Building an autonomous car utopia with Rivian's RJ Scaringe

20:51Ellis:Have you seen that as well in the auto industry? Like going to softer colors? Just, I guess, a bit more conservative take on like, I don't know, Maybe it's on the one hand more naturalistic, on the other more blah in some people's minds, or is it cyclical? Yeah, I think colors move around so much. And I think, though, in general, we do see – like the Vastorite customers always like the gray, silvers, and whites. And interiors, it's like light – like a white or light, light, light, light, like gray or black color. We've done some interesting colors in both – we did a green interior, which is amazing.

21:29but the take rate was quite low. Really? Yeah. You'd think that would be well-received considering how you... No, no. It was well-received. It's just... Even people who liked it were like, yeah, that's awesome. I love it. It's so cool and eclectic. But I'm going to go with the black. What is the Rivian color? Is it yellow? Do you guys have a standard color? Standard? No, we don't have... We have like a brand color. We call it compass yellow. It's a color of... You can see on like on our logo on the sides of the of the emblem, there's a little yellow like accent. That's our compass yellow.

22:03Ellis:Yeah, I was thinking about y 'all the other day. I was down in San Diego for a minute, drove by the shop. And then over Thanksgiving, I was up at the Peterson Auto Museum and got to see the OG Blade Runner spinner, which was pretty sick. I did not get to see the Fifth Element cab or Johnny Cab from Total Recall, unfortunately. Have you been to that museum recently here in LA? No, not recently. So what's there right now? A whole bunch of things. I guess it rotates if you really want to see it all. You have to pay to upgrade to the vault, you know, to see the whole Jay Leno collection or whatever there is down there.

22:43Ellis:But yeah, it was just making me wonder. I mean, I guess the car designs themselves are so cyclical as well. And I'm from Detroit, so I feel like I'm heavily steeped in concept cars and auto shows. Were there any old designs that really inspired you guys? I mean, for me personally, I've always been inspired by a lot of early Porsche stuff. It's like I grew up loving that, but it didn't really inspire Rivian. It inspired me to love cars, but it didn't inspire me in terms of the form factors of what we ultimately built. I think for us, we found a lot of, like a lot of our inspiration was around wanting to find these balances of things that you don't typically find together.

23:24So something that's really capable, but also approachable and friendly. You know, it's probably manifesting most clearly in the way we approach the headlight graphics on the car. So if you look at like the front of our R1S with a quad motor, it's, you know, it has over a thousand horsepower. It can out-accelerate any Porsche or any Ferrari, but it looks really nice. It looks really friendly. And a lot of vehicles that have that kind of performance, typically they start to look like they're really angry or pissed off. And we always subscribe to the idea that you can simultaneously be really powerful and really capable as a vehicle but have an approachable look.

24:04Ellis:Yeah, I think it's just endlessly fascinating how cars have faces and give off vibes. Maybe you've seen that like old Volkswagen commercial about people who look like their cars and whatnot. But am I like a mega sci-fi nerd or have you heard of the designer Sid Mead? No. The concept artist? Yeah, of course. Yeah, I know him. Yeah. God bless. Okay. I'm so happy we could talk about Sid Mead on the call. I think one of the things I love about him and, you know, for those who aren't familiar, I mean, Blade Runner is definitely on the dystopian side. but I feel like he's more associated with future utopia, whether it's like the world of Mass Effect or some of the things you might have seen in like the above space colony of Elysium or otherwise.

24:46Ellis:I love that he's representing. I mean, you don't see this a whole lot in tech where we tend to go straight to the evil dystopia. But maybe in cars, you could portray more of that beautiful, bright future. I mean, is that a factor at all or at least staying away from from stuff that might connote the wrong ideas? I mean, even with how you portray the autonomous stuff. I mean, it comes back to your question on color. Design language, color, material choices all have, you know, they reflect a certain tonality. and what we find is the design language and design decisions that speak to something that's more calming or more approachable is more consistent with certainly what we want as a brand, but I think culturally and from a societal point of view, a lot of people are drawn to that, something that feels timeless, it's not too aggressive, it's not trying too hard.

25:45Yeah, and I think as you go like the other end of the spectrum, If you go like very dystopian, it sends a signal about maybe how you see yourself or how you see the world.

25:56Ellis:I mean, clearly a lot of people want to look evil. I feel like whether it's with the black Mercedes Darth Vader helmet look or the Cybertruck's entire vibe in general. Yeah. So this is interesting. So I will tell you guys something that's not that well known. But Rivian offers a black color. Oh, you never see that out. Well, I'm sorry, Alex, but this is the truth. So it sells fairly well. It's not the most popular of our colors, but I've tried since the early days to discontinue that color. And ultimately, we keep arriving back at not doing it because there's enough people that want a black SUV.

