:18 »founder was previously in NFT« - all we need to know.
And now coffeeZilla is looking into him
Juicero mentioned
I mean... Holly shit
COME TO JUICERO
big yikes, burry the R1 and AI pin
What an age we live in…
Please dont try to get profit from your hardware if your main product is a software. Rule 5865.
Today 90% AI products are products surfing the wave.
It would be good for people who can't have a phone or for kids to learn on.
@@americansoil8260Up
@@americansoil8260 thats true, also Japan loves their gadgets like the tamagotchi, also i remenber the pokewalker being kinda popular outside of Japan, but it has to have the right price, also i think it would be a niche product.
Guess what? If they made an app, they will be directly competing with a bunch of first-party voice assistants, and I bet manufacturers will find ways to block/disadvantage them.
@@andyhu9542also couldn't interact with other apps on Android like Spotify. The metas horizonOS is just Android launcher as well, but people buy it for the VR Hardware
Their demo was a fraud. I wish their CEO will serve jail time.
Pretty much every "quick to market" piece of tech is something with android these day, from my experience. Theres nothing WRONG with that, but at least make the product actually useful and not a piece of E-Waste.
I would say there is something wrong with releasing a product that would run perfectly fine on a mobile phone and pretending it's unique just because you wrap it in a casing that looks different then a mobile phone. Its built to be nothing but a full on investor scam because no investor would pay as much as they got for the Rabbit if they knew it's could of just been an android app.
It is wrong stop excusing trash
Agreed. The fact that this was an app is the most minor of the issue I have with this whole situation.
They called it powerful rabbit os
Assuming Rabbit is making API calls, the business model is to create an expensive yet cheap to produce product which limits the use of a per use API. This hopes that the user is unable to use the device enough to cost the producer more than they sold it for. Considering that a GPT subscription costs 20$ it would take around 10 months before the user uses enough credits to become a cost to the producer. They idea would then be to create the R2 so that they can get more money.
Not a business model that will scale well 😅
Credits are cheaper then monthly subscription
@@kzalesak4 yeah but it is a good assumption for how much openai assumes someone who pays for an ai product will use it for, we don't really have good data on how many api calls would be used in a month so we can't determine the credits cost, but it would probably be a bit cheaper
I seriously doubt they have one sub / user, most likely they made a proxy API mutualizing all users' calls
Oh yeah pretty good "business model" and the owners are not insane
Now is a good time to build a non-AI startup using AI tools. But it’s a horrible time to make an AI startup.
Depends how good your marketing is; some might say, if you can get millions from dumb venture capitalists then who cares how garbage your actual product is 😏
I'm working on my own app with exactly this strategy, and it's saved me thousands of hours and dollars, so yeah i can attest to that
If its a real ai startup: good! Fake chat gpt wrapper: BAD
Naw, we're in such extreme hype right now, anything with AI in the name will get millions thrown at it immediately.
Thing is, most of the future successful non-AI products using AI are the one that already exist without AI. Like Jetbrains now offering AI.
The Rabbit R1 is basically just a limited Android device and it's software is just an overengineered wrapper around OpenAI
If they manage to fool investors into giving them money, then startup is succesful.
@@user-df5ym9dv5g For the purpose of getting the quick bag? Then yes.
For the purpose of building a sustainable longterm business? I doubt it, but time will tell us.
3:37 My guess is that the so-called "firmware modifications" were merely to support the hardware inputs... that's all. There's really no other firmware needed for a app connected to the cloud. It could have been an app.
And whatever the "modifications" are, they are probably not even in-house they just slapped on something open-source because at the end its just Linux.
Not surprised in the slightest. We are far from an app-less ecosystem. Ai is not there yet, it is too dependent on data models sitting across the entire internet. I would have loved to have sat in on their meetings with investors when they were trying to get funding for this. It tells me that either the investors have no idea what they are really investing in or the startup founders are really great liars.
they speak of AI being able to be goal oriented, yet it can't use an APK like a user uses via scanning the screen doing computing vision and then clicking on it, heck, if it used some sort of testing framework automation to use stupid Apps it would already be a better product.
