SPEAKER_01: welcome back to this week in startup with me lon harris i'm jason calacanis we've got albert

SPEAKER_00: brotherton and boris radilov jason they're the co-creators of nanogram i really like this this is a cool thing it's basically tiktok for mobile video games it's available right now it's in the google play store it's in the ios store guys thanks so much for being here thanks for having

SPEAKER_03: us guys so tell us what you're working on yeah boris do you want to kind of bring up um bring

SPEAKER_04: up the demo and walk them through it yes yes uh just give me one second sure this is i have it on

SPEAKER_00: my phone jason this is it's available now it's basically work it's like think tiktok but people can create their own video games very quickly like casual mobile games and then you just scroll when you're tired of playing one you just scroll up and you go to the next game it's pretty remarkable

SPEAKER_09: typical tiktok feed this is the mvp we put out uh a month and a half ago you're playing oh my god you get bored and you're going to the next game you play this game you get bored you're going to the next one it's very simple concept uh proving to be quite interesting but the whole brain role maxing of playing multiple games in a row is just part of the flywheel the second half of the

SPEAKER_12: flywheel is the creation aspect when you're looking for a new game to play it can be that complete like

SPEAKER_00: tyranny of choice like choice paralysis like there's hundreds of games i don't have time to research every single video game this is just like open your phone start playing a game the creation is

SPEAKER_09: even cooler if you go on your create tab you put create with ai and you have all these templates you could go off or you could do a custom template i'm just gonna quickly demo something real quick for you so when i was like 14 i started programming because i wanted to make video games and you know now with ai and agents something that would have taken me a week back then now takes 90 seconds so for example if i want to make like a 3d version of flappy bird all i have to do is go on 3d write flappy bird hit generate and now it's gonna cook up the game for me and what's really cool is while you wait for the agent you get to play other games that other people have made for inspiration uh and usually on the first prompt it takes about 60 to 90 seconds to like one shot a full game uh and after that the cool part is that you're absolutely free to remix it i'm having a hard time playing and talking here at the same time i have to forgive my poor performance but uh the idea yeah the idea is uh after that you can remix the game as many times as you want uh and we'll just probably get that shortly uh additionally you can remix other people's games so for example if i make like a really cool game i send it to jason here and he can remix it change it in a way that he likes so here is the 3d flappy bird that i made for us uh and this is very basic this is like a starting point for us and i get to play this and as a little kid uh i i this would have taken me a few days to figure out before the age of ai but interesting enough i can now do this in like 90 seconds and i can go prompt it and i can say hey can you please add guns to this i like that you said

SPEAKER_01: please very polite so this is all just ai prompting to make games what's the engine behind it we use this clawed or something or is this like some bespoke ai engine to make games what we're doing is we've

SPEAKER_09: created a custom game engine and then we're using gemini uh and a bunch of agentic harnesses and like tool calling to give it the ability to create 3d assets create 2d assets out of like pixel art and and things like this basically what we've created here is we're trying to rely on the native platforms as much as possible so we're really relying on gemini to do the tool calling for us and then we're putting a bunch of tools such as like generate 3d mesh generate uh 2d pixel art generate sound things like this and then it figures out how to call these put it in the

SPEAKER_22: game and and do all sort of stuff for the user genius can i take lon's game and build off of it

SPEAKER_25: and fork it yep yep yep this is like one can post like you know uh a flappy bird and you can say i want to add guns this you could send a picture well i mean at the moment okay so that's inherent in it

SPEAKER_29: is you just fork stuff it's like the remixing it's remixing yeah yeah it's like the remixing on

SPEAKER_00: where i can see your video and i can say make them say this instead and it's just like everybody's

Chamath Palihapitiya: idea builds on everybody else's idea yeah what's the business model here and who are you two dudes

SPEAKER_25: so i'm albert this is boris um i'm the ceo he's cto and um yeah i mean the business model is i mean look at the end of the day we're like a feed of interactive content right that's that's like i'm not even saying games because that's the direction we're moving in right interactive content and you know we have a lot of rev experiments we want to do right because this is a very new field we can't know exactly what it's going to be right but i think it's a very interesting space where you know there is no accessible interactive ad space that's really easily distributed right now right so i'll use an example here so let's take like a big advertising provider like domino's right and they want to you know do some easy ad space stuff they're going to go on short form content right and they'll have a marketing team that's great at short form content they'll post on instagram the post on tick tock right but you know at the end of the day interaction is much higher converting

SPEAKER_01: right and yeah and it's engagement so they're going to spend more time on it so domino's can come in and say i want to make an asteroids like game where the boulders and the asteroids are domino's and when you break them apart they turn into the pretzel bites and then i want to have a spaceship go by

SPEAKER_25: that's the brownie whatever it is genius and make the pizza and the call to action can be order the exact pizza you've made in the game for example yeah very cool albert and boris you guys went to school

SPEAKER_04: together you're you're in high school right now what's the story no no so i'm i'm 22 boris is 24

SPEAKER_40: we've worked together i'm glad i look so young jason i was referring to albert but okay um no joking

SPEAKER_34: you're both they're pretty young but so you're you went you're at a college did you skip college you went to college i i skipped college personally well done um i went to college unfortunately and

SPEAKER_44: i went where'd you go where'd you study university of leeds i studied computer science and uh yeah

SPEAKER_09: unfortunately it was right during covet so for the two of the three years i literally just didn't go

SPEAKER_47: to classes so i guess so you paid a fortune and got nothing got it yep yep that's ridiculous how much

SPEAKER_51: debt you got boris how much not much it's like 40 50k in the uk it's quite not much 50 000 okay we

SPEAKER_07: got to make this start at work i mean compared to some of the people that like in the u.s where

SPEAKER_35: they're getting on yeah all right albert what's your story you're the business guy you're a developer

SPEAKER_25: who are you you i'm the business guy i i'm the business guy i'm the hustler yeah i mean i've kind of i've i've been a gamer since i was six by the way i'm like one of the weird people in gen z that plays a lot of world of warcraft which is kind of like more of a millennial thing wow boris is two but we play together but i've played since i was six and is that how you met did you meet in world of warcraft no we met through a childhood friend that kind of grew up with you know me and grew up with boris and he just intro'd us about three years ago and we kind