26:37And the reason for it is not that I don't like the color black. It's just that a black car, especially a black car with black wheels and blacked out trim, has a certain tonality that's inconsistent with how we want to feel as a brand. It feels too mean. So we still have a black color, but it's despite my best efforts. And like our head of design and I have this debate all the time. So I was – if you actually go look at our black, I was like when I agreed to do a black color, it's actually black with some soul. So if you look up close, it's got like a beautiful bluish purplish flake in it. And so in the sunlight, it has more depth versus like a pure black black.

27:18So RJ, you're talking to someone who just convinced his wife to get a murdered out XC90 2026. They actually brought the black trim to the US for the first time for this model. It was in Europe for a while only. So it looks almost like a Range Rover. and it was our compromise because I wanted a Range Rover. She thought it was, she didn't like it. She wanted the safety stuff with Volvo. And so I was like, okay, the concession is you let me get the murdered out version. So our head of sales, who's just an amazing, he's a friend and he's an amazing leader within the business. He has black on black and he has it in a matte finish.

27:56And it actually looks really awesome. I do agree. It looks super cool. but anytime I'm around him and he shows up the car it's like I know you don't like this color I'm like no Gary I love the color I just think the tonality is a little you know a little edgy but it's it's like an internal joke we talk about this all the time colors colors are such a funny thing you could go too far I mean in the Tesla direction which like sometimes when I think of Tesla I think of like Death Star or like when you look at how they do their events it's like very leaning into the Blade Runner stuff. Yeah. And I think it's actually, it's a good thing that there's brand diversity.

28:32The Tesla brand and the Rivian brand are so different and manifest so differently across every aspect, whether it's our retail locations, the events we run, the way our products look, design choices we make. And it's good because some customers are going to want that look, some customers want our look, and the world needs to have some choice.

28:51Ellis:It's clear you're mindful of the consumer sentiment. And as I'm sure you've read in the headlines, you know, there's so much hesitation and skepticism, especially across the broader culture of AI and autonomous. How are you thinking about that as you bring that to market now? I assume you want to do it a little bit differently than the others. I think it's perhaps the biggest technology-driven societal change we're going to experience, certainly in our generation, but probably in the fullness of history, it'll be like a seminal moment in human history. And so, in fact, I'd say there's two very unique moments in history that are happening in our life, in our generation, which is the first is the end of the fossil fuel era.

29:39And so if you imagine history books a thousand years from now or 2000 years from now, They're going to say humans lived a certain way for tens of thousands of years, meaning we couldn't travel long distances. We could only talk to people we were staying next to. We could only eat the things we could gather locally to suddenly living the way we live today. And that was built on fossil fuels. And that happened over the last 150 years. And those fossil fuels also had dire effects on the planet. And we switched off of those between year 2000 and year 2100. and that's like where we're alive right now.

30:14So that's going to happen in the course of our lifetime. And that's a huge shift. And so that alone is a big deal. And then to say simultaneously and by coincidence, as that was happening, you know, this other inflection point, which is we created artificial intelligence. And we're like, if this is a thousand step journey, we're on step two. You know, this is like AI as we define it today is three, maybe four years old. So it's so early and we've only seen it really manifest at scale through written word, images, and videos. The existence of AI in the physical world is in the very early stages.

30:53And I think one of the first large-scale deployments or embodiments of physical AI is going to be autonomous vehicles. And by the way, autonomous vehicles up until like 2021, 2022, were not actually built with AI. These were like rules-based planners. These were human described code. This is not to say machines weren't doing things that humans do, but it wasn't doing it through AI as we now think about AI in terms of like building a model or a neural net to control how vehicles are being operated. And so the whole massive shift that we've seen over the last couple of years of AI now completely shifting how we look at designing and building self-driving systems is like we're in the early days and so like i'm super bullish on how fast this is going to roll out the way i say this all the time is the next three to five years for autonomy are going to look so absolutely different in terms of rate of progress than the last three to five years in terms of autonomy um you know like like orders orders plural of magnitude difference in terms of the rate of progress and we're seeing this firsthand with our own products just because of like some of these big innovations like transformer-based encoding, which didn't exist, you know, like five years ago.

32:12So there's just like a very different playing field. So with that said, I think autonomy is going to be like, it's going to become as critical to buying a vehicle as having tires on the vehicle. Like the idea of buying a car without the ability to drop your kids at school or drop you at the airport or run to the grocery store and pick up an order, like that will just, like the utility of the product will be so much lower that people will be like, why would I want to buy that old thing? You know, it'd be like buying a phone that doesn't have Wi-Fi connectivity or the ability to like check your email.

