GPU clusters, venture trends, and the robotics startups we’re most excited about | E1977

0:00What are VCs known for doing in the summer and the winter? uh it is not hard work and graft it is mostly skiing and hot air ballooning correct they're known for doing nothing yes now i gave a big speech to you know one of our young guns here i call them the young guns of the ones who come from out of school doesn't mean we don't hire older people we do all the time but uh look we hired you alex i mean you're ancient uh I'm never missing my age on this show again. Just getting roasted. Getting roasted. I'm scared I'm turning three. You're supposed to be supportive. Well, that's just my numbers turned around.

0:38In a heartbeat, I would give up. I would literally give up every dollar I have to be 35 again. Really? I literally would go back down to zero to be 35 again. Every single dollar. This Week in Startups is brought to you by .techdomains. Don't miss our Jam Session with JCal contest. To apply and get more details, go to jamwithjcal.tech, brought to you by.tech domains. Vanta. Compliance and security shouldn't be a deal breaker for startups to win new business. Vanta makes it easy for companies to get a SOC 2 report fast. Twist listeners can get$1 ,000 off for a limited time at vanta.com. slash twist and open phone create business phone numbers for you and your team that work through an app on your smartphone or desktop twist listeners can get an extra 20 off any plan for your first six months at openphone.com slash twist all right everybody welcome back to this week in startups i'm alex he's jason x.com slash alex x.com slash jason and uh news three days a week that's our goal this week i think we have two we're still doing the interview shows we're still doing the liquidity pod every other week i think will be the pace for that and we're still doing ai with sunny so we're trying to figure it all out shout out to our sales team here at launch the venture firm uh and this week in start that produces this week in startups because they keep selling the show out and that gives us the ability to keep investing in it and make sure you are subscribed on all the podcast apps youtube all that kind of good stuff we got a great rundown today uh alex has been working with the team you know there's about six seven people on the team here so uh what's on the docket today alex and then uh welcome back happy tuesday today's tuesday yeah july is going too fast my birthday is coming up in a couple of weeks and i'm not i'm not excited i'm turning 35 which is not milestone yeah it's kind of like the you're actually middle-aged now moment um but that's now they're hearing you're there on the pod today we have a venture capital firm that is building its own AI, sorry, GPU cluster.

2:48I want to talk about why they're doing that, costs and also services. Then venture capital trends from the second quarter, including notes on where valuations are today. The data actually shocked me, so I can't wait to talk about that. Then one of J-Cal's favorite things, remote work arbitrage. And we're going to segue from there into robots and the Twist 500. Lots of videos there, so get excited for that. And I recently learned that you can watch this podcast in video on Spotify. If you didn't know. we were part of the video yeah we're part of the video um i guess uh daniel uh from spotify invited us to be part of the video uh beta like a couple years ago so we've been doing it we used to have to do one rss feed for like for on itunes you can actually get a video version of the show as well because they separate it okay and spotify kind of came up with their own system to put them together but it's proprietary and i told him listen i don't like this i told daniel straight up and i cc'd his bd guy when they invited us i was like i don't like you creating your own standards please respect the rss standards we have a video one we have a regular one just do it from those and they were like no we built our own thing and i'm like god damn it daniel if you keep breaking our if you break the standard i'm gonna call you out on it all the time so and he's like what do i care and i'm like oh fair enough but please spotify don't break rss standards don't break the podcasting standards that's what got us here yes podcasting works because of rss and if you don't know, RSS stands for Real Simple Syndication.

4:15It's literally designed to be simple, basic, easy to use, and works everywhere. If you create a new system, it's going to be worse. But let's start with what's going on at Andreessen Horowitz. Jason, we have talked over the years about how venture capital firms are expanding into services, hiring, business development, et cetera, et cetera, et cetera. I never thought I would see a venture capital firm put together 10 ,000 GPUs as the information is reporting that they then rent back to their startups, I presume at a very low cost. It's a project called Oxygen and it's the first I've heard about it and I'm blown away by this.