SPEAKER_30: of kicked off and became best friends and now you're sitting in a gamer chair right that's a gamer chair yeah yeah it's a razor yeah see now this is what i i'm like the colombo of investors

SPEAKER_60: like i look for little clues and i'm like hey you're at the east lawn i noticed albert is sitting

SPEAKER_62: in a chair the chair is a gamer chair and it's it's the brand razor and i had a razor my brother-in-law had a razor uh mouse and it was used for playing games like world warcraft so you're very much into

SPEAKER_35: gaming are you not albert so uh you guys vibe coded all this you did it you have there you go razor

SPEAKER_38: and has weights in it too you can put different weights into it depending on the game you're playing what's your story you guys are going to raise money for this you want to come to an incubator you have raised money you did go to an incubator have you incorporated what's the story here because

SPEAKER_25: this is a great idea yeah yeah so we have raised money i mean it's a really exciting time for us we've raised from drive awesome fantastic uh well and where are you based so we're based out

SPEAKER_34: of uh london and sophia at the moment but with plans to go towards new york yeah okay great man my

SPEAKER_35: hometown um it's great just when you go there please don't join isis okay so when you get to

Chamath Palihapitiya: ellis island they're gonna be like would you like to go to the isis training camp or you like to go

SPEAKER_69: to brooklyn or whatever it is well it's a bit like that in london to be honest too so don't worry you

David Friedberg: know yeah you just have to defend yourself as young people against these uh ideologies don't blow anything up um fantastic new york's a great place for you because there's marketing in new york and people experiment and this could be permissionless so if you just tell brands hey you can just build whatever you want as a brand some young person at a brand um i don't know if you're watching the staples uh baddie do you know about staples baddie no lawn do you know about staples baddie on tiktok okay so one of our team jaco will very quickly i found it i found it and uh staples baddie she

SPEAKER_35: basically is passionate about staples and she works at staples she is like a baddie in the internet slang and she just talks about like here's the best pen and she was talking about pens the other day and i was looking for a new pen and i don't want to spend a dollar on a pen but i don't want to spend

SPEAKER_01: a hundred dollars on a pen and i was searching for a pen and she came up and um one of the pens that she liked was the zebra g 750 that i literally bought it now i didn't happen to buy it from staples but i did follow her so now i'm giving credit to staples for it anyway my point is she's getting more views and lon probably has done a bunch of research on her right now she's gotten more views and has done

SPEAKER_35: more for the brand online than anybody in corporate who spent 10 million dollars or 100 million dollars ever had because she connected with people authentically hiring can be its own full-time

SPEAKER_01: job and hey guess what i already have a full-time job i make podcasts and i invest but when you're running a small company we both know every hire matters you don't want to waste any of the seats you have at your company and the best partner you can have is linkedin hiring pro why there's a billion people using linkedin all the great talent are there if you're proud of your work you build a linkedin page and you update it linkedin hiring pro is going to streamline and simplify the entire process for you nearly 60 of companies using linkedin hiring pro you're going to get an incredible candidate to interview in the first week and you know we're looking for a new producer for the pod we did shout outs here on the show we posted it on my social media we asked friends you know where we found our next great hire linkedin and it was competitive we had like three or four really good choices so hire right the first time post your first job and get a hundred dollars off towards your post at linkedin.com hiring pro offer that's linkedin.com hiring pro offer terms and conditions

SPEAKER_35: apply she went through journals she went through pens now she's getting picked up whatever now other people office depot is trying to find their baddie it's become like a whole thing it's like the the

SPEAKER_00: fast food burger thing that you were following that last week it's the same it's the same concept the mcdonald's ceo did a tick tock where he came to the audience that was that was really bad he ate a burger but he took a really small bite and it looks like he's disgusted by his own burgers and so now every other fast food ceo has done their own version where they love their burgers and they're eating stuff in their face with them and it really became like every single fast food ceo had to do one to get the momentum from this viral mcdonald's video i think the problem

SPEAKER_01: with his video because i did see it yeah and the problem was he's very robotic yes he's almost like

SPEAKER_87: an mba who's a bit on the spectrum i don't want to diagnose people but when i say like he's robotic

SPEAKER_00: you know you get the idea it's not he's awkward and and it just it he's awkward he's not a good enough actor to pull off like i really am enjoying this and i want to eat this burger

SPEAKER_93: he's not even able to pull off he's a human no at one point he goes and this and this is gonna be

SPEAKER_00: my lunch later and that's not a hard thing to convince me that you're gonna eat a burger for

SPEAKER_12: lunch but i don't believe him i don't also in that moment i don't believe him he referred to the burger

SPEAKER_96: as product yeah a product it was like very awkward and he takes a very tiny bite he does not take a big

SPEAKER_00: bite i thought the same thing about the burger king guy honestly the burger king guy did one and i was like he also doesn't really take like a a big like a man-sized bite like he loves

SPEAKER_97: burgers the question is who let them pose this like there was probably a whole marketing department

SPEAKER_00: yeah like right there wasn't a social media person there being like can we can you take a bigger bite like you like your your burger because that looks so dainty it's like it's yeah he even shows it it's like it doesn't look like it's so clear albert when you see this that this person does not like their own

SPEAKER_01: product and there is a there is literally a salad that has been made to his specific specification by his chef off camera and he is going to spit out whatever he ate swish his mouth with some

SPEAKER_35: pellegrino at the exact temperature he wants and then he's going to eat that freaking salad that his chef made with that literally each item was weighed and put into his you know uh spreadsheet where he keeps

David Friedberg: his calorie count what you need is somebody who actually is sitting there with the chef saying put more onions on and like a you know the mustard ratio is off which is what the crock uh which which is what the mcdonald's brothers were doing albert what's the vision here um for the oh so anyway back to the permission list and we're talking about brands i think you just let the brands do whatever they want on and then you give them the ability to promote them yeah which would be just what brands are doing now brands just make content there's no permission to do that and if they get views and they do a good job it gets views organically but they can boost it um and they can pay to boost i think just paying to boost such a good idea then what about monetization from i'm the creator could i put a game

