Is This USELESS Gadget ACTUALLY The Future?
HTML-код
- Опубликовано: 15 сен 2024
- Learn EXACTLY how I make my videos:
www.enricotart...
--------
📮 Behind the scenes and nuggets on my free newsletter:
www.enricotart...
💻 My gear and tech setup:
www.enricotart...
🎆 Instagram:
/ enritarta
--------
The Rabbit R1 is here and it's useless ..but genius. Rabbit is showing us that action models and agents can be made mainstream and a different approach to the traditional integrations of AI and APIs. 2024 is going to be an interesting year!
--------
🖖 Hey! I'm Enrico and on this channel I go behind the scenes of the design, psychology and stories behind tech and making stuff on the internet. I'm a tech Product Manager, builder of things made of pixels.
Learn how I ACTUALLY made my most successful videos with hands-on, practical behind the scenes breakdowns:
www.enricotartarotti.com/storybehind?
that didn't age well.
The "genius" in the Rabbit device is their ability to scam so many people with an useless android APP inside a nice orange cage.
😂
2024: Hi Rabbit, can you modify After Effect’s code from the ground up so playback can finally be smooth? Once that’s done, send the code to Adobe and send them an invoice for 200K. Then proceed with booking my next vacation.
I just don’t think i’d want an ai to do tasks like booking planes, shopping online or making appointments. I want to be in charge because there are a lot of micro decisions that I will want to make along the way.
Welcome to the world of shallow minded thinking of the AI train hype
just like online payments first impression
Not to mention it can just be an app and people don’t want to carry another device.
Haha yeah sentiments exactly, hotel location/price, plane ticket price and travel fares.
Yeah that's to be cringed upon all the way down but I am happy that I won't be so frustrated modifying my Minecraft mod packs
..and yes they achieved donation income based off basically false advertising. People got it in hand and were like “wtf 🤨..this is garbage!”
You completely dropped the ball with this, just delist the video already...
0:44 "The best solution would be cars that transport a lot of people"
So... trains. Trains are the ideal car.
well, trains use tracks not tyres, so more like busses
yeah!
I see what you did there and get the reference. Definitely on point.
Ever heard of busses?
@@spaceowl9246 its what bee's do
Giving AI full control of technology... Isn't that how sci-fi horror stories start? 😅
Good thing that AI has no will
@@Marquis-Sade But it makes decisions, and can hallucinate in terrible ways lol. It can randomly become a menace
@@adamgreenhill110 It cant really decide anything.
@@Marquis-Sade Yes it can, that's the whole point of AI (instead of hard-coded algorithms). That it makes a decision (or prediction) based on its training data. Sometimes that leads to terrible hallucinations. If we're using AI to control our technology, what happens if it hallucinates and decides to be a villain? Google Bard, ChatGPT and Bing AI have all done that before
@@adamgreenhill110 But all decisions are set with fixed parameters. AI can't give you "bad" answers, can't talk about sex, can't talk about anything illegal, etc.
I think most of the AI expert would disagree with you on "we cracked reasoning in 2023"
At 1:50 you claim that "in 2023 we have cracked reasoning" which is a ridiculous statement. Lets not forget all that an LLM is doing is predicting the next string given a prefix string. Hallucination is a massive concern, not to mention privacy, which you did not address enough in this video in my opinion. I am also not sold on the LAM claims, I have not seen anywhere any proper technical deep explanations from their side of how that truly works and how is it really reliable. Also, how are they providing local competent LLM/ "LAM" in that silly hardware box? Otherwise, is it a cloud API they are pinging? If so, how are they asking for one time payments? Not only useless, but the product they are providing is a bomb waiting to explode, as I am expecting to massively under-deliver on its hyped claims. I would also like to add that, fundamentally speaking, sometimes doing a task is just faster than describing it in natural language, which is an ambiguous, context sensitive grammar.
You can't run a proper LLM without hardware the size of a cabinet and a couple of kilowatts of electricity.
The AI hype is just coming from all those people who bought all those GPUs to do all the stupid crypto nonsense and now want desperately a return on their investment
What, did you expect a PM to understand a technology thoroughly? Lol
@@personzorzi see your point, however you have to consider the ram of the R1 is pretty low.
