Rabbit CEO Jesse Lyu isn't thinking too far ahead

Support for Decoder comes from NYU Stern Executive MBA program. For a lot of professionals, the idea of earning a higher degree can feel like a tug of war between furthering your career and installing your business development. Well, with the NYU Stern Executive MBA program, you can fine tune your emotional intelligence, enabling you to make smarter, more human-centered decisions that can drive your business forward. And you don't need to put your life on hold to earn your MBA. Programs are available in both ways.

The U.S. City and Washington DC. Classes are held every other Friday and Saturday in New York City or one Friday, Saturday and Sunday each month in DC. Search NYU Stern EMBA to learn more. Search NYU Stern EMBA to learn more. Support for Decoder comes from Janus Henderson Investors. If you have questions about investing without paying for expensive advice, you might want to check out Janus Henderson Investors.

You can get access to over 90 days of world-class investment expertise. They offer free, personalized investment advice to help you evaluate your investment strategy. All you need to do is ask. To find out more, visit JanusHenerson.com slash advice. Spelled J-A-N-U-S-H-E-N-D-E-R-S-O-N dot com slash advice. Janus Henderson, investing in a brighter future together. There are no additional costs for advice beyond the underlying fund expenses.

It's time to review the highlights. I join by my co-anchor, Snoop. And what up, Doh? Snoop number one has to be getting the new iPhone 16 Pro with Apple Intelligence at T-Mobile. Yeah, you should hustle down a T-Mobile like a dog chasing squirrel, chasing the nut. That's a nice analogy, Snoop. On the highlight number two, a T-Mobile family is saved 20% every month versus the other big guys. Very impressive. Take it away, Snoop.

Head to T-Mobile dot com to get the new iPhone 16 Pro with Apple Intelligence. On them. Now drop that jingle. See how you can save versus the other big guys at T-Mobile dot com slash switch, Apple Intelligence coming fall 2024. Hello and welcome to Decoder. I'm Neonipa Tell Editor and Chief of the Verge. And Decoder is my show about big ideas and other problems. Today I'm talking to Jesse Liu, the founder and CEO of Rabbit. It's the startup company that makes the adorable R1 AI gadget.

It's a little handheld designed by superstar design firm Teenage Engineering. It's meant to be how you talk to Rabbit's AI agent, which then goes off onto the internet and does things for you from playing music on Spotify to ordering an Uber and even buying things from Amazon. Rabbit launched with a lot of hype at CES and then a big party in New York, but early reviews of the R1 device itself were universally bad.

Our own David Pierce gave it a three out of ten back in May saying most of the features don't work or don't even exist. And the core feature that didn't seem to exist was the most important of all. Rabbit's large action model, or LAM, which is meant to allow the system to open a web browser in the cloud and browse for you. The LAM is supposed to intelligently understand what it's looking at on those websites and then literally click around to accomplish tasks on your behalf.

There have been a lot of questions over the past ten months about just how real Rabbit's LAM was. But the day before Jesse and I spoke, the company launched what it calls LAM playgrounds, which lets people actually use a bare bones version of the system. And it does indeed appear to be clicking around on the web, although it is obviously very early and it is very slow.

So I wanted to know how Jesse planned to invest in the LAM and compete with all the other AI agents being announced that also promise to do things for you. For example, Microsoft just announced a new agency version of Copilot and Apple's vision for the next generation of Siri is an AI agent. One that will run on the phone that you already have and have direct access to your phone apps and the data inside them.

It's the same with Google and Android and Gemini and even Amazon's rumored next generation of Alexa. This is major competition for startup and Jesse talked about wanting to get out ahead of it. But what I really wanted to talk about is how Rabbit's system works and whether or not it's durable, not just technically, which is challenging, but also from a business and legal perspective.

After all, if Rabbit's idea works and the LAM really does go out and browse websites for you, what's stopping companies like Spotify and DoorDash from blocking it? You might have a strong point of view here, Jesse certainly does, but at some point there's going to be a fight about this and it is not clear what's going to happen. Here, give me just a second to put this in historical context with an example.

About a decade ago, a handful of startups all tried to stream broadcast television without permission or licenses by putting a bunch of TV antennas in a single location and then building apps that people access them. This felt technically legal, but it's the difference between all of your customers having their own antennas and putting all those antennas in a single place and letting your customers access them over the internet.

And some of these companies were seriously innovative. The most famous was a company called Ario, which spent a ton of money designing specialized TV antennas the size of a nickel, so they could pack as many of them into a data center as possible. I wrote about Ario back then, I visited their offices in Brooklyn, I saw the antennas, I interviewed the CEO, the whole thing.

Yeah, Ario got sued by the TV networks, the case went to the Supreme Court in 2014 and you will note that Ario no longer exists. I don't know for rabbits in another area and I don't know how all these companies will react having robots, browser websites, and set of people. And I certainly don't know how legal systems around the world will handle the inevitable lawsuits to come.

I asked Jesse about all this and you'll hear his answer. He thinks rabbit will be so successful that all these companies will show up and want to make deals. I got to say, I don't want that either. I do know that this is a pretty intense and occasionally contentious interview. Jesse did it back down and that means we got pretty deep into it. Let me know what you think. Okay, Jesse Liu, the founder and CEO of Rabbit. Here we go. Jesse Liu, you're the founder and CEO of Rabbit. Welcome to Dakota.

Thank you, Neonai. Glad to be here. I'm very excited to talk to you. Rabbit is a fascinating company. The idea for the R1 product is fascinating. I think a lot of people think that something that looks like the R1 is the next evolution of smartphones or products or something. And then there's the company itself, which is really interesting and you've got a connection to Teenage Engineering, which is one of our favorite companies here at the Verge.

And you've got some news to share about opening up Rabbit's large action models so people can play with it. And it's kind of an early version. I really want to talk about that. But let's start with Rabbit itself. The company has not been around that long. The R1 just started shipping six months ago. What is Rabbit? How did the company start? Here's a little bit of history of it. So I actually started a company back in 2013, which is called Raven Tag. We were at YC Winter 15 batch.

It's basically my personal dream to chase this grand vision that I guess me being this generation grow up, we watched so many sci-fi movies. There's AI stuff here and there. I guess every geek wants to build their own job at some point. That's exactly how I started the Raven Tag 11, 12 years ago. Back then, we had this idea. We had this direction. But the technology back then obviously there wasn't GPU training. There wasn't transformer and stuff.

So we worked really hard on the early days of like voice dictation and LPLU, which is natural and processing and natural understanding. So the technology won't there. We tried our best. We actually built an entire cloud system and the hardware, which is similar to what we have in Rabbit today. But the phone factor was like more of a smart speaker as if we all know back in 10 years ago. Everyone's chasing that phone factor. And ultimately the company got acquired.

So it's not a new idea for myself, but it's definitely a new opportunity that when I saw the progress on the research side of transformer, obviously I got a chance to try to change the chat GPT or GPD's API very early time. We were really impressed because we felt the timing is right because in order to do something like R1 or more sci-fi job with stuff, you really need to figure out two parts from the back.

One is that you want to make sure that by talking to the device, the computer or device actually understand what you're talking about, right? Which is the transformer, the large language model part. But we believe at the around 2020, 2021, we believe that the transformer is absolutely the right path that opening out of companies are heading to.

We believe that portion is being solved will be solved. So our focusing media to shift to, you know, after this device can understand you can actually help you do things.