4:48So first impressions from you. It's a great idea that they stole. This idea was pioneered by Daniel Gross, who founded a really, really interesting company called pioneer labs i think he's a really really really smart cat um and what they did at pioneer labs is they kind of gamified startups or really more like product launches and like you could kind of score points move up a leaderboard for releasing products and having product velocity which has always been product velocity the key to startup success how quickly can you iterate on your product yeah that gets you more shots on goal etc so he kind of gamified that and created a competitor to y combinator he since shut it down because he's working at one of the uh ai companies um he co-founded uh an ai company i think with ilia so we'll get to that in a second but what they did was they created something called the uh andromeda is that how you spell pronounce andromeda if that's not right i've been saying it wrong my entire life so this was uh nat friedman and he uh got a bunch of gpus together uh clusters of nvidia h100s and they did this back in the day and um you know it got a ton a ton of press so here i'll share it there you go so there's a little bit on it here is the um hacker news on it andromeda cluster 10 exaflops for startups from nat and daniel you know they launched in june of 2023 so why is this important well uh the same reason why getting credits from azure aws google cloud oracle and other folks has been so powerful if you give startups you know 100 000 in credits 50 000 in credits they start it really does help because they would have to spend that so you're kind of letting them experiment if they do succeed they're going to be very thankful to you and maybe they'll buy more clusters from you.

6:45So this is just a little piece of candy. It's a little added benefit to your, you know, consider this like your Amazon Prime adventure, your Uber One adventure, you got a couple of extra features. One of the features is you with a 16 C, you don't you haven't built out your cluster yet, you can use theirs. If you're going to be a going concern, you're going to have your own cluster, obviously, or you're going to use AWS is but this is a nice little piece of marketing candy they stole from daniel gross and that uh so i was like give do you have credit but you mentioned it there right like people have offered hr services like recruiting sequoia capital is known for being great recruiters uh people did product days i remember sequoia as well when i was a sequoia ceo they had a product day where they would have some major company google ibm walmart whatever come to silicon valley and they would have 10 of their companies come and pitch them for 20 minutes each you know it's like okay and for walmart to come to silicon valley and meet 20 you know a bunch of walmart mid execs to come and meet 20 startups is like they don't even know these startups exist they would maybe be able to find three or four of them on their own so to curate that for them is really magical so this is um really nice of interest and horowitz to do big win for founders big win for interest and horowitz uh so i think it's a give them a plus on this it's it's not cheap though that's what i keep thinking about because their cluster, which they're going to scale to 20 ,000 in time, does include NVIDIA H100s, which are expensive.

8:16Now, I get Andreessen being the central purchasing point because they have a lot of money, they have a lot of friends, I'm sure they have connections at NVIDIA, they got access, they got allocation, great. But it's a pretty big capital outlay to buy all those chips and then keep them around and then also let people access them. They've built essentially their own personal, what, core weave, it sounds like. So it's not inexpensive. And so my question is, Are they funding this out of fees, like management fees? Or where's the capital of this coming from? Good question. You know, these go for$25 ,000 in GPU, ballpark, and H100, I think.

8:50Yep. So anyway, you can start doing back-of-the-envelope math. This is a nine-figure investment. They have$20 billion under management, something in that range. I remember when they hit$10 ,000. I remember when they hit$15 ,000. So anyway, if they have$20 billion under management, and you get a two and a half percent or two and twenty two percent management fee uh you know it's a lot of money uh every year coming in 400 million dollars uh so and some of those funds are sunset so they're at maybe half or one percent and they some of them might be front loaded with two and a half who knows anyway they get hundreds of millions of dollars in management fees this will probably cost them you know some amount per year because you would spread the cost over 10 years yeah and who knows what kind of round tripping craziness they're doing here we talked about round tripping yeah maybe this is like maybe they're clusters on one of their investments uh or maybe they're splitting the cost maybe they let you use a certain amount and then they'll charge you you know cost so maybe you get like you know an allocation um i see this as you know like uh marketing and pr yes uh more than anything and so if you're some team and you can get a meeting with them and they're like, hey, by the way, do you want to play with our clusters?