32:49And so for like purely nostalgic reasons, you'll see vehicles that don't have autonomy, but otherwise the vast majority of the mass market will want to have it. And I think that's not like 20 years away. It's not 10 years away. I think that's happening in the next three to five years. And so you have that. And then you have AI in the physical world through the lens of robotics. And I think we often look at robotics and our imagination sort of stops at humanoid. But it's like smart robotics with reasoning capabilities and with dexterous capabilities that rival or beat human dexterous capabilities are going to have many different form factors.

33:27It's just like the physical world has an evolutionary breath to it, or the biological world has an evolutionary breath to it. I think the world of robotics will have some evolutionary breath to it as well.

33:38Ellis:I'm envisioning a pickup truck, Rivian, with a power arm that loads your two-by-fours for you. Maybe. Is that what you're picturing as well? I think every single task that a human can do today will be able to be done by a robot. Now, is that in five years, 10 years, 15 years, 20 years? That is a little less clear. But certainly in the fullness of time, in my lifetime, that will be true. Well, speaking of, I mean, I work with a lot of companies that are automating things that people really don't love, like let's say transcribing an entire interview. But I mean, going back through history, I mean, speaking of like the span of history, I mean, driving a car is probably one of the most fun things that's ever been invented, if not like arguably the most fun.

34:28Ellis:Yeah, yeah. Unless you're crossing LA at rush hour, but yes. Yeah, yeah. I agree, Alex, yeah. Where do you think it's going to settle with the stuff that people let the car do for them and the stuff that they still want to do? Because I think it'd just be a shame, you know? I mean, obviously there's a lot to talk about with safety and all these other things, but driving is so fun. Yeah, it is. Look, I'm a huge car enthusiast, So I agree with you. I think in... Well, it's just tough to balance that, right? Tech companies just want to get rid of everything. I don't know if that's true. But in the foreseeable future, call it in the next decade, maybe even two decades, I think we'll have autonomy will represent a choice.

35:11So customers will decide, do I want to drive my car or do I want the car to drive me? And depending on your situation, I think everyone will fall into – you may find yourself in both categories, one maybe more than the other. But there will be times where even for like the hardest core driving enthusiasts, as Alex said, like being stuck on Interstate 5 in traffic is not fun. And so you'd rather switch off, be on your phone, watch a movie, read a book, something. You'll see that. And then you're also going to have customers that want to, when you're on PCH and cruising up the coastline, it's beautiful.

35:49You want to feel the cargo around the corners. I think some customers are going to do that. Where it gets a little bit fuzzy is you move out beyond, call it, 10 to 20 years. And this is where a lot of culture and societal elements start to play. But there's no denying that the self-driving vehicle is going to be safer. And so then how does that get managed? does driving become more of an activity you do purely for pleasure, sort of like riding a horse you would no longer do for functional reasons, but you do it for pleasure, for joy.

36:18Ellis:Speak for yourself, RJ. Yeah, I mean, very few people, I should say, are using a horse as their method of transportation now for practical reasons. You haven't been to Silver Lake lately, have you? The hipsters are on some new shit. So the question is, does it like – take the United States. Today we have over 40 ,000 people die every year in traffic accidents. If you could reduce that by two hours of magnitude, if you can make that 100, is that the right decision? But you then – That's hard to argue with. With that, you lose the ability to drive. Yeah, it's hard. Insurance rates might decide that, right?

36:57I mean if the insurance companies see that statistically autonomous is just way safer, they underwrite that. and money and economic pressure actually pushes everyone into this future. Do you think that'll be part of it? I think that's going to be one of the driving forces, yeah. I think another big question, which obviously is debated all the time, is what does car ownership look like? So today, I'm just speaking to the United States globally, even numbers or the ratios are roughly the same, but the United States, there's just over 10 trillion miles that could be driven every year, which just for a moment, like soak that in.

37:29That's such a big number, 10 trillion miles. and about 99 % of those are in assets and vehicles that we own or a household owns, meaning all of rideshare, all of rental represent around 1 % of the miles traveled in the United States. And so as things like robotaxis become more available, do we see a shift away from owning cars and towards using robotaxis? Our view is you're going to see some, but I don't think car ownership is going to fall off that dramatically in the United States. And so what we think we'll likely happen to see RoboTaxi be a big, very interesting business. But personal level four, like the idea that my car can truly support my household, like it can drive my 10 year old to, you know, to baseball practice, it can pick up, you know, my partner from the airport, it could do all those things.