SPEAKER_01: up there and say i if you want more levels you can buy coins or is that annoying to customers and

SPEAKER_24: against your philosophy albert so again this is a big rev experiment we want to do like look there's two there's two ways i look at it you can look at someone like roblox right that's done incredible

SPEAKER_25: creator currency right however their games are a hundred percent more long form right where you're not going to be playing like you know a hundred games an hour right or more right so i mean it's very normal to have microtransactions and you know a creator currency where you as a creator can say hey look i want to add an extra level you know clear it's clear i got a million plays on my game let's add an extra level for my big supporters that want to do it right um this is something we want to experiment with so yes we're we're looking at experimenting with a creator currency you know because i think there's a big incentive there for creators right so let's say you you put a game up right short form you made it in two minutes right you did like three prompts and you get a million plays right and just so we have a game in in our minds that we all know let's just say it's like candy crush and candy crush doesn't exist right and you pioneered the first level i want that to be a big incentive for you to kind of this is something you don't see in socials you know like iterate on that already posted thing right unit of content this is something you don't do in like instagram and tiktok you know the moment you like press post that's it you know the ship sailed so i mean there is an aspect that we want to test on this but to be totally to be totally frank we're too early i think to really know how that will go and yeah we are not the ones that can decide if it will work

SPEAKER_37: or not on that i think you spend the first literal two years just trying to make the tool so good and

SPEAKER_34: so addictive that people get into it i mean it's already kind of getting there i mean have you do

SPEAKER_25: you know anything about like our attention and like kind of angie yeah uh tell us a little bit about it yeah quite cool so we launched like mid-january and we've got about 100k users so far um and 20 percent of them are what we call power users right and these guys play more than 25 games per session and the average session time is around 21 minutes and they do two sessions a day so they're playing you know just under an hour a day and going through over 50 games and i think this is really cool because i mean you saw like a brief glimpse of the feed we don't have an algorithm a which is what drives most social engagement on platforms like instagram and tiktok again the point on algorithm you know i won't linger too long is never been introduced to gaming before right drives most short-form social engagement but never been done in gaming right so we have really high session times for you know basically giving you slop right that's what it is you know until it gets better and you know what we're seeing is kind of people are already really excited

SPEAKER_47: about it yeah i think this is going to be a breakout hit um and how do you get the first couple of

SPEAKER_22: casual users where do you find users uh for the product we started this with discord to be honest

SPEAKER_25: um you know we started a discord community like we had the idea in november we started a discord got it to like 10k people but as discord is not not all 10k people were kind of super active kind of like 2k of them were and these were like our test users and we had like a running beta of 200 people up until we launched in january and these guys loved it like everyone that made games back then are still making games today so you know insane you know commitment to it and at the time we didn't have like game generation they were like doing it on their own and you know uploading it um but then you know a huge way you know and i won't really touch on ads because i mean ads are quite obvious way to get users but you know another really interesting thing is sharing games right like i don't know you mentioned a lot about socials and sharing stuff have you shared a twitter a tweet you know of course

David Friedberg: yeah i mean in the group chat etc so that's going to be in high scores and sharing and then not to mention if you have games that are um a duel so like lon and i could play versus each other or we could invite each other to beat a score so that that's where like the loop could happen right

SPEAKER_34: 100% a wordle type game and then i send it to lon and say look i i solved this in three out of five

SPEAKER_25: yeah can you try and beat it yeah well yeah i think it's really interesting because like look i think the big point i'm trying to make with this is it's not in our culture to share games right like you say you share you know reels that's true tweet but when was the last time you were playing a game

SPEAKER_34: well i will say there is one it's um people will share clips of like specific kills in call of

David Friedberg: duty on youtube whatever that's like deep gamer but that's not casual content that's still content

SPEAKER_25: and that's still super niche at the end of the day like you know you will if you are an avid instagram user you will daily send reels i mean this is a way of keeping in contact like i'm a middle child i have two brothers i send them reels every day and they send back right yes but games aren't something people share right they don't right so this is like one of two big questions we had before doing this of and by the way we have the answers i think it's really interesting but these are one of two big questions we had before doing this was like a do people want to share games because they don't right now right and b will this shift look the same as the long form content youtube shift to short form content instagram tiktok right so the answer for the first one is yes per 100 likes on the app right now we're getting because we have likes and shares that you didn't see on boris's beta version because it's a beta version that we're running right now but per 100 likes we're getting between 30 and 50 shares at the moment so huge amounts nearly one in two and then when it comes to the long form short form shift i've already told you you know the stats of we have 20 of our user base going through 25 games per session in less than 25 minutes right around 21 minutes so i mean the big thing here that was like a huge sign for us was look when you are talking about long form content and before tiktok and instagram there is a use case to spend three seconds on a video it's your camera roll genius the camera roll from iphone whatever android has as well right you can go watch three second clips that you've taken yourself or a friend has sent you and scroll through them but games are very different right yes because with a game you go on flappy bird what's the session time it's like eight minutes you know you go subway server something like this i mean don't quote me on those stats i don't know exactly what subway server session times in but it's not five seconds right but i mean so this was our big question will people basically do this hyper casual doom scrolling but for gaming and it will be answers yes awesome uh great idea can't wait to get working

SPEAKER_38: with you and we will uh have you back on the program in six months and see what your progress has been really excited uh albert and boris thanks for coming on the pod also i look at 10 000 startups every year what is the essential nature of uber or robin hood these incredible companies i was lucky enough to invest in they focused on their customers they focused on their product and then they had

SPEAKER_130: support for chores let's call it what it is your incorporation your banking your payroll your benefits your taxes all of that back office you can drown in that or you can find a partner who loves doing chores and every dot io wants to do this detailed work for you they're going to handle your delaware c corp your ein getting yourself a registered agent all with no legal fees all without delays and if you're further along in your founder journey and you need to scale every will help you hire contractors in over 200 countries there's talent everywhere but there are blockers or maybe hey health and dental benefits that's all baked right in and you're gonna get three percent cash back on every dollar you spend with your every corporate card so head over to every dot io and stop wasting time on the stuff that isn't growing your company that's e v e r y dot i o every dot i o welcome back to twist my

SPEAKER_132: name is alex and today we're going to solve one of the largest pain points in my life which is picking