@@personzorzYou aren't very smart are you ?
the guys clueless , he made it clear he doesnt undestand , pob just sellling crap somewhere
The future of machine learning (aka, what all those companies are making with the NVIDIA AI chips they bought last year) is an oversaturated "virtual assistant" market full of startups crowding each other out while vying to become the next Google.
How does this relate to rabbit's strategy?
@@1anre I'm really just talking about "The Future of AI" overall. There's going to be a ton of startups grabbing for the same brass ring, and eventually they'll all get gobbled up by whichever company actually comes up with a compelling and useful enough product (and has the timing) to become the next big thing. If centuries of capitalism has taught us anything, it always trends toward monopolies.
@@ThisSteveGuyRabbit R1 has to fail so a better version of the device will succeed.
@@1anre rabbit is just another one of them
@@sp123 Did it fail?
Rabbit ripped me off. I ordered in time for batch #2, and now their FAQ page says non-US/Canada customers will get theirs first, and I may be waiting a year hor more ?). Let’s be clear - THAT IS FRAUDULENT. It is disgusting abuse of people like me who supported Rabbit right from the beginning.
Furthermore, their Discord server is a dead link. And I haven’t seen a single review by an actual user, as opposed to RUclips influencer. Has anyone received one yet ? Or is this a massive blatant rip ?
LAM is exactly what will make this company (potentially) great. a lower price cost is to get it on a scale to a lot of users to build a greater ecosystem, the product itself is kinda useless but pretty sure R2 will be the next big thing if they know how to play their cards right.
I was waiting for your video about It!, thanks for the analysis!
I want access to the OS to be honest
R1 will do alot. R2 won't be out till 2028, as hardware isn't that easy to just put out there.
Folks will have to max out R1 which will take sometime
R1 is using cloud directly, they don't process things in phones instead leverage cloud to process it and that is where they've lowered the cost.
LAMs breaks all the time, as soon as the company changes there website they are completely useless. Everyone forgets one thing. Companies like to have a brand, they are not going to let an AI hide all of that away, they will find ways to block things like this. This will flop massively, but thats just my opinion.
r2 just will look different, all updates are stored in their cloud, so you dont need any upgraded device, only for the look maybe:)
I love the clash of the nice style and editing of your text layers with all their typos. Seriamente adoro ❤
6:44 'straregy' 👌
Nailed it.
Nightmare fuel. Why AI and tech never cares of your presentations? Because it is totally fixated on consumerism and making money to corporations, its all about you shopping and spending more, watching adds, ordering food and tickets. It will never be about you doing LESS, rather about you spending MORE.
I'm disappointed in your buzzword salad
It's like in "Silicon Valley" the TV show "customers want to buy a box"
it's suspicious how every channel is suddenly shilling this product
Talking about what excites people unprovoked now equals shilling?
Damn y'all are way out of your depth on these interwebs
They’re not shilling it. They’re not saying to buy it, but they are saying that it’s the step forward past ChatGPT which it is.
@@1anre lmao you're way out of depth on this life kid
@gloobark if we check in real life, you're still living with your parents but you come ònline to call someone a kid
@@1anre incorrect but i hope you find something better to do with your life than picking arguments online with strangers to defend the astroturfed ad campaign of product whose creators aren't even paying you. it would be less pathetic if you actually were a child but you're a grown man doing it 😆
Side topic: that OP-1 field reminds me in key style of the Olivetti Praxis 48. I love it!
Except it’s not actually AI in the way they claimed it to be in their marketing demos lol. The code’s been cracked, they use Perplexity to search, ChatGPT to understand requests and chat, and Playwright to interface with websites. Problem is, Playwright is a hardcoded script. So if UI’s change, it breaks, as we’ve been witnessing. They claimed to have created a “foundational model”, when all they did was take a bunch of off the shelf tech that they had no hand in making and smushed it all together to create the illusion of an LAM. It’s utter bullshit, and I’d suggest you take this video down as it’s aged horribly. Just another tech scam. Oh plus they’re a revamped NFT company lol. Just tech bros being tech bros.
Device got attention but builds in a locked market. Then when users can sell their own training of rabbits (which they would take a cut) you have a far more profitable business model. On top of which the users build the AI.
You can’t really do that with an app in the current Apple / Android ecosystem without handing over a lot of coin to either of those two if at all. App Store rules and all.
the music is not only loud, it is extremely distracting,
People have smartphones in their pockets they don't want another device to carry around. It's just a matter of time before phones have this exact same technology. If they wanted people to have this as well a phone they should have made it into a watch or something wearable so it isn't two handheld devices to carry.