10:06And they're like, sure, I'll play with your clusters. Here we go. Boom. Okay, founders, the jam sessions with me, Jcal contest is heating up. I've seen a ton of interesting submissions so far. We've picked two winners and we're looking for three more. If you are a founder with less than$2 million in funding and you have one of those awesome.tech domain names, head to jamwithjcal.tech and tell us all about what you're building. If you win this contest, you get to come on the pod this week in startups and pitch me what you're working on. And then I'll give you some unfiltered feedback in real time.

10:40And you'll get your company mentioned on the number one startup podcast in the world. We're partnering with dot tech domains because super innovative startups already use them names like rabbit dot tech, Aurora dot tech and one X dot tech tech, even our own founder fridays.tech use this amazing new domain name so here's what you do very simple call to action you tell me about your awesome.tech domain and startup apply for the jam session with jacal contest today at jam with jacal.tech we're picking the last three winners soon so get in there all right i'm going to explain this using jason style logic so roll with me uh you're anderson horowitz you have 200 million dollars a year in inflows just call it half of the total figure we thought of that means you have a lot of capital to play with.

11:25You spend, I don't know, $25,$30 million a year on this cluster. You get into one or two more hot deals that pay out 15, 20X in your capital. And this is breakeven or profitable with only a couple of winners out of it. So as long as this actually does help your deal flow and getting access to either a lead check or a big check into other companies, the math can bear out. And it turns out Andreessen Horowitz, unsurprisingly, is doing tons of AI deals. Just since June, they've done Hebbia, which was$130 million round, Decagon, Mistral, Valor Labs, and of course, XAI. Now they weren't leading all those, but it does go to show that they have the deal flow, they have the money, and they have the cluster.