SPEAKER_134: up after my children i have a great many children they leave things all over the house on top of things underneath things it's a mess it's a constant struggle to keep my house up to a reasonable standard of care which means that i'm constantly bending over and making bottles and changing diapers now i love my kids but oh my gosh would i not appreciate a little bit more help the good news for parents like myself out there is that there are companies building humanoid robots that are going to make our lives easier in time and even more importantly there's one company called one x that's making a robot called neo that's designed just for your home it's my absolute dream so to tell us more about it how it's going to come to market and when please join me in welcoming to the

SPEAKER_135: show it's burnt bernick how you doing welcome to the show hey great to be here looking forward to

SPEAKER_137: this one so i've known about one x for a while you know keeping track of your progress and so forth but

SPEAKER_134: one thing i didn't know was that you guys actually got started building uh robots not for the home but for industrial settings and you had a robot called the eve industrial on wheels years back but made a pivot towards the home but take us back in time to uh the robot that you guys actually put into uh you know industrial applications years back when i started the company it's been 11 years it's been

SPEAKER_141: a while um the first things that i really wrote down they still stand today it's like we want to make robots that are safe so they can live and learn among people they need to be capable like they need to have the dexterity the strength the agility that we have or they're just toys and then of course it needs to be scalable and affordable so that you can have an impact and when we designed eve it was very early and we didn't have the power density and the technology yet to make this as general as we would wish but we can make it very general for its time and it was an amazing robot and what people might not know is that i actually had eve at home for multiple years oh really so you had the robot that was designed for what appeared well it wasn't designed for industrial use cases but it was like designed for general labor that's the thing about humanoid right so to me humanoids is a play and like you want to create general labor because at scale if you look at the system at the limit what is going to be the most reliable the most affordable the most intelligent and the most helpful is going to be whatever has the largest scale and all technology goes through these cycles and maybe the simplest one to talk about is computers right started with mainframes and at some point consumer pcs came around and it got scaled to this incredible number and the ecosystem and the reliability and the cost and everything that comes with it right so now it's just this one tool this is like the hammer that solves everything and uh whether you're like typing something up or you're recording a podcast we're all doing it on the same computer right there's no specialist in the end it actually goes full circle and the market is so big that you get specialization again it's happening in computers now you have like specific computes for inference for training for simulation workloads for all kinds of things right because each of these markets are gigantic at this point this is happening with robotics or physical ai it started with industrial robots that's kind of like our mainframe now we're going through the general labor phase where you just want to create the machine that is as general as possible that can really unlock humanity from kind of being bottled bottlenecked by access to productivity yeah in the end it'll probably be all star wars and there will be different droids for all kinds anyway eve as a long wind wind way of saying like he was our best shot of like making something that's as general as possible to get started on this multi-decade journey of how do we create something that can completely and fully kind of empower us as humans to work on the things that matter and i had it at home because it was very important for me to make sure that like hey this is going to be in people's homes at some point right we need to start early but it wasn't safe enough that that's actually the thing like it's extremely safe robot for what it is and i would argue it's safer than most of the other robots out there today but it was a bit too heavy and not quite general enough and uh that was really what led us to uh use this for industrial markets because i i've been working all my life on robotics and we built a lot of beautiful robots up through the years but like they didn't have any real world impact and it was so important to me when i started one x we want to do something that has real world impact so it can't be on youtube it can't be on the lab it has to actually be out there doing things and those industrial use cases were great early applications but if you want to make

SPEAKER_161: something that is generally useful to humans you're going to need a couple of things to make that happen going off of the original eve idea into neo which is legs and i would presume modern ai techniques that allow the robot to become more generally useful by learning new things on the go so it sounds like you had a good proof of concept and then you refined it and then imbued it with the latest in kind of modern ai and as a combination you've come up with neo which we're going to put on the screen right now and uh and so everyone can see it and the result is essentially a softer lighter more

SPEAKER_134: intelligent and more frankly human robot is my read of neo is that a fair encapsulation there are many

SPEAKER_163: reasons for this let me let me start with a couple of like very important ones that i think is quite

SPEAKER_141: unique to onex we've been very all in on we have to create something that moves and interacts with the world exactly like a human does all the way down to like the smallest interaction of like how what's the stiffness of like the tissue and your skin and how does that interact with the world right if you get that right and the sensing and everything then you can take all of humanity's knowledge mostly encoded in video and other things that exist out there and these priors hold and you can use them to get intelligence off like off the ground and this is so important because the alternative is go out and create an internet size data i'm not saying these things aren't useful we also use teleop yeah it's not the solution to bootstrap your way to having billions of tokens that you can train your models on so yeah you kind of have to find some way to bootstrap and i think that's what makes me the most excited about neo these days is that we're getting to where we're clearly seeing that that works like the world model work that we've been putting out and there's some amazing stuff coming right around the corner actually on this that's the continuation of that really starts proving out that this bet is a bet that holds and we can get true general intelligence most of what you see robots do today on youtube are kind of like fine-tuned policies as we call them meaning like

SPEAKER_163: it's not necessarily one task but like you take a set of tasks and you gather a lot of data on specifically these tasks yeah you can have the robot do these tasks maybe in another environment

SPEAKER_141: but the tasks are the same tasks and they're very similar what we're building here is actually true general intelligence where you can ask the robot to essentially do anything and it will do a

SPEAKER_161: pretty good job at it let's go back we're going to get into world models in a second but the point you made about neo the robot for your home designed for that uh having similar physical attributes to humans you mentioned like finger skin tension for example that allows you to

SPEAKER_134: map how humans interact to the world to the robot itself so that way it can learn more cleanly accurately quickly from how the world operates so essentially as you make it more human it learns faster it does

SPEAKER_141: right it's just like if if i pick something up i do that in a specific manner with my fingers like i rotate if like if i rotate that one like you just rotated that in your hand right the way you use your fingers to do that it's very hard to transfer that to like a three finger gripper or a claw or whatever right it doesn't transfer so like yeah so then you would need to go out and gather all that data so that's one part of it that's just so important it learns faster if you've got an engineering team at your