"it tells me that either the investors have no idea what they are really investing in or the startup founders are really great liars" its both
yeah, if I had to build one of these that's how I would have done it. my question is - how did they get their hands on the apk? lol
The CEO spent the last 3-4 years Lying to an NFT community he ammassed with lies, then lied to them about making an NFT game with AI npcs, then opensourced the unfinished game without any of the ai npc part and then renamed the company to Rabbit INC (so yes, Rabbit INC still holds the rights to the NFT's, its still an NFT company just renamed)
This thing probably tries to bypass the aosp security features (like keeping the microphone constantly on without user permission) and collect data.
A phone does this right? Always when you talk about something and you open google.. you get adds about the stuff you were talking about. Not always! But it happens
@@Gingerlion777that's simply not true at all. The code for the mic is available right there in AOSP for you to look at. Security researchers check this stuff man. It'd be a big big scandal if that was true
@@Gingerlion777 What are you talking about? You're contradicting yourself in the span of one and a half sentence.
Anyway, people had done some experiments about this (Verge, if I'm not mistaken). According to their results, it's not that they use the mic. It's when you browse websites and hover your cursor over certain links or images. It's when you're doomscrolling and you pause on certain ads or posts. The way you crawl links. Every single character you typed into websites (even ones you backspaced off). Your phone's _other_ sensors (GPS, body activity, etc). Those things are processed by _the_ single best ad-serving algorithm in the entire planet.
They don't need your mic for that.
It's funny when I joined their discord and asked why this device with a microphone, speaker, screen and SIM CARD couldn't do phone and video calls, I was called a troll and dogpiled on by a bunch of simps that were not getting the point at all. Very *woosh* moment right there.
As much as techbros try and pretend to be above the tribalism that plagues the rest of the world, they're just as bad.
We're doing new thing now and you better get on board. Don't be a MKBHD.
I think the burning question on all our minds is: Could Rust have prevented this?
Ya playing the game Rust will prevent of thinking about this crap AI products..
Well clearly theyd be still fighting borrow checker instead of shipping code
Rust was already the backbone of this project - what do you think the "R" in "R1" stands for?
Even more obvious considering rabbits are blazingly fast
My man coping so hard with that "late stage hype-cycle"
viewers accidentally stumbled upon communism, had to end it right there.
Noooo cómo se te ocurre criticar al capitalismo 😢😢😢
Twitch chat: "Should I do an NTF startup right now?"
Me: "Yes, you should most definitely do a Non Tokenable Fungus startup right now. Fungi whose cellular structures haven't yet been fully tokenized by biomedical AI models are extremely valuable to psychedelic researchers in the European market currently."
For us cultured fellows, NTF means something else entirely. But yeah. AI-designed psychedelic mushrooms sounds quite lucrative.
The way the app renders on the screen reminds me of those old Nokia times where if you downloaded a game in the wrong resolution, there would be an empty (black) area next to and under the game in the good case. In the worse case part of the game would render outside of the screen.
I don’t understand why the game couldn’t just pull the resolution information from the device?
@@MrAdamo my guess why all Nokia apps were in absolute sizes is that the OS had to be very limited in order to meet all of its requirements, so every nonessential feature was removed. That's prolly also why photos were labeled incrementally (picture00, picture01, etc) than with the timestamp.
Also in order to fit the app on screen you need a scaling algorithm to correctly fill the screen. And all of this complex stuff was pretty much pointless since the most used resolution was 320x240 anyway.
@@MrAdamo TL;DR:Because it's a very old system that had to run with extreme limitations to fit on a small device, so they removed everything that wasn't essential to use the device.
Did they really just claim they made an entire os just for this? lol what a blatant lie on the Rabbit teams part.
