[00:02] You know, when you think about American manufacturing, mobility, technology, uh I'm thinking about Rivian. I think that company has a lot to say in in those in those areas and and CEO RJ Scaringe joins me now. RJ, thanks for joining us. [00:15] Really appreciate it. You know, before we get into all the meteor topics and we I I don't think I know the actual actual answer to this question, but why did you start Rivian in the first place? Like what was missing [00:27] marketplace? >> I was so enthusiastic about cars, but they had so many challenges associated you know, if you think about it from a climate point of view, from a air [00:39] quality point of view, from a global conflict point of view in the context of so I was like, boy, I want to help solve this so that we can have these things you know, personal transportation, but do it in a way that's [00:52] that's not so damaging. >> And I know Rivian has to now with electric adventure vehicles and you know, the successive R1, MotorTrend Truck of the Year, all these different things. Uh an expensive truck [01:05] and expensive to make and now cut to now we have the R2, right? So challenges, but also build upon some of >> Yeah, I mean R1 is our our flagship product. It was [01:17] really like the handshake with the world and R2 is really the scale product. It's the mass market, takes all the brand essence that we had in the first you know, condenses it into a slightly smaller package. It's an mid-size SUV. [01:31] But it's just as exciting, just much, you know, much more um more people can access it. >> Yeah, I I had a chance to drive in the R2 and I I recall being I mean perfect size in terms of like it's actually [01:43] bigger than I thought inside and for my sort of I was sort of shocked shocked by composed even without air suspension, right? Real steel springs. Uh but that for you to a smaller bill of materials, cheaper to [01:57] build. Uh and is that sort of the future then R3 and then maybe is you're going keep it more accessible? adventure, but [02:10] that can be much more accessible in terms of price. cost out, but we wanted the car to still feel big on the inside. Then R3 takes that further. So it's a still smaller than R2, uh, even lower [02:23] priced. And as you said, the next thing beyond R3 will be R4. Your naming convention's pretty predictable there. But, um, the the way we're thinking about it. >> It's kind of push more mass market, [02:36] >> There's often the assumption that as you go to higher volume or to mass market, a product has to lose its personality or its soul or its excitement. And it's what we're trying really hard to do is yes, the the vehicle's becoming lower [02:50] price. By virtue of that, it's becoming much more accessible to a much broader audience. But we want it to be just as exciting. And so, um, that's that's a lot of work. We have to do it with less content. So R1, as you said, has [03:02] air suspension, it has electric hydraulic damping. Very complex in terms of its chassis and its architecture. R2's a simplified architecture. We took a lot of cost out, but it's still something you can have [03:15] many different ways. And and lots of little features still has that like a wow factor and excitement factor. >> Talking about software and AI, you know, how do you see that kind of working with [03:29] more of a car company or an AI company at this point? What's happening there? >> We think that will be obviously a a segment and we we are have a We're But we also think it's really important for personally owned vehicles. And so [03:42] you can imagine in the not too distant future, you have airport and drive itself back home. It can go pick up the kids from school. It on a road trip, you can be watching a movie in the back of the [03:56] car completely relaxed. Um, and so the the The behind autonomy is it it just completely opens up the aperture for how you interact with the vehicle can use it however you like. You can use it to be on your phone, you can use it [04:10] to rest, you can use it to read a book. Um you can use it to be in conversation. And so we're we're very focused on that. Uh and we're just at the beginnings uh to launch our point-to-point capabilities, which is where you'll type [04:25] the address and and the car will completely drive you to that location. Um but all that is on a staircase to get to you know a full level four capability, which is where the car can drive itself empty. It can completely [04:38] you know operate itself without any you know human um interaction. To train you need to collect lots of data. So you need a fleet that's big enough that can cover a broad spectrum of corner cases and identify all these the [04:52] long tail of unique events. Uh you need to control the perception platform that's on it. So you need to have complete access to the raw signals potentially lidars if you're using that in our case. Uh soon we're we're [05:05] Um and then you need to build uh foundation model around that. So what we do uh uh we build what we call large driving model, LDM. And our LDM is this collection of all the information, [05:18] all these insights that we're using to drive the vehicle. And so our point-to-point capabilities that are coming is this end-to-end trained model that's built nothing like how autonomous systems were built [05:31] systems were built you know up through call it 2021 2022. rules-based systems and then the new neural net LLM type systems? You know that's sort of how humans think. Like explain to me just as simply as possible [05:45] system and how we think in terms of a neural net. What does that mean exactly? simple uh simple situation. If you're driving down the road, there would be software that would be you know describing to stay in lane so [06:00] go over the dash line. Uh there'd be software that would be of you. Uh but it'd be very deterministic. You'd rules that says, you know, based upon these inputs, these, you know, the [06:14] the inputs from the uh from the perception platform, don't hit the white line. And so where it's fragile is when you get to complex situations. So you get to a roundabout and suddenly [06:27] in a roundabout. Uh and you know, if you look at like early autonomous vehicles, let's call it like that 2015 2016 time frame, these more complex interactions because you were trying to describe something [06:41] that was so nuanced. And so in contrast uh there's none of that rules being written down. So they're, team is not describing at all how to drive. And so when the car does a [06:54] that we've we've extracted from, you know, many many many millions of miles of driving that through fleet. And so these are all emergent behaviors. [07:09] >> So like if you'd never been to New York City and you learned to drive in could drive here. You'd be like, wow, this