SPEAKER_38: company i'm betting there's a solid chance they're spending far too much time on infrastructure you need your team building your product to delight your customers not configuring your virtual network render is the all-in-one cloud platform for developers that allows you to deploy scale and secure your apps and agents with zero ops most cloud platforms ask you to split your focus between product and infrastructure or they force you into platform constraints that you know you'll outgrow in six months but just connect your github repo to render and you are live live web services cron jobs manage postgres the whole stack in one platform it's time to find out why five million developers are already using render go to render.com twist and apply for the render startup program you'll get anywhere from 500 to a hundred thousand dollars in free credits depending on your stage and who your

SPEAKER_141: backers are that's render.com twist now the second one of course is that the world is made for us yeah so like if you want to get around the home and like do all these things that you don't want to do yourself you kind of have to be very close to human and i also think there's something quite magical about creating something that the embodiment kind of connects to us emotionally uh and the companion part of the product is also so big part of the product that doesn't get talked enough about it's not just about doing the labor it's about doing the labor and being your companion throughout life and how this can kind of like help you interact better with technology and i think we're very quickly trending at least for me now that's using this every day in the direction of like i i don't talk to my computer i talk to my robot yeah it's away from screens and gets you more present in like your everyday life and i think there's a beautiful story to be told there someday about how this can

SPEAKER_160: get us away from screens well i i think also there's an enormous elder care component to this i mean we

SPEAKER_161: we've all heard about how society is graying how our elderly you know parents and friends are lonely and so to me like sure we have pets that's great but pets can't help you with dishes and you know what what if you're older need help and you're a little bit lonely well we can kind of fulfill a lot of that with one device in the case of neo and i i don't view that as dystopian at all i view it as a way that humans are taking care of humans via technology and to me that's a a good future 100

SPEAKER_141: agree and this is actually one of the main things that gets me up in the morning i think we we are at a point in history where this is not optional like we we have to apply technology to solve this because we have to be able to give her one uh essentially like dignity in in the way you get treated as your age and people sometimes mistake this for being like a replacement it is not and i like to say as someone who has one at home like it doesn't it doesn't replace my dog it doesn't my kids it doesn't replace my wife it is something new right uh and just like a dog is also something new that you add to your family yeah this beautiful companion throughout life that like it's always on

SPEAKER_182: your side it's helping you with everything it remembers what's going on and i kind of almost see it more or like my my hobs in calvin and hobs uh-huh yeah yeah yeah yeah well if you don't know

SPEAKER_161: that reference covenant hobs is an american uh child focused comic strip and a young boy has a stuffed tiger that he thinks is real and they play games and such so that's that's the context of it so essentially it becomes if you don't get that context go read it yeah go bill waterson it's it's amazing it's a great i grew up on it so that's that's very near and dear to my heart so here's the thing though you're talking about you know youtube demos and we're not going to name out any names we're not going to say figure we're not going to name any of the companies out there who are famous for doing a lot of demos online uh versus the ability to make something that works you know in your home so you have neo at home right now and i know pre-orders are up um it's going to start shipping later this year we'll get to that in a second 20k or 500 bucks a month so how good now is neo at general household tasks and how quickly is it improving against that problem set and the context here barrett is that i'm curious about the 1x world model and i'm also curious about how teleop plays into this because i love the idea of being able to call an expert into my neo to help it learn a new task but at the same time to me with my unlettered perspective here is that certainly after enough usage we'll have closed those gaps because enough people will have needed that bit of help so i'm

SPEAKER_134: curious about compared to youtube demos how we're doing today and how quickly we're improving neo i

SPEAKER_141: want to back up one step i'll answer that from like a bird's eye view so first of all you asked a question earlier which i kind of covered but not all of it which is why the home first right because this is about general labor so it's not just the whole but it has to start in the home because we want our machines to live and learn among people and we wanted to like hold the door open for grandma and we wanted to understand the social context of work and this enormous diversity is actually what gives intelligence so if you're in this narrow environment where you're seeing the same thing every day you don't actually learn it's the same for people like we know it's about to learn um so that's why kind of the home has to happen first and that's also why i think like this let's call it the early adopter program like what what we're shipping here the 20k robots literally was called the early adopter program um and it's gonna be rough i'm just gonna like tell everyone up front it's gonna be rough because it's the first time in history anyone does this but it's gonna be one heck of a journey and right now i would say in my home the robot is doing a reasonably good job it's not doing everything but it's like it's doing my laundry it's doing a lot of the tidying and cleaning it's it's doing a lot of fun stuff just around like companionship and just being around uh like opening the door if like a guest is coming over or all these things that's a party trick opening a door when the door when the door dash guy comes over is actually a lot of fun um just to see what you package so um now the way this works right now is there are two modes and we're gonna listen to customer feedback her also so like knowledge is locked in stone but like the way is there's two modes there's best effort autonomy mode where you're running the world mode and it's

SPEAKER_140: doing like whatever i ask it to do and it'll do best effort and you can ask me to do things with just speech you can talk to you just talk to it and it's extremely general right the thing that blew

SPEAKER_141: my mind the most is solar day was like i was just asked the robot like hey can you pick the post-it note on the board over there and read it to me and the robot could do that that's not the training data like that's pretty magical to me right that's actual true general intelligence now next time i ask it about the same thing because it was so fun hey look at this like neo go read the post-it note it didn't

SPEAKER_140: get the post-it note so it doesn't always work that's my probabilistic ai dang it and this is actually

SPEAKER_141: the magic of the world model because what it gives you is it gives you this incredible general base layer of intelligence where you can have the robot sensibly just based on voice it'll have a sensible approach to essentially any task now all you need to do is to have a tribe and then it learns and that's a lot of fun um there are tasks where this is extremely fun and just works like it's very good at laundry because if it screws up folding my shirt i just ask you to do it again right a lot of risk if it's taking grandma's like ancient vase out of the cabinet to put it on the table i would probably not do it that way um so there are tasks were easily repeatable and where like failure is not a problem then the robot is actually very good opening a door is not a good example because we

SPEAKER_163: make the robots just like us right to be soft compliant low energy lightweight so like it won't hurt itself it won't hurt the door it can just try it against it so for these tasks it's starting to work extremely well for tasks where you don't want it to fail then we're still relying largely on