38:20It doesn't just sit waiting for me to drive it. I think that will likely become true. But if you need basic mobility, ride share on an autonomous robotaxi is going to likely be very, very cheap. And so it'll become the new entry vehicle. Rather than own a car as an entry car, you say, I'm going to use a ride share service that has a variable cost that's low, 50 cents a mile, something like that.

38:46Ellis:I smell a fight. RJ, you didn't want to come get me? You sent the car to the airport? Pick up flowers on the way to the airport. There we go. We'll figure it out. Tech can solve it. RJ, before this interview, I was watching your keynote at your all's recent Autonomous Day in Palo Alto, where you were talking about a lot of this stuff. And I heard you're calling this a large driving model. And I'm wondering, did you come up with that? How How many ad agencies were involved in coining that term? And what technically is a large driving model? Yeah. So the actual, the technical teams came up with it.

39:26And it's just...

39:29Ellis:What kind of laugh are we hearing right now? It's just the most honest description of what it is. It's a foundation model and that's large. For driving. For driving. So the team had a bunch of clever ideas. We just decided to keep it simple. Yeah, maybe it's helpful to give a little context on how self-driving has evolved. So self-driving started as a concept with some of the work that the DARPA challenge was doing. This is almost a little less than 20 years ago. And the way these early systems were done in self-driving systems up until the early 2020s, including what we launched with as a company, which you now could like characterize as these autonomous vehicle 1.0, people say AV 1.0.

40:17These systems would have a perception platform, so cameras at a minimum, but likely cameras plus radar and maybe LiDAR, depending on the capability of the system. That would perceive everything around the vehicle. It would classify everything. So it'd say like that's a car, that's a bike, that's a person. And then it would associate vectors to all those classified objects. So what's the acceleration, velocity, and X, Y, and Z? And all that information, when they get handed to a planner, and the planner would then make a whole series of decisions around how the vehicle should operate. But the planner is like a human-written piece of code.

40:54And this isn't like one line of code. It's a pretty complex code base that would try to describe, try to codify how to drive a vehicle. So human code, what even is that anymore? Come on. Yeah. So that was the way it was done. And then a big shift came along and you'll hear this called many things, but essentially like the, probably the most commonly referred to way to describe these more 2.0 solutions is an end to end approach where the, where you're, instead of having a, a rules-based planner that's prescribing how the vehicle operates to have a large parameter, like a big model that you're building that's trained off of a data set of vehicles that are driving.

41:38And so this large model is a neural net. It's a foundation model that is learning how to drive through observation of many, many, many billions and billions of driven miles. And so that's what we shifted to. Of course, we're not alone. This is what Tesla's doing. I'd say anybody who's serious about self-driving is now building these end-to-end foundation model-based approaches. And to do that, it requires a few ingredients. You need to control the perception platform, so there's no pre-processing of any of the, like, you don't want a camera from a third party that classifies objects. You actually want the raw camera signals.

42:19You want the raw radar. You want the raw lidar, if you have it. You want to feed that into a data architecture that allows you to extract interesting events, so triggered events. You want to take that offline, so it needs to be sent ideally over Wi-Fi to a cloud environment where you can then process it on a lot of GPUs and you start to build weights in a model that, and of course, understand what the model is. It has many, many, many different nuances and observations it can pick up on that essentially determine how the vehicle operates. And that approach, and this is why I said the the next three and a half, three to five years are gonna be so different.

43:00That approach is remarkably different than these AV 1.0 solutions. And so we launched with a, we had a mobile eye front facing camera. We had some of our own stuff around it, but it was very much a rules-based environment. We started re-engineering the system around an AI-centric approach, late 2021, early 2022. We completely threw away everything. Like we didn't use any of the original sensors, any of the original code base. It's a full new platform. and we launched it in the middle of 2024. And so it's built a lot with a lot more similarities to like an LLM. I say similarities, it's not the same as LLM, but similarities to an LLM foundation model.

43:39And so I think there's going to be more than one, less than five large driving model type foundation models that exist, at least in the Western world, in the United States. I think China will have its own set of foundation models for driving. I don't think many companies have all the ingredients to do this really well, but it's by far our biggest focus. And so we're quite bullish on it. As I said, we think this is going to unlock very high capabilities over the next couple of years.

44:16It's not just something you made. It's the privilege that you get to work with your hands. It's building something that serves a purpose. Proof that you have the grit to keep going. At Timberland, we understand you take your craft seriously, and we do too, which is why our products are built to the highest quality. We put in the work so you can perfect yours, with purpose in every detail and crafted with intention. Timberland, built on craft. Visit Timberland.com to shop. Your planet is now marked for death. Marvel Studios' The Fantastic Four First Steps is now streaming on Disney+. We will protect you.