12:05The only thing we don't know is, is the cluster driving the deal flow or not. But I wonder if anyone's going to compete with this. Do you see Sequoia stepping up to do similar playing? I mean, you can rent this stuff. So I think this is like a short-term PR win. you know there i my belief is that um we will be overbuilt with hardware and infrastructure pretty quickly now i am in the minority in that belief but i've seen this movie before with fiber and server builds out and storage build outs when people see all that profit and they see all that money being spent a bunch of people rush in to capture that money and then what happens the margins get burned away so while i still think nvidia is just a tremendous company they'll have a lot more competition open source software other software solutions will reduce uh the need for more clusters they will become more efficient i believe i could be wrong obviously we're in uncharted territory here so i leave the possibility that i'm wrong and i am in the minority on this but i think we might get to an overbuilt situation in the next you know call it two to five years and if we're overbuilt then i think it's going to be like storage like do you hear startups talking about storage anymore no did we hear people talking about storage in 2005 to 2010 yes drop youtube google photos icloud everything was about how much storage how fast was it backup speeds and now it's kind of all been abstracted hasn't it so i think we'll be we'll stop discussing this hardware uh shortage and this hardware push probably you know towards the end of the decade it'll be forgotten goldman has already said in a in a recent research report that actually that the chip shortage the fact that people couldn't get even h100 that they wanted to buy them is easing already now that could be because nvidia has new hardware coming out that's going to be better but i'm curious if that shows that the arc here might be a little bit more compressed we might otherwise think but you know the depreciation on these things has got to be crazy once the next generation comes out so i wonder if you could even put that out over a 10 year time horizon or you'd have to do five years at which point it gets a lot more expensive because money is expensive now so you know they're in a lot of these things when people do press releases you and i as former journalists doing random acts of journalism here in commentary you know we've all been part of this where they you read the you know the details of announcements and it's like oh up to 20 000 clusters or you know eventually 20 000 like it's like they may have a hundred right now and they may have said like okay yeah if we keep raising more money on the management if we keep seeing demand we'll add to it the price with that so who knows what they're actually spending on this maybe they spent 10 million on it so far and if they win a couple of incremental deals then they add the next set of clusters the next set of clusters and who knows if they're like standing did they did any of the details say they were standing this up themselves or are they just renting them at aws or google or oracle and they have some relationship with dell or somebody i i think that they're hosting it themselves but i will double check that and get back to you on it but here's a question that i have thinking about entries and horowitz and its scale because you just mentioned their aum there's there's overall size and i think they went the ria registered investment advisor route over time so it is is andresen too big for the venture game is that why they're doing stuff like this to kind of just they feel like they're overfilling like the venture cup if you will they're doing so much more i wonder if we should not really consider them trad vcs anymore they're not traditional vcs in that they've gone for scale when you go for scale you hit the average right the good news is the average in venture is better than most averages so if you can deliver beta the average with a chance of alpha when i went out and raised this last fund one of my lps said you know you're doing 100 startups a year and i said yeah you know i think we'll get to 200 you know by the next fund i don't want to rush it i'm not like trying to trace the 450 investments but that yc are doing i want to do me you know i think 100 is the right number uh so i keep track of it two new investments a week you know it makes sense on average uh but you know i do want to keep growing and they said well the seed stage does better than the series a stage right so we've seen people try to get in earlier now doing that means a lot more work so you've got a front and center uh view now because you're on the inside and you came to the off sites and you know you're you're involved in what we're doing under you know obviously as an employer of the company under nda and everything so what i learned from that lp was hey if you just tell everybody you're going to hit enough surface area to hit the average the average for returns you know call it cash on cash uh multiple moik multiple uninvested capital is like three and a half for uh you know early stage and then when you get to series a it goes down to two and a half three and you know it goes down from there so hey listen you know you if you could actually make a convincing argument that being average and being an index fund of seed stage which is what y combinator is right it's it's i was just about to say that's kind of the goal it's a mutual fund it's kind of cool because then as an lp you're like well they may not hit this crazy outlier but they may not also return half my capital back or one x so kind of good because these lps like predictability uh that's what they're looking for and andreason horowitz has created a product for sovereign wealth funds for endowments that feels predictable so great job to you know ben and and the team over there and mark because they created the index and so if you want to put a billion dollars to work or 500 million dollars to work in venture you know going to sequoia if you can get in or going to andresen well you're going to hit the average or with sequoia you're going to do better and with andresen you're going to hit the average sequoia may not have room for you to put 500 million in uh andreason takes the 500 million right so let's say sequoia takes your 100 million but they don't take 500 and andreason takes the 500 now you're like well i'm going to get a better return on my 100 from sequoia great i'll do that and then i'll put the 400 with andreason i'll get it's not going to historically they're not going to hit the same returns as sequoia but they're going to hit the industry average or better hopefully and you still think that that they'll hit the industry average or better given the size of their funds kind of hard not to because they have a great brand there are great people like david yulevich and others working there who have yeah david's great um and people respect them and you know they they they had a really clever idea when they started it we're only going to hire rich people who are founders spit take okay and you know what i the back channel during the dei era of uh silicon valley was uh you know now we're in the mei era and we had this conversation last week i thought that was a very productive discussion now that like dei is out of favor but in the dei era people were like andreason harrits only hires white men that was like the vibe for partner yeah and remember he talked about partner washing like everybody's a partner oh yeah but you know if you only hire ceos who have already been