SPEAKER_141: scaling our fleet also internally to work on these tasks to get them to where there's an extremely high probability of success and also that you have a very good way of ensuring that if the task is going to fail you identify it early and you stop um this is another thing that's incredibly cool about the world model because it for it works very similar to how we do when we're thinking right so you said you have kids so if you're going to go around like pick up the coffee cup with hot coffee then you will immediately kind of like simulate forward what can go wrong like through your mind immediately shows up like you pouring coffee on the kid or like all these things yeah yeah yeah and then you select the safest kind of like trajectory to achieve your task based on all these constraints and this is literally what the world model is doing it is looking forward simulating like if i take these actions what will happen and that's kind of like a search and then like hey here's the best way to do this yeah and we're

SPEAKER_163: working very heavily on the safety side of that current and like how do you ensure that the robot always takes the safest possible paths to do things and how do you ensure that if there's um unduly risk that the robot does not do the task now this is still working pro this is still working progress

SPEAKER_203: but it's incredibly important progress but uh because safety has two aspects it's the aspect

SPEAKER_163: where like the robot just is physically not able to harm you because it's light enough and low energy and soft and all these things that we've been working on for a decade and that's kind of like that's

SPEAKER_141: that's the separate thing because that's just you prove out that this is safe standard certifications and then you have the ai part of things where right now for customers we limit to some extent what the robot

SPEAKER_163: is allowed to do because it's not allowed to cook for example it's not allowed to cook etc right no no hot liquids no like dangerous items these kind of things but of course long term we want to be able to do this with a very good safety profile and that's really like i think now the bleeding edge of like

SPEAKER_141: i think now the bleeding edge of like what's getting worked on on the ai side to ensure these things can be done safely so you in my uh oh sorry please as you said like so in my house that best effort ai mode is what i use when i'm home and it's a lot a lot of fun when i leave for work i just take off my new app and i say that i'm leaving do all the daily chores when i come home it's all done and there is sometimes tell up involved in that to make sure that everything gets done super well

SPEAKER_210: yeah and i don't care i'm not there i don't care either i just i just want the laundry to be done i want there to be clean socks for every child when they're needed you know like i mean that's i that's

SPEAKER_161: why people are like oh 500 bucks a month i'm like yeah but do you know how much this would save my life like it's it gets worked pretty quickly now i want to talk about the world model for a second you guys have written a lot about how vlms are not a sufficient answer to general intelligence for physical ai aka robots for folks out there who are listening who are not as deep into ai as you are can you explain what a world model is and what it took to build the one you guys have released and

SPEAKER_140: apparently are going to be updating very soon vlms are essentially take a language model

SPEAKER_141: and then vault on some actions that a robot can do so and how does this work then in practice right so okay you take a screenshot essentially of the world so yeah and you have some text of what

SPEAKER_163: you want to achieve and you have a picture of how the world works and then you make a plan

SPEAKER_141: and then you start executing this plan and you take a new screenshot of the world and do this and it doesn't really capture the dynamics of the world and the beautiful thing about human intelligence is that we understand how the world works down to like we can visualize what will happen when we do things and to do this you need to capture like the spatial and temporal dynamics essentially that's a complicated word but what it essentially just means is we live in 3d and we care about time we see how the world evolves in 3d over time language models is like a 2d screenshot it doesn't have time

SPEAKER_163: it doesn't have 3d um and it works incredibly well by the way like language models are incredible i freaking love them yeah yeah uh it's just it's not the full solution for general intelligence it's a very narrow type intelligence that works really well on a subset of problems

SPEAKER_141: and it's very exciting to see that when you actually train these world models which then essentially what we're training them on is can you predict what will happen so very canonical example would be if i take this and i drop it what happens that doesn't sound too complicated you can do that and that's like it's not magical like it learns how physics works essentially but when you start doing this at scale there's a lot of magical things that happen so for example back to why the home and why among people to navigate a social situation to get your coke in the fridge the robot needs to understand

SPEAKER_163: how people will behave oh and the world model we can see that we can see now in the world model that like people appear in the world model right because the robot is thinking about what people how they will

SPEAKER_141: look what they will do etc right people appear in the world model and behave like people this is kind of like uh an inception type argument here but like it's kind of like an agi complete problem because to be able to fully understand how to interact with the world you need to be able to fully simulate how people work yeah um and it is just in my opinion it's the it's the natural next

SPEAKER_163: step in intelligence right so it's not that we're not using language clearly language has is a part of

SPEAKER_141: our intelligence but language is not the base one of one of our intelligence that is our sensory visual body and how we interact with the world and you can see this with kids right they learn how the world works and then they start to express this through language but slowly yeah so uh it's uh it's it's really exciting and we are starting to see like i said like the first kind of breadcrumbs of like things with robot data like this working better than the pure digital data getting us increasingly confident that like the future of ai will be models that actually train on embodiment not just the web now we still train on the web it's not like instead of it's in addition

SPEAKER_171: to yeah but it's currently a missing component that i'm very excited about so when you get the early

SPEAKER_161: adopter neos out into the market this year and i i do want to ask you about how many when and so forth in a second but putting that aside for now as you get more of them out there that increases essentially the neo footprint in the real world to test its vision model against reality and i presume greatly increase the the flywheel of learning that you guys can then bring to bear on future iterations

SPEAKER_141: to its intelligence right hundred percent like this is one of the most important things we do right and why we need to get it out there as early as possible that's why i i'm excited actually that

SPEAKER_161: you're saying it's going to be a little rough for the early adopter because that means you're getting that out probably as soon as you can to get the learning started to make it better so how many neos do you need in homes to have the right uh influx of data to to learn and improve the world model and embodied intelligence as quickly as you like is it a hundred neos is it ten thousand i don't have

SPEAKER_163: a good sense of scale the honest answer here is that no one knows because it's not been done before