And the company that started 10, 11 years ago, RivenTek, we were actually one of the first company that we designed across API structure that, you know, after the recommendation, after understanding the query got sent into different, you know, API system has a detector to understand, oh, maybe you're looking for a restaurant, you know, maybe you want to play a song from this streaming software.

But, you know, I guess 10 years ago, there's a great opportunity of API. So there's a lot of company working on APIs. And if you remember, like 10 years ago, instead of going to valley, like everyone was talking about maybe in the future, the entire operating system will be just HTML5s, right? But that didn't live quite long. So I think now when we're looking after 2020, like the API business is not really like, you know, major business for most of the popular services.

So we also want to take a evaluation of, you know, whether we can build like a generic piece of agent technology, which is really hard because I believe the current AI is all generic. Obviously, there's a lot of people doing vertical stuff, right? You can build an agent for Excel, you can build an agent for legal documentation process.

But I think the biggest dream, what really make us excited is like a generic part of it. It's like, can we build something that without pre training without knowing people want to do what? And they just tell whatever they want and would be able to smart enough to handle all the tasks. So that's why we felt the opportunity was right. And we started rabbit. Yeah, after COVID, yeah.

The idea that agents are going to be a big part of our life. And in particular, general purpose agents that go take actions for us and the internet. I've heard the city from all kinds of folks from start founders like yourself to the CEOs of the biggest companies world. I want to come back to that. That's a big idea.

But I just want to stay focused on rabbit for a second. How many people work your rabbit today? At the current moment, we're roughly around 50 people, 15 to 60 people if we passed the interns. But when we started the company was seven. And by the time we launched our CES was 17. Yeah. So just by growing the team within like four or five months, it's quite a challenging job for me. Yeah. So CES was a big launch. We were there on David Pierce's party. Yeah. The rabbit was introduced.

You gave demos in a hotel room, I think. And then you had the launch party here at the TWA hotel. JFK, which is very cool. The things were now, but you've been growing. You said you started 17 people in January. CES, you have 50 now. What are you adding all those people to do? Most of parties just engineers. We have a very small group of design slash hardware design or ID that we started from day one.

And most of the new folks are working on AI and infrastructure perspective like cloud. Basically, we would not only ship the hardware, we build the entire rabbit OS for it. Right. So I think the major work is always being going to be in software apart. Yeah. How's the whole company structure as you go from seven to 17 to 50 after we have obviously to decide how to structure rabbit. How is that structure now? How's it changed?

We are primary located in Santa Monica. We have a device team, really great folks in Bay Area. And we have a couple of research engineers here and there. So it's kind of like a mostly in person, but some was a hybrid system. And the way that we find our people is mostly by internal referring. So we're not like, you know, spending money chasing for like agents agencies to do the hiring. Most of the good folks that we basically do an internal recommendation.

Yeah, but how are your 50 people you have now? How is that organized inside the company? It's really flat in a sense. You know, we have different department. Obviously, you know, the hardware, OEM, that part is in Asian. You know, we have our ID team in collaboration with folks in Stockholm, Teenage Engineering in this case. And we do our own like graphics and marketing all that in house.

And then for the software part, we have the device team that they need to work with the OEM OEM. And we have the cloud team. We have the AIT. That's basically how much team we have. And each team. There's obviously cross overs. And we basically working project base. So there is no like really like hierarchy like crazy hierarchy going on. I mean, the biggest company I ever was back in the Raven. I believe by the time we got acquired, we're like 250 people.

So this is like still within my comfort north to manage like 50 people. So yeah. Teenage Engineering is clearly a big part of the rabbit story. They designed the R1 and the co-founder. Yes, for a cowthoof. Is your chief design officer? How much more hardware are you designing right now? Are there iterations to come? Do you have a roadmap of new products?

Yeah. So the way we work together, obviously, this is not the first time we collaborate with the collaboration back in Raven. First of all, Teenage Engineering is my hero company. It's basically a fanboy dream become true story for me. I really appreciate their help over the years. The way that we work together is very intuitive. There are obviously many ways that consider to be like the proper way of designing a project like this.

But I think we're out of the ordinary way of doing this. I can give you an example like back in the Raven. All we did is that we had probably two meetings in person, a couple of phone calls, no email, no text messages. We set up a secret Instagram account that we just shared sketches and we just hit like on that Instagram account. And that's how we designed the previous Raven project. This time is even quicker. I think I shared this publicly.

I think we spent like probably 10 minutes on deciding the R1, how it's going to look like and you know we have quick sketches here and there. And ultimately, I pushed the SBB back for using the current color, which is the orange from Raul.

We do have maybe like two or three projects in our mind, but I think by the end of this year, our current focus is to really get this lamp pushed to the next level. So yeah, stay tuned. I think one thing people will realize is that this team do hardware really quick.

Because when we start sketching the R1, it was like last year back in November and we introduced that by January and we start shipping by April. So if we want to launch the next project, it's going to be like roughly in, I don't know, six to eight months timeframe, certainly not like a year or two.

But that being said, I think I was having my own community voice chat yesterday, I was talking to people about the current R1 because I really don't like the current consumer electronics like one year per generation by default, regardless. Right.

Like we've seen that from the smartphone companies and doing annually release for all this stuff with minor changes. When we start designing the R1, the entire rabbit OS runs on the cloud. That means that the this piece of hardware, even though it's 199 and not the latest chips. It's really capable of offloading the future features to this device. So I don't think R1 is like a one year left spend device and so that's our community thought they think we can tweak so many things about it.

So in that sense, we're not in a rush to drop another version of it, but we do have different phone factors in our mind at the moment. And it is yes, we're actively working on those designs or as chief design officers, you're working on something else. He was literally in our office last Sunday, which is three days ago. Yeah, we are actually working together. Okay. How much money do you raise so far?

That's a good question. I wouldn't be accurate, but it's somewhere around 50 million total in the whole lifespan. Last part was 35 million that by sound venture and also host venture and Amazon Alexa found its synergies. So last round was 35 and if you consider the all the money together, I think it's around 50.

Can I look at the amount of money that other AI companies are going out to raise open ads right as we're speaking open and I just raise the biggest round ever in history to go build obviously a foundation model digital God, whatever same Altman things he's doing. Do you think you can compete at 35 million dollars around?

But I think talking about competition money is one part of it. I've considered myself a veteran because I've done startup before I know how it works. Certainly money is very important probably most important in the early couple of years, but I think when we talk about competition, we ultimately want to ship products to consumers.

Because the way I look at it is that people are not buying electricity. Electricity is basically controlled by hearing California, South California, Edison, right. You have an address you have to pay for it regardless of how much it actually is you're using. But I think people are ultimately buying microwaves cars, motorcycles, televisions, people are buying products powered by electricity.

So research wise, I can say very clear we have at this moment of rabbit. There's no way that we can compete over opening and drop pick and deep mind and Google. But how can we play the game? We become partner of everyone, right. So our one is hosting every single model, the latest model from these guys.

And we offer their capabilities combined with our product and additional and rabbit OS and all the features offered to our user. So there's no way we can compete over on a research perspective, but we should product fast, right. You saw opening I just released the instant API as they call it. I was actually invited to the meeting, but I'm launching the lamplight ground yesterday. So I couldn't be there in person.

But they're offering a API for people to build a agent for it. But yesterday we dropped a lamplight ground, which you can go to any website and just do it by voice. So I think competition is different magnitude. I think money is definitely important. We hope that we can raise more money. Of course. But I think right now, if you talk about competition, we have to play smart.