*makes yet another AI assistant* "innovation :D"
R1 Tamagotchi mode coming next month, you heard it here first.
That's an insult to Tomagotchi everywhere! An AI Tomagotchi would sell more units than the R1!
Rabbit founder had an NFT videogame project before this, that's all you need to know about R1.
Prime should look at how Qualcomm is marketing their new chip by saying is AI by just running AI programs lol
"AI is in its Juicero era" is a fucking cold line holy shit
His very dim minded chat sniffing "unfair" copium at the end 😂
What really bothers me is that my portfolio manager is probably falling for this crap
The cloud? The whole point of the cloud is to remove the device limitations. If they had some hardware accelerator, sure, but the cloud? I need more details
There had to have been a technical reason they wanted root or near root permissions on an android platform. I'd also like to know the reason but I don't really care all that much.
I mean it's a good way to always listen to your conversations typically android would have to you grant permissions for that.
I guess for doing inputs in other apps?
It’s unfair that Microsoft spent 70 years building up resources and talent to be able to position themselves to write foundation models but I can’t throw something together by myself in 2 months to beat them
Neural networks/fuzzy logic was just a theory until late 90's.
@@woldemunster9244 but building capital and talent to pursue new growth opportunities wasn’t.
This reminded me of that great "You wouldn't download a bear would you?" wallpaper. If we're talking bespoke AI integrated devices, it's going to be google/amazon with their google home and alexas that are going to actually work. It's most likely going to be behind a subscription since there's some associated cost, with some free period first to kill all competition. If small LLMs actually end up working fine, maybe we'll see a generation with enough local compute to make it worth it the upfront cost compared to just running everything in cloud.
I like "small Large Language Models." ;P
I'm so fucking pumped for the bubble burst.
$12 per drink... you can buy a tree fr that price 😂
In South Africa, $12 is literally 4 Starbucks lattes damn
Currently, R1 acts like a Multimodal AI that can process text, generate audio, analyze images through vision technology, and provide descriptions. This functionality is similar to the Gemini app available on the app store.
However, despite the capabilities of Large Language Models (LLMs) like this one, a key missing piece is the Large Action Model (LAM). While Rabbit R1's LAM promise to handle complex tasks like food ordering and flight booking, etc, their implementation seems to be lagging behind.
I admit I bought the audio player equivalent of this: a new sony walkman. It's a mini android device, but I like having a separate device to handle my music library.
Makes more sense to get a high quality standalone music player than a standalone Siri though
See the difference is that Walkman claims to be a portable audio player and it does that job well. So there is nothing wrong with that.
I just want a pin with a cam and mic that pairs to the phone as a shortcut for an AI assistant.
That's basically a modern smartwatch
the thing that sticks out to me is that the device itself doesn't seem to be using any AI, everything has to be done in the cloud which ANY device could do... if these devices were actually doing local processing I'd be more impressed, that's what I was expecting; considering we've already got phones perfectly capable of sending queries to the cloud to be answered by AI models... Where's the AI hardware? I've got old AI chips sitting idle due to the fact that they were built as a crypto mining rig for distributed AI cloud processing, it would be nice to be able to use those rigs as a local server to run models on, probably can't tho due to the hardware being produced before the software and what it would require.
I made a voice interactive ChatGPT thing that talks back - a plugin I designed to access OpenAI API (ChatGPT, Whisper, etc.) easily through the VoiceAttack program alongside user created voice commands for games and Windows. I'd never claim to be an AI developer, I get that this is just a deployment of existing OpenAI tech into another app, but then... Does that make me a Wrapper? 😅
This is what the Nothing CEO is always saying... Innovation in tech stopped years ago... Now is just small upgrades and no innovation...