SPEAKER_141: i can give you some first principle numbers yeah we know that we know that pre-training on a data set roughly the size of youtube gets you very far with respect to general intelligence now of course that data doesn't actually have what we call agentic behavior and it doesn't have physical interactions with like forces the inter the agentic behavior is actually extremely important so let me spend 30 seconds on it when we train on video you only know what's going to happen next you don't know what was the action that the agent took so like you know if i'm going to pick up uh my phone over there i think first of all i have a goal i want to achieve i want to pick up my phone so i know the goal and then i decide to take an action to go do it and then you see the result so these are like the three different types of data video only has the last one robot data has all three it has the internal state of a robot what was it thinking what was trying to achieve what was the actions it attempt

SPEAKER_163: decided to attempt here you see the result so this data is way richer so we hope we can get away with

SPEAKER_141: way less data but if you look at youtube as an example about 10 000 robots you will be having the about the same influx of data as youtube has so that's a good baseline that's not a small number

SPEAKER_161: but not an insurmountably large number because you could you could get half of youtube with 5 000 i presume it kind of scales up and down and the youtube corpus is so big like that's actually a

SPEAKER_160: pretty impressive amount of data inflow from just 10 000 robots compared to all of youtube

SPEAKER_141: and just to do the correct story it's not all of existing youtube but the upload rate to youtube which is growing is about the same as the influx from 10 000 robots oh okay i see i see i see okay

SPEAKER_143: well that's that's still doable that that's very doable it's very doable but it doesn't stop there of

SPEAKER_141: course like the goal here is to create something that is so intelligent that it can really accelerate the progress of humanity right you will have robots helping us build out all our infrastructure make sure we have enough manufacturing enough compute data centers power

SPEAKER_163: infrastructure everything is done sustainably because we don't have to cut corners on cost and labor why not do everything sustainably progress science like these ai models will not really help us solve all the outstanding problems in science or especially medicine and and without doing lab work right your ai has to design the experiment run the experiments qualify and make sure the data was done correctly iterate on this like that that that's how you do research

SPEAKER_141: and today it's extremely bottlenecked in that models are getting good at coming up with suggestions but they can't close the loop and check whether it worked well we have to go into the real world for

SPEAKER_239: that and if we're going to do that we're going to ape human institutions like laboratories and so

SPEAKER_161: forth so it makes sense to have a humanoid robot doing that it's just it's funny because when we're talking about neo today i'm thinking about having a helper in the house because that's what i need and i i'm fixated on that because it's what i want but in your example of improving the intelligence of these general purpose robots over time to do more and more things to me they're moving in some sense away from the home and into a dark factory or a dark lab where there's not humans per se it's just robots doing what we need them to do so it's interesting that we end up training them at home to be more generally useful out on their own it's almost like we're raising children as a as a class of robots and then sending them out into the world uh what's the time gap between you know neo getting good enough at the home stuff that you're like this is ready for everyone to buy one and when we can have humanoid robots helping us automate laboratory work to accelerate science is that a short time period

SPEAKER_141: or is that relatively long it's quite short um the home is actually the most complex really why it's it's such a good place to start because you have to solve it then you just like you jump into the water and now you need to learn to swim right so like um i do want to make sure that we really strive to satisfy all of our customers and the demand and that's kind of like what's constraining it now so exactly when we roll out in some of these other markets really depends on how quickly

SPEAKER_161: we can scale manufacturing okay so let's talk about that um before we jumped on you guys just opened up a new facility in san carlos you're gonna have i think you said design and even manufacturing under one

SPEAKER_141: roof absolutely everything on a one roofer so research development ai production service and in production we also mean manufacturing development and r d right we build the machines to build the machine um everything under one roof and this you know actually this is one of the things that i'm very proud of with one x and it's quite unique we literally go raw material raw materials even further than raw materials we develop new alloys and materials all the way off to the foundation walls on the i side under one roof and wow for zero to one technology this is incredibly important because the development speed is just going to be a function of how quickly you can iterate across the entire stack and if you're reliant on suppliers it's going to be extremely hard but even more importantly almost all of the great discoveries that we have made on how to make robots safer more affordable and just in general better they are between the lines of these disciplines of science right so it's when someone on the hardware side says like hey i see you're struggling with that if i do this i think your model will learn better uh or like oh you need these tolerances in assembly that's really hard you don't actually need that i can just use a neural net and we can learn how to calibrate this and you can like loosen up your tolerances and we can produce cheaper and it's like this incredible kind of almost like bell labs types think yeah of science and uh i think it's the most exciting it's world

SPEAKER_161: to work actually it's one of the few i think really good examples of when in-person work is not just better it's much better and i i like that you're having that effect but at this new facility in san

SPEAKER_134: carlos when you're at early stages of productivity how many neos can you make per quarter per year

SPEAKER_141: is it 550 500 the factory that we have in hayward that is now up and running fully and that will be the factory that delivers home robots this year okay that can do tens of thousands per year the one that we're building in hayward now can do hundreds of thousands per year tens of

SPEAKER_161: thousands per year is a lot given our earlier point about youtube but i'm curious is there enough early commercial demand from early adopters to keep your factory running at that pace or are you

SPEAKER_160: guys more looking to sell like a few thousand uh to get them out there to start the the faster

SPEAKER_141: learning process there's demand but of course the product quality needs to be there yeah so right now we're being very cautious and uh we're essentially doing very fast ramp of like a batch then we evaluate in the market because we do the point we're not at end customers yet are in homes right and things are under nda and stuff but like we of course we're testing this so it goes into goes it goes into homes and some older industrial applications and then we get data we figure out how well this works we figure out what's not what's not optimal we rev the design we do a quick batch again so we're kind of like running our factory at like full steam retool full steam retool and we're doing this until the product is rock stable you only get to do this once like this is the first time in history that anyone does this and ships robots like this and we want to make it right yeah that's also why we haven't really given a specific date on shipment we've said we're shipping 2026 and we will we're good on track for that but it is when it's ready because you'll do this

SPEAKER_161: once no no i i i that makes perfect sense to me i can wait six months it's fine i will not wait six years but i will wait six months um i'm a financial nerd bant so i'm really curious about the twenty thousand dollar price point and um everyone knows economies of scale bring down the price of things and so forth you're doing smaller batches but does twenty thousand dollars cover the bill of materials does it cover bill materials plus labor is it profitable is it super unprofitable i just don't know where to like where to peg that number in terms of economics for

SPEAKER_253: humanoid robots so i have to take you back all the way to the beginning then because if you don't