They are good on the research. We are good on converting all the latest research into a piece of product that user can use today. Let's talk about what that product is today. So right now you have the R1. You can buy it to beautiful piece of hardware. It is orange. It is very striking. It has a screen, it has a scroll dial. And then it has a connection to your service in a cloud, which goes and does stuff for you.

That costs $199. Are you making money on the sale of each individual R1 unit right now? Correct. So what's the margin? What's your profit on R1? It's a very good margin, even though I can now tell you the details, but it's over 40%. To make it over 40% on the hardware margin of the R1. On a hardware margin. We run a calculation. We might have to redo the mass because yesterday, they literally after drop the lamplight ground, the server crashed like multiple times.

So we might need to redo the calculation, but again, first of all, in the beginning, we are making money. Now we have this more powerful features moving forward. I think I haven't heard a company that went bankrupt because they got a popular service that is so popular that they couldn't afford cloud bills. I think if you build a good product. I can draw that line for you. So it's a buck 99. You're making over 40%.

So that's between 80 and $90. Right? It's not 50% which we $100. So it's a little less. So between 80 and $90 in margin. That margin has to, you do have to pay your cloud bills. Right? So is that margin all being fed into your cloud bills? Obviously we have this dedicated instance with all this cloud competitors. Right? I mean, don't get me wrong. Like the Amazon AWS. We're hosting on AWS.

And there's AWS Google Cloud Microsoft Azure on the LN partnerships. We have an Olympic opening AI and Gemini. So don't get me wrong. That's a lot of companies that like to make a lot of money. I just want to be like, they're not cheap to partner with all these companies. They're not cheap. But what I'm trying to point out is that they are competing so fierce in a way that they have a lot of good benefit for the early startups. I have to shut off for our companies. Sure.

So they really want to figure out a way to help you on board and maybe making your money in the long run. But I think at this current scale, we can totally handle it. Yes. So we got great deals from them. So if I buy an R1 from you, you take $90 a margin or $80 margin. At what point, how much do I have to use my R1 to turn that negative for you? Because everything I do with an AI, that's a token that token costs money across multiple services. You're bandwidth cost money. It all costs money.

How much does a single R1 user have to use their R1 to take up $90 margin or $80 margin for me? I think if a moderate user using in a non-robotic way or non-malicious way, it's going to be really hard to break down negativity. But a two years worth of usage, one year, six months? I think it's definitely over a year and a half. I'm not sure about two years because there's new features going to implement into this, including landplayground and teach mode.

But yeah, so I want to kind of like share my understanding to this is that yes, we did the mathematics. We are mature. No problem. We wish we can sell more, which we're hoping that we can sell more. That's going to definitely help. But the target of this whole launch strategy is not set for making like X amount of money on like first six months. I think there's other companies that really greedy about how they want to launch their product. I'm not going to even mention the name, right?

That won't work. That won't work. I think if you look at any new generational product, if the founder and the board decided to set up a strategy that less quits every single penny out of the user, it's not going to work. Because we know AI is very early and we know that there's going to be a lot of things that go around. In fact, I believe that every company, regardless of if you're big or small, if you work on the latest AI stuff, as the first two weeks, it's going to be disaster.

Because you're going to find a lot of the misbehavior by the AI. You're going to find a lot of the edge cases by the model. So I think the whole thing is too new. There's no way that we want to like charge for subscription. That's even like worse. I don't like that strategy in general. So even though this sounds very concerning that, okay, you can easily twist my story or someone might twist my story.

But like, oh, rabbit is doing everything great except they're going to bankrupt no matter what, right? I think it's a very stupid way to think in that sense because a great innovation, you have to focus on the innovative part first. Then when you figure out the money part, if we start figuring out the money part, now these making sense. Yeah, now these making sense. I think there's another people in the industry that they have a great understanding of everything.

And then they decided to release a wallpaper app, charge for 50 bucks per month, right? Hopefully that works, I guess. Yeah. You can go talk to that guy and you say, hey, there's no way you're going to bankrupt because your money checks, all these equation checks. If you charge for this, you're going to be making money. But that based on the perspective that the whole logic needs to stand up, right?

So like I think I'm not really wasting a lot of time of my time at this point. I'm trying to like basically fine tune you a little about about the mass-matting equations to make this more like 20% 50% obviously as a startup, we need to survive. Yeah. And I think even though we have a like a rollercoaster right since lunch, but we're growing and we're surviving and we're still pushing the features that now the other devices can do, which is a very, very good sign. So yeah.

We're going to have to pause for a quick break here. We'll be right back. Support for decoder comes from service now. AI is set to transform the way we do business, but it's early days and many companies are still finding their footing when it comes to implementing AI. Service now partnered with Oxford Economics to survey more than 4,000 global execs and tech leaders to assess where they are in the process.

They found their average maturity score is only 44 out of 100, but a few paysetters came out on top and the data shows they have some things in common. The most important one is strategic leadership. They're operating with a clear AI vision that scales the entire organization, which is how service now transforms business with AI. Their platform has AI woven into every workflow with domain specific models that are built with your company's unique use cases in mind, your data, your needs.

And most importantly, it's ready now and early customers are already seeing results. But you don't need to take our word for it. You can check out the research for yourself and learn why an end-to-end approach to AI is the best way to supercharge your company's productivity. Visit servicenow.com slash AI maturity index to learn more. Support for decoder comes from Vanta.

Whether you're starting or scaling your company's security program, demonstrating top-notch security practices and establishing trust is more important than ever. Vanta automates compliance for SOC2, ISO 27001 and more, saving you time and money while helping you build customer trust. Plus, you can streamline security reviews by automating questionnaires and demonstrating your security posture with a customer facing trust center, all powered by Vanta AI.

Over 8,000 companies like Atlassian, FlowHealth and Quora use Vanta to manage risk and prove security in real time. Get $1,000 off when you go to vanta.com slash decoder. That's right, that's vanta.com slash decoder for $1,000 off. They're not writers, but they help their clients shape their businesses' financial stories. They're not an airline, but their network connects global businesses in nearly 180 local markets. They're not detectives.

But they work across businesses to uncover new financial opportunities for their clients. They're not just any bank. They are city. Learn more at city.com slash we are city. That's cit.com slash we are city. Welcome back, I'm talking about the rabbit founder and CEO Jesse Liu. Right before the break I was asking him, how many AI tokens a user would have to use in order to cost rabbit more money than the device brought in.

His response was that rabbit had done the math and it was fine. Then he started calling out his competitors and others in the space. Let's get back into it. So one, I don't think anybody has ever linked criticism of humane to criticism of Mark has is wallpaper app on our show before well done. I think Mark has a very different view of where his expertise is in what went wrong with that app and maybe one day we'll talk to him about it.

But my question for you when you talk about growth and you talk like the unit economics of the rabbit is on some curve. The hardware becomes unproperly for you just me having a rabbit for longer than 18 months becomes unproperly for you. That's the moment that you would charge a subscription you would say to continue using this thing. It can't be negative for our company and that's the thing that I'm pushing on here.

I think there are multiple solutions to that question. One is that obviously if let's use our one for every user for more than 18 months there's a couple of solutions. One is that we're going to launch the next generation device. Try and maybe multiple devices still profitable from the hardware to I think we have this prepared since day one from last week we roll out the RF a teach mode to a very select group of testers.