With Samsung pushing AI, I bet that they'll have all these features in their smartwatches in a few generations.
its a replacement, not a secound thing to carry around, a smartphone is full of distractions, the r1 rabbit will help older parents or my future kids to simply do things without scrolling allday on the phone, 100% it can send messages to others in the future
why so many blind people wanna compare the rabbit with a smartphone ??
I just want an AI that gives me anything I want. Books, specific information, files. Generate images movies, edit things.
This is how you take smartphones era and change them to be dumb phones.
Waited for the iPhone era to over for years now.
I totally agree. And I wonder how these action models will work out with Apple’s iOS products privacy-centered design.
You mean their design that's full of NSA backdoors?
You got to be kidding. Privacy-centered? That's a good one! You mean profit centered right?
I think we are past the age of thinking apple is a privacy centric company. Its just their way of marketing.
My irresponsible shopping habits can’t wait for my late night shopping sprees in a form of a monologue
What?? They have carriages without horses in New York? Since when? The future is now truly here.
But joke aside ;D this AI rabbit is a great idea, but it also needs a few Tamagotchi and Alexa features to really bring it alive and make it cute and interesting to interact with for the big break-through of this device.
The constant background music made it hard to enjoy watching this... 😢
Agree!
I disagree
I also disagree
I also too disagree
Seems interesting but impractical, at least for the time being. What if you ask it to book a place in downtown LA then it sends you to skid row? How does it make that decision? It's a simple solution which would be to converse with the user about the locations while showing them images. Maybe the best approach would be to login to the account bookmark the recommendations then present them like a personal assistant would. It sounds promising though.
Its not designed for criminal city lol, 88% of the population doesn't live in the west, and amongst those 12% atleast 8% lives in relatively safe neighbourhoods and wouldn't have these kinds of concerns, these are micro tweek and updates not a huge impact on use cases for most of its users.
I don't know, this sounds extraordinarily dumb. I would never ever in my life, no matter how good this gets, use something like it. The smartphone does literally everything this shit does and more. This whole horse vs. car analogy just simply doesn't work.
They did not want to be build an app because they do not want to exist as an app in another OS. The reason for the creation of the device is to test a device with their operating system. So if there is an error is to create a device that is unlike current phones. They could have use a manufacturer like xiaomi, or anybody to build the device and test their operating system.
Now you understand why you have to check a box that says: "I'm a human". :D
I honestly think this would integrate with the phone in the future.
I know a little doorbell company that sold for $199 per device then was acquired by Amazon for 11 billion dollars not long after launch that might disagree with your statement there
Techo bros be like:
Useless.. oh my friend i have so many ideas for this.. but probably because I have ADHD and I really could use a natural speech AI assistant that specifically doesn't have access to every single piece of my life through my phone
I am one of the 300 going to the pick up party, I was sold on the training. Figured it alone was worth $200
Useless you say, we'll see, especially if I can train it to upload videos, design web page layouts from existing ones, import and update data used to make maps. Experimental noted, category, not booking trips, think bigger.
Exactly my thoughts....as soon as it can edit (video, audio) and mock up websites, then I'll be sold.
All that can be on your regular smartphone too. Why need another device?
don't bet on this one yall. At least not in the long term. A large aspect of web security is to try and stop bots from automatically navigating websites. APIs or a centralized ecosystem (like what Apple does) is the bread and butter of the internet - don't bet against it
wasnt letting AI on the web the big no-go? well... we are human, figures we'll put ourselves on fire to see what its like...
so the future is to create a huge botnet?
I’m sorry, like you’re brilliant, sir. It feels like you’re reaching out because if you’re telling us these applications that Google is not telling us that is supposed to get us excited, why isn’t Google telling us what it is? It seems like a marketing gimmick similar to pantyhose, lipstick, man, wipes, etc. I don’t understand. It looks like a marketing gimmick identical to pantyhose, lipstick, man, wipes, etc. Just an opinion. My second issue with this is if this becomes something where it clears the playing field, and everyone can do the same thing. Wouldn’t this stifle competition?
This channel is the greatest gem I've found so far in this year. The depth of analysis is remarkable.
Calling it now rabbit is going to be very successful.