Loved the stunlock at the end hehe
The youtuber Dave2D made a video where he theorized about why these things were released as hardware devices and not as apps, and why they are so unfinished. His theory was, that in the forseeable future, Google and Apple will announce their own LLM smartphone assistants which will be deeply integrated into their ecosystems and will therefore be much more useful than these devices. Meaning that that will render these tings completely useless. So releasing an app would have been even more useless. This way, they at least sold the hardware to some customers
That makes it even worse, tbh.
Thanks for the e-waste, guys.
So….. everything we have already
Still waiting for the "useful" part for having "AI" on phones.
I pre-ordered an R1. For me, it was a calculated bet on the possibility of its capability as an accessibility tool. The machine vision with the camera was what brought me in. My hope is that I can have it read medication directions for me or tell me which way my gate is at an airport since I can't see the sign. (I'm legally blind)
Was I duped here? Maybe, I haven't got mine yet. But $200 was low enough that the floor doesn't hurt that bad, and the ceiling is I can interact with the world better.
Hi my name is Judah.
And I was kinda wondering what exactly you were looking for in a product from a person with sight disabilities.
I have expertise in the area of machine learning and artificial intelligence. I'm super interested in your need set.
I think the way to go with the Rabbit R1 is make it very accessible, maybe through an API, so people can get into it and make their own applications for the device. Then let them share these apps in an open source way and suddenly you attract a lot of tinkerers and hackers that make the software better while the hardware is provided by teenage engineering.
Lmao what a rube
The only I guess “cool thing” that the R1 can do is that it has a LAM (large action model) which is just an LLM with additional abilities. For example you can ask it to book a trip for you and it’s supposed to go on the browser and do the clicking and what not and book a trip. But it should still just be an app.
Mini Discs are still pretty popular in Japan, well more so than you would think anyway. I've almost bought one a time or two.
I wonder if some new AI can be predictably better to be worth of training, or it will be just hit and miss and mild improvements of halucinating chat bot we already have.
Capitalism isn't to blame for the Rabbit, Capitalism is to thank for the fact it wasn't tax-payer funded
To be fair they can be using some sort of NPU under the hood to help with for example do speech to text(STT) localy, or something of the sort.
Also if said tech for 179$ bucks includes a data plan and integrates with say smart home solutions(say leveraging that good local npu assisted STT to give commands to your smart home) can be a good deal.
Never would have bought it for myself though
But the problem is that STT is already done locally on most modern smartphones, what's the point of having a separate device to do the exact same thing.
It's also going to the cloud 99% of the time whereas modern smartphone assistants like Google Assistant and Siri do a lot of stuff locally.
The Rabbit does not come with a plan, you put your own sim into it. The Humane AI pin does come with a data plan but they are also charging you a monthly fee for it and it's vendor locked. Also, you can already do smart home integrations with your phone.... just use an open source alternative like Matter or Home Assistant, maybe even something like ITTT if you don't want to go open source. These are all already solved problems and devices like R1 and Humane are just dumb alternative solutions that make absolutely no sense.
@@CHUCKNORR1S Well local STT is not that great, unless you count something like Whisper or say OpenVoice(aka mozilla) STT which are in a legue of its own and aren't really fast on phone hardware. So having something like a dedicated NPU do the heavy lifting for the top-notch fast local STT can be great for this use-case.
As for local integrations, i was implying using say openhab+rhasspy kind of thing in tandem with this thing(aka basically using it like a "batteries included" speech interface to the system), which would enable you using low-power no-gpu setup for your smarthome hub with voice assistant.
tbh I would use r1 if it could be more useful. Like understanding how to set reminders and alarms with full customisation of dates, recurrence, etc.
rabbit is a data collection scheme
reading their reseach page i believe they are doing selenium like actions when the app tries to book a travel and for that they would need system wide permission to act as a HID to read/act on behalf of the user and for that the app needs elevated rights on the phone's OS. the irony is google/apple will include this in their next generation anyways
Playwright. They generate Playwright actions using the AI and run them in Docker containers. It's clever, but also not crazy hard to replicate.