I would love to give you the access so please reach out to us later on we'll see if we can help you set it up. But we roll out of the very small group of our testers roughly around 2025 people to be honest and then over the last 72 hours I saw more than probably 200 more than 200 lessons or agents has been created through teach mode.

If you look at the current Apple ecosystem or Android ecosystem I think the hardware is not going to be the number one money contributor. It's really hard to make the top of the margin of the hardware anyway.

So at some point you want to convert that into services and software right that doesn't means that you're going to charge subscription for the device. What I think is very promising is that we're going to slowly roll out the teach mode to beta testers and hopefully by the end of this year we can like.

Open the teach mode as we promised on day one so all these lessons created work or or rabbits or agents created by each independent users or developers they can be considered as a new generation of app store on that we commit big money using the App Store economics of taking 30 percent. I don't want to invent any exactly I think it's very I think I'm not trying to invent any new business model. I think as a startup is very risky to invent your own business model.

But there is a very great business model all there which is the app store and that's contributing like 70% for the any income right. I'm curious just as I've played with our ones and and looked at the device I've always wondered how on earth we're making money at a buck and that makes sense to me.

When you think about what the rabbit is actually doing ask it a query shows you're in a beautiful animation on the screen which is adorable and it goes off into the web and uses a bunch of APIs and now the new large action model which is the news right yesterday you announced the large action model playground people can watch it work.

And the lamb click around on the verge website just to read headlines which is neat is that the back end of this I ask the rabbit to do something in the cloud it goes and clicks around on the web for me. We have to separate two different systems here maybe three different system here as talk before yesterday because yesterday is really a great milestone.

Before yesterday what happens is that you talk to the R1 we have a intention triage system which basically we go with convert this audio to a text with send that text to our LL providers and then we have an intention triage system from there like after the LM understand the intention we send to different APIs or different features there are a lot of feature which is on device right like set a smart timer or something like that or there's like a simple question.

We think that there's other services or model probably answers better than default LLM so sometimes we send a particular query to propaxity sometimes we send a particular query to Wolfram Rafa so it's you can understand as you know intention triage system is dispensing on on this to different destinations.

And then the relative features will trigger but after yesterday which we have this playground and that's a first stepping stone towards what we really want to create which is a generic cross platform agent system it has to be generic which on this case it is a generic it is not cross platform yet because it handles on the website it will be cross platform very soon but with this generic website agent system essentially you can just talk to rabbi like hey.

Go to ABC website or go to somewhere and then how me do this so that's exactly how we wish to design a product and I think everyone in the industry is heading towards this direction which is you say something we understand you and we help you do it and what happens as we put a windows on the rabbit hole that you can see is that the agent will break down you know different steps I'm going to Google first I'm search for the verge I'm clicking to verge home wax.

I'm trying to find this title as you request it I'm clicking the button to share this and in theory can change multiple steps even steps follow up queries to the system so I give you an example I tried I I think I showed this to another reporter is that hey go to read it first right and search for what are people's recommending for the 2020 for best TV 4K HDR get a model then go to best by.

Add that to my card if best by is out of stock then search on Amazon if they both are out of stock get me the second recommended model so you can actually change different queries and you can pause it you can add you can tweak it you can find it so it's really just like a playground you can freely explore the system and the system is fairly good enough to do like daily tasks and people are you know obviously developers and our hacker in swastics what hackers of course are.

Giving us like impressive like showcases there are people using the land playground to create a app by just by talking to our one because there are third party AI destination that you can just you know use prompt and create a app and download the code and stuff like that so it's really amazing to see all this great showcase just like within actually

precisely 24 hours yeah so I want to make the market between yesterday and in the day before it right you announced the rabbit at CS in January with the lamb. But it wasn't there why announce it without its.

Fundamental enabling feature it is not accurate I want to take this opportunity to address that if you go to the connections now we have seven apps by day when we have four apps those are the first iteration of lamb which is not a generic technology we never on the CS claim that you cannot go to Amazon or something right we said we're

working towards this piece and today there's four apps that you can connect we're going to add more services and over the past couple months we add we did add three more services so as if today there are seven services in total then we keep working on the current lamb playground and when the time is right we swap it so there's a lot of debate say lamb wasn't there that is not true I can trace back to where this rumor starts is where there are people hacking to the R1 they saw R1 is

fundamentally powered by Android system on the local device and obviously that should be the case will be more sketchy if it's non-enjoyed so at the bottom of it is Android system and they don't the code which you can do that in fact every good piece of hardware in history has been hacked someone goes into this and you know gel break the R1 which I guess every piece of hardware is joke

breakable at some point so they and I'll obviously that's the factor to us if you build a software and no one even bother to gel break it's probably not a good a phone factor anyway so people gel break it find out the Android code they don't the Android code to another media and they say hey there's nothing about AI here there's nothing about lamb here of course because all the stuff is in AWS so that's where the rumor starts and then there's a lot of you

know media and they just take that piece I read that the apps you started with Spotify or toward Ashley or a few others those are APIs right you were using very you were actually opening Spotify on the web yes and clicking on it yes yes because why what you mean why there is no

that's the most brittle way to spotify in the cup there is no API there is no API so you had a spotify can run on spot that's a that's a partnership that's a partnership go to Spotify read their documentation yeah there is a specific line is that you cannot use API to build a voice activity application so is really so spotify right now on the R1 when I asked a play song it goes in open Spotify and the web so close to the window yes and then you're re-streaming the audio to my device through your

site correct does that's the spotify know that you're doing this yes and they're they're okay with that we have a conversation they realize this is agent behavior and we said look we ask user to log in on your website and there 100% vigilant user and they're paid user and when we do the trick we help them click the button I've always been very curious about this I'm I've been dying to ask you these questions so I ask my R1 to play a song somewhere in AWS a

virtual machine fires up opens a web browser open Spotify logs into my Spotify account using mic credentials clicks around on Spotify pushes a button to play a song and then you capture that audio and re-stream it to me everything is accurate except we don't help you logging you have to log in for yourself and we don't save your connection the part where you are re-streaming audio that spotify is playing to your virtual machine to me that

you're doing that we are basically gave everyone a virtual machine which is a VNC which is lead which is 100% within policy right and you have the rights to access that VNC and on that VNC we basically work directly on website just like today's landplay so we're not getting the audio from server from Spotify for somewhere else we're basically going to Spotify website and play and do the things for you and play that song for you

okay but where does the song where do the bits go the bits come to the virtual machine and then they come from the virtual machine to my rabbit correct so you are you re-streaming the song to me I am not re-streaming the song to you I am basically presenting the VNC directly to your R1 how to explain how that works maybe I'm not technical enough to understand how that works you're presenting the VNC to my R1 correct so it's it's running locally on my computer with no UI

with no OK I see with through so I walked into the cloud computer R1 is the client to a cloud computer in Spotify is playing on that cloud computer and the R1 is taking that on yeah OK that raises like a million extra questions right yeah yeah I see where first of all I see where you're going before you go deeper I just want to say first of all that we're not using API second of all to say that is not there that's false

claim because because we have all this services if you really pay attention to their documentation there is no API for like door dash there is no API for over right but I just want to be clear that's a choice those

companies have made to prevent companies like rabbit yes from automating their services and disintermediating their services from the user so as you think about these agent models going out onto the web however they're expressed whether it's the lamb whether it's whatever you're doing before the