44:55as a family. Light them up, Johnny! Marvel's first family is certified fresh on Rotten Tomatoes. That is fantastic. And critics say it's one of the best superhero movies of all time. Oh! Marvel Studios' The Fantastic Four First Steps, now streaming on Disney +, rated PG-13. What time is it, Ben? It's Marvel time!

45:22Ellis:How region-specific are the models? do you think? I mean, I was just picturing like the difference between training in LA and training in an old European city feels pretty different with a lot higher degree of unpredictability over there. Well, this is what's so unique about the neural net based approach. In a rules based environment, any difference becomes something you have to tune and calibrate. So if you build a rules based environment and you change the vehicle, or you change the sensors, or you change change the environment, like you said, like you go from San Francisco to, let's say, New York City, it requires a pretty heavy recalibration, localization for the vehicle, for the sensor set changes.

46:06In the case of some sensor set changes, a lot of it would be like a melt and repour. Here, think of it as just like you train your own brain. If you learn to drive in San Francisco, and let's say you had poor vision and you learned to drive without glasses, and I then took you, gave you glasses, so enhanced your perception stack and maybe even found a way to bolt a radar under your forehead and then put you in a different city, you'd still be able to drive. You'd still, you need to pick up some nuances. You'd have to learn those. So there's some training, but you'd actually adapt pretty well. And so the skills and the capabilities generalize far more in a neural net-based approach because there's not a, it's not a fixed rule set that you're following.

46:49as part of the code base. I think even more importantly, though, is the model continues to work and all the work that you put into the model continues to work as you change or enhance the perception platform or the compute platform. So imagine, you know, I jokingly said, imagine you put a radar in your forehead. Well, imagine not just improving your vision. So let's say your vision increases by 5 to 10x, including low light and bright light performance. You add multiple radars, you add a LIDAR, so very, very long distance visibility, which is great for training your camera perception stack, your equivalent of your eyes.

47:26But then imagine you could increase your compute by 10x. And then suddenly, like the patterns you're observing, and all these like nuances that are your current brain probably wasn't picking up on will suddenly start to see. And so it's really remarkable how with this neural net-based approach, the platform just continues to get better as you make hardware improvements. And none of it's throwaway, as it was in the past. Like as it was in the past, you increase your camera capability by 10x. The whole thing has to be redone. It's a whole new platform. Let's keep the AI theme going for a second.

48:02I think you guys are having maybe your next, I don't know if it's fair to say ChachapiT moment, but you've got the R2. coming out this year, right? Is it still coming out this year? This is the... Yeah, it's just... I'm in Illinois right now. It's just down on the line. It's... So you were just checking it out. It's going to be about what, like 45 ,000. This is like the real mass market play. Long time coming. I'm drawing the analogy of Chetchepeet moment. Maybe that's not fair, but like, it's a big moment for you guys. And I'm wondering how that's going. Yeah. So it's the, maybe the most important thing we've launched to date.

48:35It takes all the learnings from R1, of which there's many, you know, that, and when I say all the learnings, all the things we made mistakes on in terms of ramp or in terms of production system design, all the areas we can improve in terms of cost structure, all of that's embedded into one vehicle. And then the price point of it just generates, will generate a scale of demand that's much different than what we have with R1. And, you know, so if you look at like the, actually the closest comparison is Tesla. You know, Tesla prior to Model 3 was not a very high volume company. They had Model S and the Model X.

49:15It wasn't until Model 3 came along that the business really started to thrive in terms of cash flow, you know, in Model 3, Model Y product set. And so for us, that's R2 and then subsequently R3, which really drive this significant volume and also help rationalize all these really large fixed cost investments we've made. So we have a very, very large engineering team that's developing vertically in-house all of our electronics, all of our software. We just spent some time talking about all of our self-driving platform, inclusive of all the in-house perception and compute, like we do our own silicon that'll launch with R2.

49:52So we have just a very heavy fixed cost base or R &D cost space. And then we have, like, we're building a big infrastructure for the direct-to-consumer vertically integrated business. We have multiple large distribution centers across the country. We have, obviously, lots of sales locations across the country, lots of service, lots of service infrastructure, like parts distribution warehouses. So, like, we don't use dealers, which means in the front end, we have to spend a pretty large sum of money to get all that infrastructure built. And then it gets paid for with big structural cost advantages.

50:27once we get to scale. We don't have to pay a third party, you know, 10 to 15 % to sell our products for us like we would if we had dealers. The stores you guys have are very nice. I just went to the one in Pasadena. They're all very nice. You got a nice one in Venice too. Yeah. Somebody drove through the Pasadena one, like a car wrecked through the front glass. Yeah. Oh, wow. Now that's a mega fan.