It's clever because the only money they need to pay is when using the AI once for each custom user action, and can run them multiple times without involving any AI again. Problems arise when a service updates their UI or UX :)
The supposed purpose of the R1 is to not only do the Siri/Google voice interaction but also the "what is this?" bit and learning to control stuff that you show it like a smart macro. It can supposedly bluetooth connect to devices to do that learning, but I believe it's preferable to USB connect it. That direct connection with a computer is likely the point of their bespoke design: Getting around the typical security on a phone that prevents controlling a computer. >D
The Juicero machine was built to the spec of a mechanical engineer's wet dream- see AvE's video - and they initially released them at $399, which was a definite loss. I'd love to get my hands on one and part it out. What's crazy is that it didn't work by rolling a bag, but rather apply flat pressure across the entire surface. So the analogy sucks, because Juicero as a comparison is like finding out your todo app was ran by handcrafted mechanical computers.
The money was in having a DRM-juicer for $8/pack subscription packages of questionable fruit to stick in it. Which, it turned out, were even easier to juice by simply squeezing by hand - DRM not required.
Cory Doctorow's War on General Purpose Computing talk's concepts at work here.
Google's I/O is coming, Apple too with WWDC, they will heavily focus on AI, this was the last chance for Humane and Rabbit to get novelty AI hardware out until Apple and Google have superior versions just baked in into the phone's OS with no extra charge.
Yeah. These features are better if they are built in the os rather than being an app.
Last Google I/O was heavily focused on AI. Can they add more bullshit? Of course.
I don't expect to be impressed by theirs either.
Any model has to have a way to interface it.
And a android application is a interface.
I honestly think the Large action model idea of Rabbit and the R1 could be huge, however it doesn't work yet and could still be a phone app
Did you guys notice upscaled zoom?
Looks weird and dope at the same time😅
Having their own hardware allows them to circumnavigate any rules of App Stores that they don't want to follow. It also allows them to add things like payment systems and their own App Store specific to the R1 and any future hardware platforms. They most likely don't want to be beholden to Apple and Google, especially if they are seen as competition.
App stores or payment systems don't have anything to do with hardware. They likely went with an AI-in-a-box because it would be easier to sell than an app.
Rabbit R1 was designed in collaboration with Teenage Engineering, a company with an absurdly successful history of producing overpriced devices with basic app-like functionality running on bespoke hardware that’s arguably less fit for purpose than an Android or iOS phone running an app with the same functionality - albeit more typically in the music production market. They’re exactly the company you’d go to if you want to wrap an app in a pointless hardware device and market it to an audience of people with more money than sense and a superficial fascination with the aesthetics (rather than the substance) of geek culture. For anyone familiar with their range of products, this sort of product design is no surprise coming from them.
The whole reason for buying the R1 is their advertised "teach mode" which will be released later this year, which will allow you to train the AI to use any app or website autonomously via their "Large Action Model". Even if that could technically be an app, it's an amazing proprietary technology worth paying $199 for, if it works as advertised
m$ copilot is free
"Engineers came together.." lol no they didn't. Some morons around a shiny table came together and sold this to VC before any engineers were involved.
We have to think how Jesse was NFT salesman before dropping the ball on that and going for Rabbit
Ok, so where can the app be found?
Minidisks were the shit, they combined floppy disks with CDs and make a full metal tank of a player. I wish I knew what I did with mine.
I mean this type of dedicated purpose hardware device does have a market, but at least combine it with that 3d holographic desktop display if you are going after the clipper/tomogachi crowd.
I have a reason why it's not an App: if they made it an App they would be directly competing with a bunch of first-party voice assistants. And phone manufacturers will find ways to block/disadvantage them, or even just change terms of service so people cannot develop competing apps. With custom hardware they can be sure their software is not dead on arrival.