lamplay grant it all of those companies are going to have a point of view on whether agents can use their services in this way that's pretty unsettled yeah and I'm curious you know you have a few services they might have just said OK let's see how this goes but over time you're going to enter into a much more complicated set of negotiations that will actually be probably determined by the big companies making deals right you can see how open AI or Microsoft or Amazon

would make a deal to have door dash accessible by agents and door dash would say we've made the steel you can't be accessible how do you solve that problem it's not a problem for now we'll see how this problem involves but I remember when Apple is relatively not so big I mean not as big as today when I read the Steve Joss book there's one chapter he said OK go talk to Sony from tomorrow when's 99% right remember that

moment so at some point this level of negotiation needs to be happened I'm not sure if we're leading this or someone else is leading this but this is the working proof that we're not using API and I don't think the services are not building API just because they're trying to prevent people from automating the company it just because API to them is not making money and they for sure will love to have set up the negotiation in some face later when we grow bigger

but I guess you know we try to reach out to Uber with this before lunch they're like who are you you're too small that's it we don't care and so then when you give Uber on the R1 now that's opening the Uber desktop app no the Uber website which is very jenky which is very useful right now that's what I'm asking that's sorry what I meant by desktop app is in the web browser you're calling an Uber yeah yeah if you're

running on Android why not open an Android virtual machine and use the Android app it is a little bit more technical to achieve that which we are working on the other platforms I think I showed very slightly group of people of a working prototype that lamb is operating on the desktop OS such as Linux with all that local apps so we're definitely heading to that direction is there a possibility they can detect the fact that these are not human users

but in fact agent users I guess there's always a way that you can detect but I think the question is this is actually a very good topic that we're talking here is that you know think about captures sure lamb click lamb playground or any capable AI models now can go there and solve tech space captures so they're old system to prevent automated systems like this are currently fouling and this is an industry effort to push everyone in the industry rethink about now with this AI now with all this

agent how their business going to reform or how their business going to how all these policies needs to be changed I do agree this is a very complicated topic so but what I can see is that this is not this is not rapid doing some you know really fancy magic here like every company is doing this we

have other agent company like motel even the GPs are doing this right so this is like a new wave emerging for all this old services that they have to think about but I can tell you my personal experience dealing with scenarios like this like when we first start building one of the first

smart speakers back in like to some 13 you know like all this music label they don't care they don't care until everyone's building smart speakers they're like okay we have to re-solve the whole copyrights for this particular phone factor I guess end of days about money right they want to sell

the same copyrights to as many phone factors they want if there's a popular one so we're okay to have this kind of negotiations but you know certainly like you said there's bigger companies that are doing similar things or even more advanced things that needs to be addressed I give you another

example like you know Siri and Microsoft there's a feature called Microsoft recall right which they pull back that feature now and I think they relaunch it yeah which is very aggressive that is taking screenshot of your local computer this is what I saw was happening in the early days there's going

to be a lot of like different takes and tries and eventually people will reconcile and agree on single piece of like terms and agreements but if you check how we automate the website to their interface the most important part is we don't create fake user we don't create spam user we don't

log in on your behalf and you are you the way I help you to do things is by help you click the buttons and mouse it's equivalent of if I want my buddy to help me I'll give you example so if I'm busy I'm about to head into the meeting I want my buddy to help me order a burger from

DoorDash all I need to do is I unlock my phone I pass my phone to my to my guy and my guy help me click that and in this process I'm not sharing my credentials to my buddy right I'm not telling my phone password I'm not I'm not telling my door dash password I'm not even sharing my credit card

info all he do is just add to the card and click confirm that's it so this guy is equivalent of the first generation of lab which is unfortunately we don't like it so that's why we order so hard now we have playground which is much in our technology well let me ask you about that difference

between the first generation of lab and the playground the playground sounds like the thing you've always wanted to build right you actually have an agent that it can look at webpages yeah understand them take action on them the first one the first it might have been a lamb in the broader definition

but as technology was expressed as like testing software there was moving in an automated way through these interfaces right you weren't actually understanding the interfaces you were able to just navigate them and well yeah that's like pretty normal robotic process automation

stuff where you just building on that kind of technology while the lamb came into existence no no no okay where we're going yours and bar like right so even in the first versions yeah and but you could only understand well so for example the question I've always had is what happens

if Spotify before the lamb exists because I understand that the claim is that this version can understand every website but if Spotify changes its interface well or door to ask changes interface rabbit was kind of getting tripped up right I'll tell you Spotify change interface all

the time right and I think in the past six months five months since the first lamb was adding the Spotify with the connection since launch I think we probably put the Spotify under-minus for like maybe two times one hour in total that's a very hard proof yeah it's a it's a hard proof but I

just take it for what it's worth I don't I think that means it's not good enough right uh the Spotify app on my phone never goes down from maintenance and if the claim is the agent can go take actions for me I have to rely on that at a hundred percent and so I think the question for

me that I have this whole thing is the delta between what you want to do which is have agents go and crawl the web for me and the reality of what we can do now actually the middle ground is APIs right the middle ground is not so brittle right like I okay I that makes more sense to me the agent would

instead of using an interface design for my eyes I'm using interface design for computers I really want to laugh hard okay great uh two things I disagree that the Spotify's not working good Spotify's been working amazing sure five months maybe two times we put on the

maintenance a total amount of time put on the maintenance probably on a one hour you can ask any our one users that's not through API which is impressive that's through agent that's through agent to handle I get that's present for an agent I'm just saying

the you said it's not I said it's not good enough it is not good enough where's the curve where it's a hundred percent okay that's my second part yes API is a hundred percent but you're relying on they give you the API that's stable that works I'm the user right like that's what I'm getting at

as the user why should I care users doesn't need to care we need to care okay we need to care and we need to care because we checked what are the good APIs we can use don't get me wrong for packs the API has been great sure open-air API you know break every every there are two

on they said we observe an issue you can follow the it's chat gbg down there's like a very detailed how how many you know how many breaks per day it's you know more than I guess more than 10 average that tragedy t api breaks or instable whatever it takes we have a notify so API first of all API

is not stable it is not stable sure and you have to chase for this services people won't we want to offer this music feature and we think Spotify's best experience overall and we want to chase for this partnership and we're still chasing for this partnership but to talk from technical perspective

why I said I don't like api is because think about alexa alexa speaker are all using apis and you literally have to go there and negotiate because like I said today not everyone's opening apis a lot of the traditional services don't have api and then startups for startup

there's impossible you when you go talk to them they think you're too small right we did that we just did that to everyone they think we're too small they don't care so we can't get api does that means that we're not going to figure out the option way to make it work no hell no we're going to make it work and this is exactly how we make it work so we care about user to use this feature we don't care about how to do it in fact because we know that you don't care how this been done I don't

want to spend six months eight months suit up to talk to Spotify people and other people and one by one let's do that right so well the the promise here is you're going to eventually have a general purpose lamb that is just using the web for you right at least you hand your phone to buddy which is

why you can make a the rabbit device and just talk to it and it goes off and does stuff in the general case the enormous death star that everyone sees is that apple has announced substantially the same feature for Siri on the iPhone yeah and apple can get the deals and apple can pull developers

into an api relationship locally on the phone with Siri and apple honestly can just burn money until it chooses not to build a car or whatever it wants to do and getting people to buy another device that doesn't just fall back to the Spotify app on iOS when it breaks seems very challenging

how do you overcome that because if the technology isn't a hundred percent better a hundred percent of the time that feels like a challenging sale yeah this is fun part of the game ready how do you win the game first of all speaking for myself I've sold my company before when I was 25 I don't

want to build another app I should chase my same dream because I I really think that the the grand vision that I have and our team was working on is actually the current direction everyone's chasing and it just feels so bad if you don't chase the same dream no matter how hard it is really

and in reality we feel blessed and happy to to say the exact situation because we don't have any serious competitors from startups to be honest well there's everyone is one and they seem like a pretty spectacular failure right you mean you may launch the lot of money

and a big team mobile partnership and a subscription fee and time magazine and all that stuff and it doesn't seem like that has gone very well yes so I said as if right now I don't think we have serious competitors from startup and then when we talk about competitors obviously there's