50:49Ellis:Yeah. I bet it was a Tesla, right? I don't even know what kind of car it was. I love Rivian. it's gonna kill you to see all these ai rapper startups raising at like billions valuation with like 100 million arr and here you are like dealing with like thousands of supply chain parts and have been grinding for i mean i don't know how long how long is how old is riveting as a company it's like 16 years old 16 years and i was listening to you on ben thompson's pod uh in december saying you've refinanced your house to raise the initial capital yeah yeah so like you have been grinding on this thing you're in like one of the hardest categories on earth as far as like tech and manufacturing and i'm curious what you think of when you see all these ai software companies uh and it's just like yeah path to billions of dollars like immediately based on traction but also like a ton of hype and how the space you're in uh is just way more difficult and it seems like it's only gotten more difficult right yeah it's been difficult but anyway i it's like uh i think in time i mean we i'm so bullish on what we're building that um you know the way i the way i think about it is the likelihood of success is at an all-time high and it's sort of um you know i compare that to like day one when it was just me refinancing a house and starting like with$100 ,000, no employees, no team, no tech, no anything to go build a car company.

52:23Like at that point, the likelihood of success was quite low, probably near zero. And so if I contrast it with today, that the brand has resonated really well, the products are really great, and R2 just moves the needle meaningfully forward. And I think some of the work we're doing in technology, particularly as some of the things we already discussed with our in-house software platforms and a lot of the work we're doing around AI and self-driving, just puts us in a position that I think as the market share stacks, the market share percentages get completely restacked, which I think will happen over the next 10 to 20 years, I think we stand to gain a lot of that.

53:08And if we're successful in that, everything takes care of itself. The value of the business will catch up to that. And I think part of it is it's a really hard business. It's hard to understand. It's hard to explain making these multi, like very long multi-year investments that you don't see the benefit from for quite some time. It's not something the markets are typically used to backing. But I'm having fun.

53:35Ellis:Speaking of the tech crossover that the auto industry is now feeling, we all know the hardest part of being in tech or tech adjacent is dealing with gadget reviewers who post videos of your products on YouTube. As the new products come to market, how would you describe the difference between these stalwart car reviewers and the new age gadget influencers who are looking at your car as a big gadget? How are they different? They tend to – they just focus on different things. This chassis is awesome. Yeah, I think like a fun one that's like maybe shows the difference in perspective. The very classic car reviewers don't like a lot of the future technology.

54:20Like there is a big hurdle, emotional hurdle of having a multi-touch-based interface versus lots of buttons. And as you put them, like the gadget reviewers can't imagine having something other than a multi-touch. And so – Powered by Unreal. Yeah. Yeah. So it's sort of like – it's a quote like Bob Dylan, he can make some people happy some of the time. So like that's part of the challenge. You can't make all the people happy all the time. Like part of the challenge of a car is you have – if you're developing a vehicle, you have – we've estimated it. call it somewhere on the order of 30 to 40 million decisions you have to make as a company around the product.

55:01And you're going to make lots and lots of decisions that not everyone's going to agree with. But the goal of the business is to make those decisions really consistently. So the product feels, you know, while there's, you know, four or 5 ,000 brains that are working in parallel to develop something as complex as this, you need that many. Otherwise, it would take you too long. If you had one person do it, it'd take you tens of thousands of years, which doesn't work. So you need to have many people work in parallel. And so the goal is to have it feel like one person made every decision. It's really consistent.

55:32The person who designed the trunk was the same person who did the door seal, which was the same person who did the bezel around the screen, which is the same person who did the UI. Of course, that's not the case. And so as you start to create frameworks for how you make decisions, inevitably, there will be people on the chart that like want different things and i i always love the question like why don't you do this as if like we didn't think of some of these alternatives like you know and of course like we had to make trade-offs we and nothing's accidental it's not like uh you know we just like threw a bunch of parts in a bag and shook it and out came a car like we like we made a whole series of very intentional decisions not every one of which everyone will agree with.

56:14Ellis:That's why it's fun to be a gadget reviewer is you get to pretend that you're smarter than all the CEOs. Well, I think part of it is for me, I just got had to get comfortable with the fact that as long as many people like most of the decisions we make, and the products feel really thoughtfully done. I think we're in a good in a good place. I think where you have a challenge is if very few people like most of the decisions you've made, which there have been products and cars that have had that, where that like the the audience that actually likes what was conceived is so narrow. And so that's part of the challenge.

56:48I think sometimes reviewers may more often than not fall into the category. If you made everything that they wanted, it'd be a very narrow subset of customers that would buy it.