It's made into a device, because soon if Google or Apple come up with their own AI for their ecosystems, they can restrict the growth of rabbit. So rabbit is trying to not fall into that trap by having its own device to run its AI. And the events of Google and Apple are right around the corner. I heard this explanation in some video but can't remember.
the main reason they don’t make it an app is the large action model, the thing that preforms actions like playing spotify, order uber. it needs a place to preform those actions. apple/android aren’t gonna hop on board with a ml model that just starts preforming actions like you. hence the custom device, yes it’s android, but a modified android to allow a model to hook into the ui.
Right so google will spend 2 seconds integrating that functionality in their phones and this entire product/company is dead. Also you got to remember that their "LAM" is pretty stupid and pure pr hype. It's just an OpenGPT Model which they mildy tweak functionality of to add hooks to apps. Hook to apps is something we could do at least 10 years ago with bluetooth devices and assistants like amazon echo. To claim that it an exclusive thing they developed is bs.
It really doesn't. All the app should do is take microphone input, send that to the server and wait for a response. Everything can be processed in the back, and the only involvement of the AI is to generate the Playwright actions; rest is just running it in a Docker container when the user triggers it.
even if it was just gpt wrapper, it costs $200. So at $20/month of openai subscriptions, you'd recover the cost in 10 months. This is ofc assuming they're using openai gpt4
@@phoneywheeze9959 they are using open gpt 3.5 from my understanding. Also who will actually use this thing for 10 months. finally they can always change their payment model in the future this is just the initial scam price.
That's just the same what I said right on their youtube sponsored video.
What is unfair is M$ using other peoples' (for the most part, open source) code on github to teach their LLMs. I think they were referring to that, not that M$ can teach LLMs per se.
Open source means anyone and everyone can use that resource, even those with a lot more resources capable of actually doing something with it. Part of the reason why closed source exists.
@@Leonhart_93 It's not as simple as you may think. Almost all of these come with different licenses, that posit different terms and conditions under which said code can be used and redistributed. Pretty sure M$ did not care about that one single bit.
@@nowave7 It's a possibility. But then part of having rules that work is having the ability to enforce them AND the ability to detect when they are broken.
Otherwise counting on the benevolence of the users just ensures that the machiavellian ones will come out on top.
@Leonhart_93
Not for your own closed source program.
Most open source licenses require any program using it to also be open source. Also not all public repos are open source to begin with.
AI is working in a grey area where using software as training data isn't treated the same as using it directly.
@@matt66146 Yes, because they can't detect it, which means they can't enforce it.
Juicero: An $800 CapriSun dispenser
If phones with the stupid assisstant button still exist it looks like free real estate.
I'm guessing they're looking into the possibility of replacing smart phones with AI powered devices. Kind of like a more Star Trek style HID where you mostly just talk to the computer. I'm guessing the smart phone won't be displaced until we get Vision level capabilities in the form of sunglasses
I was laughing during the presentation… such an obvious “I’m Steve Jobs look at the useless shit we are presenting like it’s WWDC”
It's a scam while legally not being a scam.
Many fundraisers just run with the money and get in trouble.
BUT, if you actually deliver a product, you legally are covered.
custom hardware would be worth it if:
- runs much faster: for example some edge inferencing hw is used (this eventually turns moot, because phones are improving at this)
- runs much longer: huge battery paired with very low spec android hw => one week uptime would be superb
- has a different form factor desired by people
Rabbit R1 has none of these.
feels good to be an Android developer right about now
Every AI startup is just mere wrapper with 95% of budget went into PR.
I love that one Twitch comment calling it "SAAH" - Software as a hardware
Honestly not surprised, from the first time it was introduced I failed to understand whats the use of that, considering all of the things it offers is already present in a phone
I think part of why corporations love AI is that pretty much only they can do AI. In the era of simple apps and servers, there is always the danger that a smaller company, or even just a single person, creates and launches a killer app based on a great idea (yes, maybe using AWS or Azure in the background, but still). Try to do that with large language models... AI is arguably furthering inequality in society is not so mild ways.