Sure...
Lets sell a 200 product meaning we are selling to the middle class. Yet use it to book a 2000 trip that no one can afford lol
Still don't get the hype chief
Will Large Action Models solve Captchas as well? Would they cause bots to be more rampant on the web?
bots can already solve captcha
@marcogenovesi8570 is that how the LAMs have succeeded in navigating the web from Jesse's demo?
captcha is not to stop bots, its to train bots. So it will just be bots training bots and creating garbage data.
So i should just wait until an app is built using LAM?
It wouldn't be an app, because no developers really want to work with Google or Apple so that all your work would be stolen or copied and then you would give up 40% or all your money for every app sold. You and your user would lose all your privacy and your data would be sold to their partners. $200 is a steal for this functionality and the ability to make my life easier and more robust. I have big plans for this thing and ChatGPT is not cutting it anymore. Just like most of the people in the world, the execution is the biggest problem with any project. It looks inviting to see how this device handles things. It could be great or the worse thing. Only time will tell. Right now I am I am going to be an early adopter of this Model T and see if I will send my GPT horse out to pasture when the time comes. Safe Ai-ING everyone. 🖖
Have you paid money to order one yet ?
The next decade is going to be insane with AI devices and how fast AI will grow. Itll be crazy to look back from 2034 to today on the technology of AI. Or we may regret it
Kim Jung Un: "Hi Rabbit, please help me nuke the world now."
This video provides an great overview and explanation of new AI technologies. Really enjoyed it
"Great at reasoning" lol
Whats funny about that?
@@Marquis-Sade It's extreme hyperbole -- at best.
@@chasebase76 Just try it out
Wait no, this is exactly what I want. I want to speak and have my AI actually order my matcha frappucino from Starbucks while I'm driving and have the thought, or book the cheapest ride sharing to my home, without my having to open all the apps, login, compare prices, do it all myself. It saves time, that's what I want AI to do.
there's an EMACS command for that.
Europe doesn’t want this dumbass device 😂
They probably say the same about you.
USELESS, it definitely is not
The device form factor helps to highlight the paradigm shift in UI and application design
On the iOS, iPadOS, watchOS, and macOS App Store, Apple charges a 30% fee for apps and in-app purchases. However, the fee doesn't apply to free apps.Jan 8, 2023. I personally am not one for paying a monopoly if I want to go my own route for my company. Maybe sell the software but not an app.
Source for the film at 0.8 secs?
It's pretty much the same reasoning behind why build anthropomorphic robots when other form factors would be much easier to make: they would be able to easily use any object built for humans
they could already do that without being human shaped, in fact being human shaped may hinder them in some scenarios since they'll never be flexible like an actual human. Lets face it, people just want to make humanoid robots, regardless of if its practical it sells better.
@@ゲンソウ Well, a human shaped robot could both drive a car, open and enter a door, use stairs or open a tap. A robot designed specifically to drive a car instead would only be able to do that. I'm not saying one form factor would be better, they just have different advantages.
Hi Enrico great video as always!
ps: I feel like the audio should be flatter, with less peaks
Device explanation starts at 5:47
How can we optimize UI for AI interpretation?
.. so now we live in a car-infested hellhole 😊 it could have been so good !
It looks and acts just like a Pokédex
You were searching for the right format and content for your channel? THESE!!!! are the videos i (we?) want from you!!
It's not difficult to train ai. It's not as though it only works one way
So this thing is literally just a Baritone bot for the whole internet!
(Baritone is an AI model that was trained to play Minecraft better than a human can, using only the on-screen information that would be visible to an ordinary player. Rabbit's "Large Action Model" seems to use a similar approach, navigating websites through their UI like a real person does.
Baritone is a client mod that has information about all blocks/mobs/structures/etc that are loaded in the world. It doesn't use computer vision to interpret on-screen pixels. It's not an "AI" in the sense of machine learning, but rather relies on hard coded rules.
@@AbruptAvalanche Then I must've misunderstood how it works. I always thought it had pre-programmed knowledge of the game's elements, but interacted with the world through computer vision.