Apple there's there's every big companies all there including OpenAI first of all I think this is good for us because it validates our direction it's absolutely correct and I also are curious about what are going to be the definitive route for the generic agent technology because different

people in the industry might have different ideas right there are still debatable state there is no evil for agent systems yet there's no like very good evil yet and there you can see a lot of different research houses and companies trying different routes obviously there's API routes like

GBT's which doesn't really take off there's newer pure neurosynbolic routes there's a hybrid routes there's like all this multimodality so we're still in the face of everyone trying their own recipe and hopefully that can become a definitive recipe including Apple I think the benefit for Apple

to do that is that yes they understand the user better much much better than any any companies out there and they have infinite money theoretically infinite money and they have the very close ecosystem the way that they're rolling this out is that they have this SDK called App Intent right so

different companies or app developers need to choose to enroll or not enroll with that to have the new series control stuff I guess my relative advantage as a small group as rabbit is that we move fast yeah we move fast and we keep growing I think if we put all the cards on our table

we have this spectacular launch we we are the most sold delicately hardware yet and we have make good profit we fix all the day one problems and the company actually caught trouble the size so we're growing we're moving fast and now we drop this I think like you said put a marker between

today and yesterday I think today I can say a lot of things that you can do on our one you cannot do on an iPhone right I believe eventually everyone will be able to come to the same solution that all the device can do same kind of kind of like similar stuff but I firmly believe at least

this remaining half the year or the Q4 of 2024 and probably the Q1 2025 it is still a game of you have something that they don't have versus you have you guys all have a similar stuff who don't matter so I think relatively we have a good six to eight months ahead of start like we have our

little room here but obviously I also believe when big company wants to kill startup they have a million way to kill you that's just a reality I think people keep talking me to me and asking questions what happens if the risk is too high what happens if the company dies I really

don't think that all these questions matters because we're on this course we're gonna we're gonna see the end whether it's good end or bad end and I don't think like any answer to this question will change our course to be honest I can go to here and tell you and be a crybaby like this is

super hard this is impossible everyone the answer can kill us easy or the YouTube reviewer can kill us by posting it doesn't it doesn't change the course because we are doing things we're launching we're shaping things right we're moving forward so it'll be interesting to see what Apple came from

I was on the Apple iPhone upgrade program so I automatically get a new iPhone every year by paying the same monthly fee but I really don't find any reason to upgrade that because people are talking about rabbit being launched too early now you have the company like Apple if you go to the what is

that called a sunset bullet war in Los Angeles where it's close to here or I guess mission string in San Francisco you go to any major cities you see this gigantic posters a billboards that I Apple puts there right iPhone 16 iPhone 16 Pro what are the other lines underneath it says Apple

intelligence is it ready is it out no yeah let me talk about growth for a second you mentioned your quadrupled I'm guessing you mean by employee size yeah you told fast company last month the R1 is only being used daily by 5,000 people it's at higher lower than you expected first of all

you saw that article from I guess birch no it's fast company that's what it says yeah no yeah but there's no but there's a birch says R1 only has 5,000 users daily which is from which is from the bird that's a quote from you what I said there can be missing to operate in

what I said is that if you go look at the data doc right now you probably will find 5,000 people using R1 at least 5,000 I'm just gonna quote you fast company lose said right now around 5,000 people using R1 daily I said I said it can be misinterpreted okay yeah first of all I think

we saw a very steady growth of all the people interacting with R1 and each time with new features there's gonna be more people using it I gave you some numbers that I want to throw to you and maybe I can share like a very detailed like usage sometimes in the future first of all there are about

5% people that they have there R1 they're not happy that return less than 5% sure which is a very good number and I think the top features that people are using are asking questions and visuals and visions and all that and we really are hoping for people to discover more use cases but

unfortunately we have like 4-7 apps on the connections right that's well the bottom neck so if you check for the total query most of the cases you ask a question you forget about it right so it's not about how many times you ask R1 is that it's about what kind of tasks you ask R1

and is R1 actually gonna help you so I guess yeah where unfortunate it seems that that's a missing interpretation so I would say number what's the what's the daily active number well will it's either correction tomorrow what is it I will I will go back and get you a very

accurate number but I can tell you yesterday our server actually crashed so I think is it double is it 10,000 is it 20,000 oh yesterday our cost actually let me actually let me check right here because I can check right here this is why I like founders on the show it's this is this is why

I love having a founder on the show okay so past one day is 33.76 thousand okay so almost 34k yesterday 34,000 okay yours yesterday okay yeah what percentage of your sales is that yesterday yeah 33 33,760 people what percentage of your total sales is that I think we

delivered more than 100,000 unit and that should be like around 33% 34% sure that makes sense and that I'm assuming yesterday because it was a launch of land playgrounds this is a big spike yes what were the days before that so past two days 52 0 6 so if you minus 33 that's another

20,000 wait I'm sorry I don't think I fall you said numbers so I don't think I followed them past two days said again so past two days 52 0 6 that's the total of two days correct okay and one days with the land playground one is okay like I was just correct so you're saying it's 5,000

active users at any time not daily correct okay and then you're getting about 20,000 users daily and then we'll see if that's up because of the land playground correct then there's so article by the verge just use that title 5,000 which is wrong I can tell you that's wrong that's very

wrong that's missing information right there well you tell fast company and then we will update it but we ran off your quote in the magazine so we feel good about that he wasn't there and he he or she that journalist wasn't there and that's not what I said in a quote okay

we're gonna have to take another short break for a minute here we'll be right back sport for this show comes from arm have you ever thought about the technology that makes this podcast possible to listen to on your phone in your car or on your laptop and then there's

the data centers that make it all work one company is at the heart of it all it's the same company that powered the smartphone revolution and is helping define the AI revolution that company is called arm arm designs compute platforms for the biggest companies in the world

so they can create silicon and solutions to power global technology arm is proudly nasdaq listed and became a nasdaq 100 company within a year of its IPO arm touches nearly 100% of the globally connected population 99% of smartphones are built on arm major clouds run on arm as well as all

major mobile and pc apps now arms engineers are tackling the insatiable demand for compute and power efficiency that AI is creating AI enabled arm CPUs are able to provide the compute platform for the global AI revolution in the years to come but for now relax and enjoy this podcast

it's very likely running on your own arm powered device visit arm.com slash discover to learn more support for the show comes from toyota for many of us driving is just what you need to do to get from point A to point B but why not think of it as a reward instead make it an experience that

captivates the senses by driving a toyota crap the toyota crown family comes with the quality and reliability that toyota is known for along with bold and elegant exterior styles the toyota crown sedan has an available hybrid max power train with up to 340 horsepower and comes with