Lol the leaked apk is using one of my open source libraries.
If it runs in the cloud why does it need custom firmware or kernel level device permissions?
edit:
I'm gonna go with my cool-kid Tamagotchis, aka Digimon
yeah. if you're linking your account, apps like Spotify and uber probably have an api which would be much more easy and efficient to use rather than scraping
The logical thing I could see being done by a dedicated mobile AI device like that would be leveraging Google ML Kit to run on-device AI workloads as the primary, hardware-optimized feature. If they're doing anything more "bespoke" than that, it seems a bit silly and unnecessary. There are awesome tools emerging, and it makes sense to have new hardware for implementing it in mobile devices. Not sure exactly what they're doing under the hood, but it sounds like a reasonable concept.
What we really need is a decentralized AI on the blockchain.
Minidisc was not a smash hit, but it had a pretty good lifecycle overall.
It’s only unfair if you NEED to do it. The second the response goes “well I don’t need it to live” then it become a luxury.
Juicero hand squeeze wasn't more juice: It was %95 of the juice but it was faster (1.5 min instead of 2min)
Devin, Rabbit And what's next?? 😮
It’s the kindle paperwhite… I am hoping that’s what they are thinking when they came up with this. Why get something like the kindle when you have your phone or a tablet already? Only problem here is it’s highly inconvenient to own an R1 versus just interacting with your phone.
At least Kindles have an actual use-cases. It's lightweight, not as battery hungry and the screen is great for reading-and perfect for cases where you don't want to give your kid a smartphone for one reason or another.
This however has nothing that couldn't be an app with a few tweaks to the UX. It's cloud-based, has worse battery and hardware, and doesn't bring anything new hardware-wise.
The fact that R1 is essentially an Android app is not a problem in itself. Android was created as a system for cameras and to this day a lot of cameras use it. Sony camera software is also an Android app, you could run it on an Android phone with enough tinkering. The problem is their hardware doesn't offer anything else than an Android phone, or any other smartphone for that matter. The Juicero analogy works so well in this case, as they created an entirely superfluous piece of tech that does less than existing solutions.
Tell me you’ve never heard of Teenage engineering without telling me
Rabbit R1 is the Juicero and AI is the next bubble.
I could see this product being sold for 10% the price. It could have a market as “baby’s first AI Pokédex”
To be fair if you like the physical gadget itself, 200$ is not too crazy.
Lmao, the best summary of AI startups I've ever heard.
Everything Teenage Engineering does can be an app. Except for the desk.
someone wanted to make an AI startup with a device and it's like the engineers said, okay we will make an android device that has an app on it and everyone just went along with that. at least they brought it to market.
12:11 I would bet the overlap is high.
MiniDisc was huge in Japan but it flooped in the rest of the world
Now I want a furbie so I can teach it to scream whenever someone mentions AI
from the very start i was wondering why the hell you would design a device for this. just integrate the functionality into a regular phone. and it not having a subscription just screams data collection
The point of this is to get to the future that was in Star Trek. Just talk to a computer and it can do any task simple or large in real life or in the computer world.
Only it's such an early stage that it doesn't really have a way to interact with anything yet.
It's weird how you are ok with products being only being good in the ceos imagination. The product isn't star trek it's a crappy device built entirely to scam investors and run away. They are not innovating anything they are using OpenGPT the same thing the majority of AI companies are using. Plus they are even lazier then Humane because at least Humane has some interesting gesture and hand detection tech, this is literally old crappy android phone that only runs one app.
The first iphone wasn't revolutionary because it sold a metal box and said oh well we will have xyz features later. Was revolutionary because it already have all of it's features developed.
@@404maxnotfound you're right this product is trash.
I don't understand who would get excited about something like the R1, maybe scammers looking for a new way to scam people.