56:58Ellis:Oh, that's the classic power user dilemma, right? I mean, is there anything more specific that's kind of been that's been uh contentious with those folks why can't i sideload my n64 emulator onto your multimedia interface one that i would say we made a decision on and we were really intentional on it and i might you might be one if i were to do it again i might revisit it it's interesting is we have for the flagship product we have an r1s which is a seven passenger suv as a sister product, the R1T, which is a five-passenger truck. And we did a lot of work studying garages. And the reason it's important to study garages is because if you have an electric car, you're going to likely want to be able to park it in the garage so you can easily charge it.

57:46And to do that, the vehicle has to be under a certain length. And so you have a choice as to where do you put in that length if you want to fit in 90 % of the garages in the United States. And so we developed a bed, a tailgate on the bed that has a gooseneck hinge. So it doesn't hinge down like a normal tailgate. It hinges out. So when the tailgate is open, it's longer than your average truck, which is great for holding like a dirt bike or a motorcycle. But when the tailgate's up, the bed is four and a half feet long rather than more typically the five feet. So the bed is shorter. And a lot of folks were like, why didn't you put a five-foot bed on it?

58:25And we said, well, like drive through your neighborhood and tell me who owns a truck. And they're like, what do you mean? I'm like, you can know who owns a truck because the trucks don't fit in the garage. They're all in the driveway because the beds are long. And so like we were very like logical to say we want the truck to fit in your garage. So we made the bed, you know, the gooseneck hinge. And I say all that and it's like I could do all that like intellectual argument as to why it's easier if you can park in the garage. You can have the charger in your garage. You don't have to put it out on the outside of your house and all these things.

58:53But at the end of the day, it's still like too much explanation. It probably should have just made the five-foot bed and accepted that the cars would fit less garages. But as someone who parks my R1T in the garage and it wouldn't fit if we'd had a five-foot bed, I sort of like it still. So it's one of those.

59:07Ellis:Tell your social team the 50 top tradeoffs with the R1S or whatever you want to do. I mean, for a certain audience, there's like literally nothing more fun than this stuff. You know, with I mean, with the complexity of the car, I mean, you're like doing the James Cameron thing out here, like somehow trying to like have a single vision with hundreds and hundreds and hundreds of people. And yeah, I mean, of course, it's a bummer when the truck looks a little bit funny to some people because of that trade off. But I do think there is a world where those decisions actually improve the brand because they show how thoughtful you really are.

59:43Ellis:I mean, it goes back to, you know, the Steve Jobs antenna gate with the iPhone. just hold it differently or whatever. But like a lot of people are like, fuck yeah, fuck yeah. I'll hold it differently. Yeah. It's not like we just made the truck bed at random length and we're like, oh, darn it. Everybody else was doing five foot. We did four and a half. Seemed cool. RJ, you brought up charging and I'm glad you did. So your team actually hooked me up with an R1 for the weekend. I want to say it was like a year and a half ago and I took it to Joshua Tree. And it was one of the most stressful trips of my life because you all had not built the charging center in Joshua Tree.

1:00:17Yeah. It was under construction.

1:00:19Ellis:Sounds like user error to me, Alex. No, like it was just like frantically trying to find a power source at our friend's house that could like charge it with enough mileage to get to the next charging point on the way to LA. That's the worst I've been in that spot myself. Oh my God. It was horrible, but like the thing was fun as hell to drive and we took it off-roading and all of that. Like that experience was amazing, but the charging experience like almost completely ruined evs for certainly my wife um for me like i understand like the infrastructure is getting built out but i actually have this theory about ev adoption that one of the big gating factors is just how shitty all this charging infrastructure is and also the fact that like landlords of rental properties are not incentivized by the federal government or anyone there's no like tax credits to install ev chargers and like, you know, I would, you know, we're renting our house right now.

1:01:11Like I would probably have an EV if my landlord would just put it in, but I'm not going to pay to put it in. Right. And I think a lot of people, especially with like the shift in millennials, not owning homes and how property ownership is changing. Like people are really locked out of this category because of the charging infrastructure. And I'm wondering like, is there anything you all are doing specifically on that besides your own, you know, charging stations? Do you think anything will change like structurally to fix this? Because to me, this is like, this is the thing that is holding it back for me personally.

1:01:40Well, you articulated that well. And I agree with everything you just said. It's, this is a really important thing to solve. And so because we so, we so emphatically agree with you, we have created our own charging infrastructure. So we call it the Rivian Adventure Network. Unfortunately, this is an area where there's been like wild underinvestment in charging infrastructure, so much so that like our network is one of the largest in the United States. Tesla's is by far the largest. We're about somewhere between four and five percent the size of Tesla's network. And outside of Rivian and Tesla, all the other networks have frustratingly low uptimes, meaning if you go to the charger, the likelihood that it works can be as low as like 50%, which is a real pain.