SPEAKER_141: think about the scale and affordability from day one how are you going to manufacture this at scale then you're not going to get there because you essentially get locked into a corner because you're making specific decisions about the design and actually even the technical direction so if i'm going to make a parallel i'd say do you want to make an electric car or gasoline car you're not going to make a gasoline car in small volume and then when you're like now we're going to ramp let's make

SPEAKER_163: an electric car no like you spent the last decade making the wrong technology it doesn't work so we

SPEAKER_141: really thought about this from day one there's going to be billions of humanoids on the planet

SPEAKER_163: meaning you need to think about is there any special raw materials in there that are rare

SPEAKER_141: even further like how much raw material is in there neo just weighs 66 pounds rather than 30 kilos almost a third of most of our competitors well that's a third of the raw materials

SPEAKER_244: that really matters when you're going to make a billion yeah because it's dramatically reduces the

SPEAKER_210: number of ships that have to bring gallium or whatever you know mining refinement like at a

SPEAKER_141: billion you're starting to even think about like oh that's a significant percentage of the world's aluminium and magnesium like how are you going to refine that right so my first thought was when you

SPEAKER_161: said billions was do we have enough metal available to us because and then i think about how many cars we make per year but it's not a it's not an insubstantial demand on you know raw materials resources and commodities like that's a lot of new product to build clearly it is so yeah but you have

SPEAKER_141: to think about that and then you have to think about like tolerances right how accurate does this need to be because it sets your um the cost of like refining essentially at these volumes you take raw raw material cost and you refine it into finished product and that's refinement step can actually be very cheap if you have the right output right if what you need is something that doesn't need to be that accurate so that has to be car part core part of your design and what we're doing here with these unique motors that we have developed that allows us to pull on these tendons instead of the

SPEAKER_132: classical gears one of the earliest discoveries of the company even before uh eve went to the industrial

SPEAKER_141: applications yeah yeah yeah this is like neo is just the nth generation of that technology right uh that is extremely lenient with respect to manufacturing like we can get away with various very few parts uh very loose tolerances no special materials and if you then do all this and you start also working really early on how to automate assembly and of course we are using robots in our factory robots building robots neos are building neos um text me a picture of that when we're done thanks then we we can get very low on cost that being said it's an incredibly complex system it's probably one of the most complex systems on the planet uh so um it's going to take us some volume to get the cost down to where we would make money on that and have a margin but it is a sustainable business

SPEAKER_267: which is the most important to me i'm just really glad to hear that because 20 000 is the point

SPEAKER_161: in which i can probably just buy one you know without without having to like do math but if it was 50 then i'm going to say i'd rather have you know a year of my kids college paid or something um so if it works at 20 hot damn that i can see you know the the real volume coming out and just being being tremendous i'm i'm really excited about this and i take it given the work you've done on supply and components and stuff you're relatively de-risked from china in terms of supply chain

SPEAKER_141: issues there are certain key components that's very hard to get uh magnets being the obvious one that everyone's ah yep now we're deep into magnets right because like even the first year of onex what i was working on was like designing new processes for magnets because the way we do our own motors um and we have some great companies there that we've been working with in in china on like co-developing these specific uh new ways of doing magnetics um we're working towards also ensuring that this is something we can do in the us yeah but and there are amazing programs around this that are getting spun up so i'm very hopeful that this will be something that we can handle in the future uh but it's not there yet uh it's it's one of the many things that need to happen but i mean outside of that most of it is raw materials that is generally accessible in open market right so copper aluminium

SPEAKER_162: etc steve it's cheaper in some of the countries in asia but it's also available here and you don't

SPEAKER_161: have to worry about boats shipping times tariffs and other things all right um last question before i let you go baron you guys uh initially raised from as far as i can tell like some norwegian investors you went ahead and expanded your capital base with investors from around the world uh how much capital do you need to get neo through its launch this year and then into uh its next generation um is that another couple hundred million is that billions of dollars i'm curious the short answer is we have

SPEAKER_141: enough money to ship neo okay uh the longer answer is i think the impact that this is going to have on humanity is not yet fully understood it's probably going to be one of the most impactful technologies ever everything we consume is physical in the end and intelligence becoming general and physical is going to unblock essentially a new way of life right um i am very excited to push that to happen as fast as possible because i think it can build a very very beautiful world where we all have what we need and where we can all focus on what makes us human instead of uh doing physical labor so that we can get paid and that's a long way of saying we will probably raise money because we cannot accelerate the path yeah but we don't we don't need it to to ship to consumers that's the reason i'm just trying to figure

SPEAKER_161: out how much you need like uh because if you were a sas company i could literally just like sit down think about your number of employees and your current cac and like i can do the math robots

SPEAKER_234: are so different so well let me say it like this like if you want to go max max here

SPEAKER_141: it's we need as much compute as everyone else right we're going to train the best models now we're going to train way better models than everyone else because we're going to have we're going to have better data i'm not saying we're going to be smarter than them there's a lot of smart people there but we're just going to have better data um and then we need to also build out all the manufacturing so it's going to be a very capital intensive race that's the that's a short way of doing and saying that uh we do have a very potentially unfair advantage here though which is that we can labor

SPEAKER_161: arbitrage our own scale because you're making your own general purpose robots no i i mean have you thought about just calling up jensen and saying look i want to i'm in a hurry i'm going to i'm going to buy so much compute you should definitely give me another pile of money so i can just go faster now like i i guess i'm i'm happily impatient if that makes sense i would like this future to become even faster so i i just hope that the investors out there get it and are like knocking on your door constantly trying to give you more money because i don't think we can get to this

SPEAKER_280: future fast enough i'm just i'm just excited i love the positivity it's gonna take them a future

SPEAKER_161: so uh come back on the show when neo starts to go out to uh non-nda people come i'm really curious to see how the reviews come and um i'm gonna start asking my spouse if i can uh if i can throw my name down but i think she's gonna have something more practical in mind but baron thank you so much for coming on and explaining this to me and walking me through it um if people want to learn

SPEAKER_282: more what's the website where should they go well next up tech clean and clean and simple this is a lot

SPEAKER_135: of fun thank you for the show i appreciate you