@@InventorZahran It doesn't at all use machine learning specifically. That's just kind of misinformation caused by youtuber's mistake
Easily clarified and dispelled
it’s cool but one device is better, I already got my keys my wallet, my phone, my sunglasses
I’m 64 years old and the point of this entire thing is to simplify. Forget all your tech horse BS and think of the learning curve you have using all these app based programs. I hate them! Then think about no more passwords! God I hate passwords. I need the info in the apps, but I could care less where it comes from! To get what I need to simply make my day simple is better.
Convenience should not be underestimated.
Glitch, Portal ..... Reboot
I'd like to see how the strategy plays out. In my opinion I think licensing out the technology would have been more savvy. Let the other companies build their own ecosystems on top of it and allow the technology to become more ubiquitous. This could potentially have become the new Microsoft.
They can still do that with the OS, but the hardware play caught more people's attention
AI LLM isn't hard to explain...
See your virtual keyboard on your phone guessing the next word of your sentence?
The AI is doing the same for paragraphs.
There. You can copy paste that.
Oh and it's the same for the pictures AI.
AI could end up being the biggest load of rubbish ever, and everyone realises it's just more of the same, when we actually want less.
"In 2023 we have cracked reasoning." I have never unsubscribed to someone so quickly.
Why though. You don't think people reason ?
Isn’t this what Alexa can already do, but I won’t let it do, because I’m afraid of mistakes when my credit card is involved, it may be the future for you but something’s need to stay in the past, all I see is an easy way for fraudulent activity, the future was keyless automobiles but that turned out as just giving the keys to the cars thieves.
Hackers and spammers are going to turn this into everyone's nightmare
Show is why or how?
You are severely confused on the level of “A.I.” this thing provides. (Might have to jb some) ,but as said many times… it can’t do anything MORE than what we already have.
this is ridiculious on a few levels
That’s the first thing I thought too. They will license or sell the model
Needs to include the ability to make AI art and have various pets to be your AI assistant (perhaps one that can evolve and grow). Then you have my buy in! What it does is barely better than a phone and other products will do it. If you want to make a gimmick, make it fun with the aforementioned additions.
I am left handed too, and it inflames me whenever I see a product like this that did not consider left handed use
Did A.I. make all those spelling mistakes in the graphics?
One step closer to ultron
It could never be an app, the entire point is using the same human-machine interfaces to do stuff, because this cannot be locked down like devices can (and have been, especially iOS). This thing does not care nor need to support a OS version or comply with any asinine app store regulation. This is an external tool operating the same interfaces that the device HAS TO expose for the human user
And it absolutely rocks and gives tinkerers the real tool they've been looking for, right?
No, it's not the future. Everything these devices do, a smartphone can do better. And you already carry your smartphone, so why would you carry a second device? It's a gimmick, just to gather the attention of the media and less knowledgeable tech enthusiasts. If you're developing a dedicated device, it better do things that a smartphone can't do, with the addition of sensors or mechanical attributes that fit certain situations and wouldn't be feasible in smartphones.
Does anyone know what music is playing in the video please? It starts from 6:00. 🎶📝🙏
car analogy is weak
cars pretty much replaced horses n carriages.
this is not replacing anything.
So if it's just using a website behind the scenes from how it's taught, what happens when the website changes it's layout or pops up with a captcha?
Seems ok for quick tasks but seriously noone is going to rely on it for anything bigger. People want to make their own decisions, not going to let it choose any random hotel and flight for a thousand dollar holiday. Gimmick.
there are bots for presentation like TOME
Not grateful people with disabilities. You need to push the button to make it work. Oh well… At least the paralyzed are small demographic in terms of not mattering at all.
Excellent analysis and summary. Thanks! I just pre-ordered in batch 6. Also looking forward to this giving me an edge at work. Haven't been as impressed with LLMs as this approach to developing an LAM. My information technology professor from a few decades ago would be so pleased to see intelligent personal agents finally emerging.
Sorry for your loss.
so the bottom line is that in the future, the apps still meed to be built so that the LAM can figure it out the apps. Only then we won't need to use the apps?
inb4 concert tickets get even more obscenely overpriced because this AI can get around anti-scalping bot measures on a massive scale
A possible solution could be to auction concert tickets. Would that work?
@@Biosynchro wouldn’t that just raise prices even higher?
@@glitterishhh Concerts are luxuries. In any case, after an auction, there is little incentive for scalpers. I would like to see this tried. I'd also like to see it tried for movies, too.
and like car blocks the streets today will ai block the internet...
Useless.
Can't they just make it an app tho