1:02:32Imagine you're on a road trip and you found it on the map and you stopped there and it doesn't work. So we have a few things we've done to address this. So one is we're building our own network and that's growing as we speak. And your Joshua Tree trip this year would be amazing because we have this, like probably our best charging station is in Joshua Tree. It looks nice. I just drove past it a few months ago. Yeah, it's really cool. It's a very, very nice setup. So that's one. The second is we adopted Tesla's charging standard became a, they actually made it standard and open. And we've adopted that so you can now use Tesla chargers, which is a big network.

1:03:10And then the third thing we did, which is maybe less obvious, but informs the first thing, is we bought a company called a Better Out Planner. The acronym is often used to describe it, ABRP, a better out planner. And it's one of the most popular independent apps for finding chargers. And so when you use it, you can see like, like, the reason we bought it is not just because we think it's a nice app, which it is, it's great. But it gives us really good visibility to where all the worst charging situations are. And so think of it as like, we can identify the charging deserts, so to speak. Joshua Tree was one of those before.

1:03:50We've now fixed it with our node. And so it allows us to really prioritize how we deploy capital to address these charging gaps that exist within all the driving places you'd want to go, all the places you'd want to drive to. And so that's been great. And so not only that, we have a charging score. So we rate A, B, C, D, F, like how good is a charger? And you can sort by it. So you can say like, I don't want to go to a B or lower charger because I don't want to deal with the fact that it might not work when I get there. And so that's some of the things we've done. And then when it comes to home charging, of course, as you point out, it depends on the, It depends on where, but in many areas, we're starting to see building codes require building owners to put charging in.

1:04:49This isn't DC charging. This is like the super fast charging, but it's still level two charging, so overnight charging is fine. And I think in the next couple of years, this will start to shift pretty dramatically. real. Charging will go to something that was a bit of a pain point through the combination of our network, which is growing, Tesla's network, which is already big. Hopefully, a few of these other independent networks growing. That charging becomes ubiquitous. DC fast charging for road trips. Home charging, apartment building charging, office charging becomes very accessible. And it it is in my life now, like for me, I like, I don't even think about it because there's, I have a charger at the house.

1:05:32Um, I have an apartment that has a charger. Well, you're the CEO of Rivian. I would hope you don't have to. But like, even like on my road trip, like, like it's only like the truly obscure places that are not covered. And even those, most of them are getting covered or have been covered. Rivian adventure networks doing a lot of that. So I think most of those charging deserts are sort of solved. So it's, it is a lot different than it was a year and a half ago. So you should ask us to loan a vehicle gun and try it again. Yeah, I should. Get a quad. Ask for a quad. Drive it like you stole it and enjoy it in Joshua Tree.

1:06:06Ellis:I'm so fascinated by this because usually in tech, the things that users do once a year don't like kind of make or break the product or their perception of it in their mind. I mean, I have a feeling that most people wouldn't actually encounter a range anxiety issue very often. Is that true? No, it's very rare. Yeah, we have the data. Right. What is it? Extremely rare that we have a customer actually run out of charge. I mean, that's got to be pretty frustrating to have that be... I mean, is that still one of the top barriers with getting an EV? I think it's more of a consideration than I would think.

1:06:48What I always say to people is like try it like like we have a few ways that people can loan car like uh test drive cars we have like an overnight test program big part of that is just like experience how easy it is to charge i but i think in time that that starts to fade uh but as alex said you have one bad experience it does it does cement an idea into your head and it's like oh i don't want to deal with that and the stress of like being stranded is high especially if you have kids um or you're hungry or you know those kinds of things well rj i will ask your team for a quad and i will drive it like i stole it i know you got to go we appreciate you being uh on the show and taking the time with us good to see you again and and alex we'll make sure it's black on black oh it's got to be murdered out yeah it's got to match the both black everything dark tins all right we'll see you yeah bye

1:07:44Ellis:And that's it for this week's show. I'm realizing now I should have said in the intro, buckle your seatbelts, ladies and gentlemen, because that still applies to EVs. For now. Sorry. Don't forget to like, subscribe everywhere you get podcasts. We've got a website with all this stuff, access.show. And you can find us in video and see how much RJ actually looks like Clark Kent at AccessPod on YouTube. Again, if you like this episode, please leave a five-star review or thumbs up and share it with a friend. It really helps. You can find me at Hamburger on Twitter and at Meaning.Company for all your storytelling needs.

1:08:24Ellis:You can find my newsletter at Sources.News. Access is part of the Vox Media Podcast Network. This show is produced by Hooked Creators. Bye-bye. Au revoir. Au revoir.