an available bi-tone exterior finish to help you stand out on the road and the toyota sycnia gives you the space you'd expect from an SUV with a stylish design unlike any other whether you're a daily commuter or weekend road warrior you can make any drive a thing of beauty with a toyota

crown you can learn more at toyota.com slash toyota crown family toyota let's go places support for this podcast comes from huntress if you're a small business owner the threat of hackers isn't just a threat it can affect your livelihood small businesses are easy targets for hackers

and huntress wants to give businesses the tools to help huntress is where fully managed cybersecurity meets human expertise they offer a revolutionary approach to manage security that isn't all about tech it's about real people providing real defense when threats arise or issues

occur their team of seasoned cyber experts is ready 24 hours a day 365 days a year for support they provide real-time protection for endpoints identities and employees all from a single dashboard their cutting edge solutions are backed by experts who monitor investigate and respond

to threats with unmatched precision now you can bring enterprise level expertise without needing a massive it department huntress can empower your business as they have done for over 125 thousand other businesses look them handle the hackers so you can focus on what you do best

visit huntress.com slash decoder to start a free trial or learn more welcome back so you heard all that back and forth about rabbit's daily active users and CEO Jesse Lewis saying he would get back to us for the better number ask the company to clear it

up and it turns out with Jesse actually said to fast company was that at any given time rabbit has 5000 users the fast company article has been corrected will correct ours as well and will use Jesse's number of between 20 thousand and 34 thousand daily active users which is still

substantially less than the hundred thousand r1 units sold that's where we left off let's jump back in now that we have the number what we'll run it but my question to you is you've got to sell more r1s you've got to get more people who've already bought them to continue using it and you

you are in fact whether or not apple intelligence has arrived or not it will arrive in some fashion in the coming weeks there's a report just a week or so ago that johnny ive is working with sam altman and open a i on a hardware device something will happen with humane something will happen with

google something will happen with samson as that universe of competitors expands it feels like the core technology you're betting on is being able to automate of the mc with a large action model right you're going to open up user sessions for people in the cloud and then your lamb is going to go click around in the web for them and that will get you out of the challenges of needing to strike API deals with various companies with other kinds of deals copyright deals with various companies

whatever you might need is that durable right the idea that this will keep rabbit away from needing all of the deals the big companies will just go pay and get because that's the thing that i i i think about the most i can i can think of 10 companies that came up with a technical solution

to a legal problem and even the technical solution was amazing the legal problem eventually caught up with them yeah where confident that this technology is the current technology route that it will work yeah and i haven't yet to see another approach that actually make any generic agent system

work in any other manner that doesn't mean that we're locked in to one technical pass if you talk to any company it's not probably not a smart idea to say hey we we just bad this for next 10 years the technology changed so fast you have to adapt but right now i think we're off the good start

we launched a concept with playground with free of charge that you can explore so that we understand how this system can be improved in fact i believe the speed can be improved very fast but we're not here to say hey we we stuck into this we do have like we do have patterns about this but we're not

saying hey we think this is a correct pass to go i don't think anyone in the AI in the industry can give you very definitive answer but i hate if you just do this here's a structure this gonna guarantee you the best result in a long run i think that's not a good way to think of it

but yeah i agree everyone in the industry are experimenting something new and a lot of company that we saw are gonna like you said running to some sort of legal problems you know there's music generation platforms there's i mean this feels like the story of the AI industry probably right

there there's like a youtube training video can be used by this or that you know there's all sorts of things like this yeah but i think it's not just the builder are adapting the industry are gonna adapt to the builder too at some point there's gonna be a conclusion that okay this is

a new policy this is a new kind of like terms that we do are you are you building to that goal i think again this is just the kind of the big question i'm thinking about all of these things basically every AI product is a technical solution that is ahead of wherever the legal system is

or wherever the business deals are it's some point Spotify might show up on your doorstep and say you know we're not gonna allow agents it has to be a human user and we're gonna change our terms of service to say it has to be a human user do i have to say it whoever might say it

are you ready for that outcome are you can you do you have the budget socked away to go lawyer up and fight that fight no at the moment we don't have the resources to fight that fight and at a moment that's not a real threat to us because they said we're too small fair enough when do you think the

turn hits i don't think that it's a dead end for us right like no i'm just saying when do you think it's a turn when do you think that's becomes a conversation about whether you can have agent you can use it as a human users yeah that's that's exactly what i'm talking about like i don't think

that they are not willing to change their terms sure and i i think unlikely they're gonna put terms like it has to be a human it cannot be there's a lot of automation tools out there already right so like where there's no turning back i think what they would like to work with any

companies including us is that when they see a popular demand from this new kind of agent technology they want to charge for it we ask our user and us to pay for them and that's a business deal that's more like a money terms that's what i can see but yeah for now we're not breaking any

of their terms and agreements right and if they change the term and agreements tomorrow we'll we'll take a look and we'll see how we adapt but the agent is out there yet already you know it's there's a lot of agent running already so i think there's no turning back and it's very

unlikely they say hey we're gonna stop like agent using our services that's not a that's not gonna happen let me end with two kind of big thinking questions and i forgot to ask you the decoder questions think on the the longest time when you can let's assume everything works out and

it's all solved how much time and money is it going to take before the general purpose agent you're trying to build is a hundred percent reliable and can just do all the things that we all imagine them being able to do i might have a different opinion here okay i think foundation models like

opening eyes you know obviously they're raising for crazy amount of money i think we take benefit from what they've been worked on right because their primary services is selling the their models as api's which saves a lot of money we don't want to recreate the wheel retraining like an OM

it might not as scary as a lot of people might think i think there's there's a huge gap between converting the latest technology into a piece of product versus pushing for a more advanced technology obviously i'm very proposed to to do like high-end research you know we we want to have like a

research house here set up as the same scale as like opening an enemy even though like they already like far far behind but i think what we're trying to do right now at this current scale because here's the money we have right we don't have like one billion dollars we don't have like

two billion dollars we have this very limited budget is that how can we convert the latest technology and research and build to a product that we can ship early and collect feedbacks and learn from it so a lot of people have different definition of a g i i don't really talk about this

term because i think so many people have so many different definitions for it but i do think that a i understand what you say and a start can help you do things and maybe you know here we're talking about virtually help you click buttons and stuff there are a lot of companies doing human

oil i enjoyed that they're actually giving a hand a leg for for the a i to do things i think it is the entire humans effort and a lot of the resources can be shared instead of each company has to go race for it is amount of money and take that amount of time to achieve the same goal so

it's really hard to say but we know we need more money and resources that's for sure but i think you see how efficient this team has been performing you know from seven people 17 people till today like we risk obviously much less than humane or any big companies all there

i think it's actually one of our advantages that we can do things in a relatively cost efficient way and fast yeah yeah timeline wise though again assuming everything goes your way is it a year from now that it you can build on all the foundation models and all the other investment

and this thing just sort of does whatever i ask on the web is it five years what do you think i think the a m model will get very smart very fast but i but i think we're talking about a generational shift i think obviously we don't want a 2024 piece of technology operating on e-base

website which is basically designed back in 1990 right so i think a lot of the infra needs to be refreshed and the the biggest gap as i can see here is productionized so i think in our role map we think that it's very likely that we can get all this separated piece