is a bit harder. It's a you know, neural net with new information being like, okay, those yellow things are [07:24] really aggressive and like suddenly you have a new emerging model, but you'd be able to operate. And um from a self-driving point of view it's to do is collect a broad enough set of data from enough locations. So of [07:37] by virtue of having consumer vehicles, our vehicles are everywhere. So we have coming from Seattle, we have data coming from Miami, we have data in in Missouri. from Miami, we have data in in Missouri. And that gives us a really diverse set [07:51] of data and set of driving behaviors that we can then use to build this generalized very large parameter model, you know, with with all the appropriate waiting. You know, think of it as a big multi-dimensional [08:03] knowledge of driving. >> Imagine dropping that >> driver into New York City might have some issues how complex it is. And where do you see robotaxis going at least for Rivian's [08:17] >> With the introduction of robotaxi what what I think will happen is it creates an opportunity for people to decide more people to decide to not own a car. people to decide to not own a car. And you'll see that [08:30] But I don't think it's the end of car ownership. I think maybe it goes from 95% to 90% but I don't see the whole country deciding to use ride share services and like unplug and plug in car [08:44] seats simply because we have the wealth and the ability to do that. there's certain things in your life that you just want to own. So robotaxi own still own the vehicle and it's [09:00] more useful in your life and maybe you go from a two or three vehicle household to a two or one vehicle household. But we still see vehicle ownership being a key part. Then in the middle there's going to be all kinds of innovation. [09:13] going to be all kinds of innovation. There's going to be know you and I maybe close friends or neighbors and say hey why don't we own a work and it comes back and then takes me in the middle of the day. [09:28] you know you know spouses can can use vehicles differently so I think we're to emerge. And we'll see that in I think in the to have these two bookends because they're the things that are most [09:40] we know what Uber looks like and we know what car ownership looks But I tend to think it's probably going to change by don't know. And so from Rivian's perspective we're participating on both [09:54] consumer brand. We still deeply believe in think owned level four is going to be really interesting. [10:06] interesting business and so we did a partnership with Uber. It was a 1.25 billion dollar deal and we're creating a robotaxi version of R2. And I'm I'm enormously excited about [10:18] for places like New York, you know, San Francisco, downtown Seattle, like these more urban dense environments, it's a it's an incredible unlock and um in the experience of being in a car that doesn't have a driver, there's a lot of [10:32] You you don't feel uh you there's a sense of privacy that's >> I think it's really cool also with regards to people who elderly, people who have sensory issues, having the ability to both have a car drives you or [10:45] >> Ride hailing is going to be great for elderly. Yeah, for sure. with Doug Field. That's a big coup there. I mean, I've I've I've met Doug a few times at at Ford events previously at Ford. I mean, the guy is just a [10:57] remarkable person. Talk to me about what you think is going to happen now with a where do you see that company going? >> Doug joining is is outstanding. It's um when I first started the company uh and I was thinking about what the you know, [11:10] how the company would play out, I would say uh to to folks that were able to work with somebody like Doug Field in building this. to make a long story short, I ultimately I I've known Doug for years and so I [11:24] I'm thinking about you know, how I want to design the organization You know, the company's growing quite a bit. Um I'm looking for somebody like you. Uh and I said I'm like help me think that [11:37] somebody like Doug. And he said, oh, that's interesting. And so instead of getting somebody like Doug Field, I'm going to get the Doug Field. >> Yeah. Well, one last question. I want to talk bigger picture [11:50] is hardcore into it. It's part of the the future of the business and where the and look at sort of the where we are as a as a civilization and a society, you know, are you concerned that some of these frontier models are just moving [12:04] too fast or no, we shouldn't we shouldn't hold them back. We we need to um but we don't want to we want to be careful too. How do you do you How do you square that? There's just such a broad spectrum of incredible benefits [12:17] that has me really excited and as a product person, the ability to develop better products faster accessible broadly in terms of pricing [12:30] uh is really interesting. And so you can sort of dream for a moment where you could imagine like incredible vehicles, incredible products that are more accessible, a broad spectrum of consumer goods that [12:43] start to become cheaper because we're able to make them more efficiently. Um people being freed up to focus on higher value added work and types of things that they're passionate about versus [12:55] some of the things that are frustrating and not so much fun uh in current jobs. I know so many folks within Rivian they're like, how this thing I used to have to do that was so frustrating I now don't have to do. I can automate that. [13:07] there's lots and lots of positives. Now there's challenges with a new technology you know cybersecurity challenges in addition to being able to manage and control AI and be thoughtful about you know what's um [13:22] in. Uh the fact that job functions are going to change and there's going to be required for us to rethink uh like how we're educating our our you know educating society broadly but [13:35] even reskilling our existing workforce. Those are challenges um but I think the the positives definitely are you know far exceed the you know the challenges. could kill us all at some point. >> No, I think [13:48] you know AI's a it's it ultimately is a tool. So we have to recognize it's not a uh sentient being. It's not um you know it's not a it's capable of it's it's all based upon how you incentivize the model. So it's [14:03] it's capable of doing bad things if it's directed to do bad things but it's not it's not it's not a sentient being. So I I think there's like we have to break like recognize we're building a tool and treat it as a tool. Uh it's not a [14:17] it doesn't have feelings. Uh we could train it to behave as if it had those things, but but we need to continually look at it >> [laughter] >> Yeah. [14:30] for joining us. We appreciate it. >> Thanks.