of technology we have like land playground teach mode and rabbit OS at some point maybe next year kind of like emerging to a new rabbit OS 2.0 and that actually will push a huge step forward towards this generic goal but my general take is that a m model is smart enough

but the action part is a lot of infrastructure there's a lot there's a huge gap between research and productionized so that's what we learn yeah so i would say like i'm very optimistic in the three years term but i think like i said right now and starting of next year is everyone trying

you know different different approaches and we'll see which one works but i think we're confident on the approach when i'll wear take right now yeah and then i just want to end and ask about form factors obviously the rabbit is a very distinct piece of hardware people really like

the design we've seen just a lot of interesting glasses lately the idea that we're gonna we're all gonna work cameras on our face and someone's gonna build the display do you think that's correct i was wearing the meta-ray band's estranged i was like why would i wear these all the time i'd

rather have a thing yeah i'm not against any form factors in fact i really think that there will be a lot of form factors but when we were trying to design our one the reason is that we know it's not it's not gonna be a smartphone because we know people gonna do a lot of other

things on smartphone which the current AI cannot do so we deliberately avoided smartphone factor talking about pins with lasers and glasses i have different comments for each phone factor because there's no universal rules here because for let's talk about pins right so like i think my

general pushback for making it as a pain with a laser like humane i think first of all i think it's really cool but i think you know it's too risky you are trying to offer a new way of utilizing your technology used to have user use software and that's already new to them and you

don't want to just introduce a sci-fi kind of type of gear so two new things stacked together that's too risky so i if you look at our one is like very familiar design you know there's a button you know you're gonna push you know we'll probably can scroll there's a screen you can look at things so

the R1 phone factor is very conservative in a sense that it directs the software it's just like you know people haven't figured out how to interact in virtual world and all of sudden back in 2016 there's like 200 different companies making goggles right and they all fell so i think i'm very

very conservative on the hardware phone factor talking about a glass that's a different story i think your skull actually grow to fit the frame now they are way around because i used to wear like prescription frames i know the pain like your skull is growing to fit the glass frame now they are

way around so i think there is really no generic fit on the glass frame i was having fun with my design team joking i'm like maybe if we do the glass we probably do the dragon ball style like the power power reader or whatever it is but i'll google glass one factor but but i'm really like

i can't wrap my head around i'll have to put like a frame you know like that doesn't fit yeah we'll see i think i think even the current smartphone i think is perfect i really like the the state of a glass or screen phone factor but the real problem here is not about the phone factor

well the problem is about the the apps right because now we see all this agent technology AI stuff they're doing things that app are doing and they're doing things that apps can do so i think the problem is with apps i forgot to ask you the main fucking question is my fault you've had a

number of startups you've done a number of things you have a big idea here how do you make decisions what's your framework for making decisions i'm a very intuitive person i like to trust my intuition on big directions like you know what's gonna happen uh you know in a long run but meanwhile i'm quite conservative that i hate to predict things so i think people when they people when people you know replay this episode they will hear probably i got really tricked by some of your questions

it's just my brain couldn't work for like predictions is that i don't like to make predicts like what happens if this happens that happens what do you think like i think when i manage my team i tell people like we make decisions based on current fact and we find the best solutions to it if you

spend too much time at least if i spend too much time think about what if apple knocks on a door what you're gonna do and what if this a happened then b happened then c happened what you're gonna do most likely you're gonna get a different strategy right because if you think about

if b is a solution to a when a happens you just do b but there are other type of people they're like hold on have you ever thought of when a happens then then d happens then e happens then f happens are you still gonna do b if you think in that way probably not right so i just choose not

you predict a lot of what ifs and i make short clear concise decisions based on current fact and in fact if you do the recap for what we're launched back in the c es it was probably the best timing the price is probably just right the color probably just right right and the decisions of

not negotiating spend six months negotiating with t-mobo is probably just right you know like i make on current decisions and that's my style and i talk to people everyone talk to me i told my team you know everyone in my team they can find me anytime talk to me anytime i spend a lot of people

a lot of time talk to my people and it's just we're in general just a very real team like down to earth and i really don't like some of the other type of a startup that they kind of like spend too much time enjoy the feeling if you understand what i'm what i'm indicating but there are a lot of

people that they they say oh i'm a founder i'm cool you know like no i've you know i've grown enough to get rid of that well i probably the same way as if i'm 21 22 but now i'm like 34 startups really tough it's a war it's about survive right it's really really tough and

it doesn't really matter if others want to do something like whatever you have to be survived and just survive by your own is tough in any sense so that's why you know a lot of people ask me i got asked a lot like okay what if they they do this what if they do that well end of day there's

nothing you can do you have to do your thing and they will react to it i think it's fair to say that with rabid other startups like us biggest company like apple they react to us they react to us in a very hustle way right like very unusual way that they have this new phone but all of things

are still not there well we're making very small dent right but that even doesn't matter i think for us we care about our customers one thing i want to say is that yes there are a lot of misinformation there are hates there are all that feedbacks criticisms if you talk to the r1 user

they're happy that's what i care that's what i care otherwise you know there will be a lot of returns there will be a lot of refunds we have less than 5% return like put that term in any consumer market electronics device it's a good it's a good benchmark and we're gonna keep

releasing all the stuff and in fact we pushed 17 OTA within five months the other company pushed like what three two three four five companies so i think i really hope people can see us as you know we're a bunch of underdogs yeah our solution isn't perfect but it is david versus golly on day one

because it's a reality and don't expect perfect stuff from us because we are not perfect right we raise a very little amount of money and we're a small team but we move fast what we can guarantee is that when rabbit shows you something you probably couldn't even find somewhere else just like the

hardware just like the playground or even the very junky day one version of lamb where the first company that has apple music can be streamed to our device yeah right this is absolutely because you're opening it on the web yeah yeah i mean i don't get legal documents to my door maybe

i will get one but maybe they think we're too small but yeah we do things in our way i guess that's what i want to say we're really down to the ground team like that's my style jessie thank you so much for coming to coda and i'm still getting to answer these questions

i really appreciate it yeah thank you so much i'm here i'd like to thank rabbit co jessie luke for taking the time to speak with me and thank you for listening to decoder i hope you enjoyed it if you'd like to let us know what you thought about this

episode or really anything else at all drop us a like you can email us at decodertheverg.com we really do read all the emails you can also hit me up directly on threads i'm at reckless 1280 and we have a tiktok check it out it's at decoder pod it's a lot of fun if you like decoder

please share it with your friends and subscribe over your podcast if you really like the show hit us with our five star review decoder is a production of the virgin part of the voxamu podcast network our producers are k-cocks and mixed out our editor is cali right are supervising producer

is lame james the decoder music is by break master cylinder we'll see you next time cloud database ops can be tricky several lines makes it easy pioneered by several lines the sovereign debass concept is a new way for corporations with

multiple business requirements and to deploy their workloads in mixed environments to exert more control over their data stacks it enables ity ops teams the ability to deploy and orchestrate databases in all environments removing lock-in risks and giving their organizations the orchestration

benefits at scale that they used to get from traditional debates but without its portability and access trade-offs learn more as several nines dot com slash vox and we're back with camera presents secret sounds work edition color guess this sound

it's a mouth click so close that's actually publishing a website with canva docs next color uh definitely a mouse click nice try it was sorting a hundred sticky notes with a canva whiteboard we also would have accepted resizing the canva video in the 10 different sizes let's know yes way one click can go a long way love your work at canva dot com