successful that means they were successful and started their companies 10 15 20 years ago like ben or market so when you looked at the composure of their actual partners not the partner washing that everybody's doing in silicon valley you had you could see it was you know going to tilt a certain way because we had such a dominance of if you look at the big exits how many weren't men you know yeah historically over the 20 years uh so but the reason they did that was they didn't want to have to pay them big salaries ah i was going to ask so when david yulevich comes there and you know whatever david's worth i don't know his personal net worth but he did fabulously well david doesn't need a million dollar salary a year i don't know what they pay everybody but my understanding was they just told everybody just take a little bit of cash you're in it for the carry that will take that cash and we'll build the hr department we'll build this cluster we'll build marketing we'll build pr and you know he was very indresin was very enamored and influenced by michael ovitz so michael ovitz was there for a bit read have you read the book who is mike ovitz i have not read that book actually just read it's a great read it's one of my favorite long list of business books it's not a long read either and it's a great audio book for when you're walking with your weight vest so put your weight vest on get that zone two workout now that you're a 35 year old old man uh not not for another 13 days okay okay now i'm counting down at all you put it all together i think they then move those the all those management fees they don't need them like mark and theresa is not in for the management fees even if they have hundreds of millions of management fees it's not going to move the needle for him he doesn't need money he's post money so if you only hire post money you're in a good place so there's different strategies when i'm building my firm i don't hire post money people i hire young guns not that we do it based on age but i hire people based on i hire a lot of people out of school why do i like to hire people out of school and give them their first chance to be in venture because they haven't been corrupted with the venture capital system the venture capital system teaches bad lessons double click on that because i could i could unpack that different ways but i don't know exactly what you want to hone in on when it comes to that so what are vcs known for doing in the summer and the winter uh it is not hard work and graft it is mostly skiing and hot air ballooning correct they're known for doing nothing yes now i had a gave a big speech to you know one of our young guns here i call them the young guns of the ones who come from out of school doesn't mean we don't hire older people we do all the time but uh look we hired you alex i mean you're ancient uh i've never mentioned my own on this show again just getting roasted getting roasted i'm scared of turning three you're supposed to be supportive well that's just my numbers turned around in a heartbeat i would give up i would literally give up every dollar i have yeah to be 35 again i literally would go back down to zero for 30 to be 35 again every single dollar you could leave me with zero dollars and if i could be 35 again i wouldn't swap you all the money in the world to give up 18 years so the the reverse of that makes sense absolutely yeah yeah yeah yeah okay i hear that oh but going back to venture young guns so when you have the young guns you know i was just telling uh one of our up and comers who i just thought has great potential i said uh can i ask a question like how serious are you about this deadly serious i said great you want to be really legendary at this yeah i said can i ask you like did you have like a big weekend you go to a rave did you go away for the weekend did you have tons of like things to do and just no i said if i was you when i'm having brunch on a saturday when i'm having my brunch on a sunday pop out your laptop send do 90 minutes of work 90 minutes while you're eating lunch having your coffee i said you do that every saturday and sunday just put in that 90 minutes on a saturday 90 minutes on sunday i said you will look to me as the founder of the firm much different than the other 11 people on the investment team or any firm you're at because i will see this activity on the weekend and i'll be like whoa you really got an edge on not only your competitors inside our firm but against other firms i said it's the easiest thing in the world to win adventure work hard because the rest of the vcs are taking off six weeks they're all in italy right now the the joke amongst lps is they they're following their gps on instagram and the lps are in the office because they work at harvard or calpers and they get four weeks vacation a year and they're like i just gave this person 10 million dollar lp commit a 50 million dollar lk commit and i i can't keep track of all the places they are they're in ibiza they're in greece they're here they're there they're coachella they're they're at this thing stage coach around everything so this means though going back to the andreason point about hiring people who are effectively post-economic it means then that you need to find people who are wealthy going to be good at venture and are not caught up in the venture capital way of life because i presume andreason doesn't run a uh a relaxed venture capital they've never struck me as chill venture capital farm actually oh yeah and the reason i would say that is because here's what happens i mean as you if you're if you are truly successful in your career you have a network you're great at signaling and you're efficient so somebody like we'll go back to david ulovich since i don't know any of the other partners names there but and david is friend uh or we're friendly so you know you get someone like david he's got a crazy network and he's built businesses himself so when he can actually i don't know david i think works hard but let's just say somebody in his model did take off two months a year or even three months a year when they are effective they're going to be definitely effective and they're going to have deal flow in their network that's the most elite so you can make up for it with those kind of advantages got it but there's a big trow in between yeah and what is in that that valley between the truly elite people who have established networks and are really good at what they do and i would put myself in there because i do have a pretty big network because of the podcast and just hustling for 30 years and a young gun is people who are actually young guns who are acting like they're david ulovich or ruloff or me or sax or chamath or pick the person or jim brayer you know you have people who are truly successful at it they have the networks established you can't play the role of senior vc without having paid your dues that's what got venture into trouble yeah well the trough then the way to get across the trough is just endless work i think it is the is the summary there and actually this i know we need no substitute for hard work is a phrase for a reason well this is correct we need to get on to q2 in a minute but this has actually been something that i've been thinking about personally because you know having a second kid and i've been here long enough now that I feel like I have my feet underneath me and I got to know everybody and working on my newsletter.

26:39And then it hit me that I right now actually lack big goals for what's next. Cause I've been, I mean, when you have a baby, as you know, your life gets shattered into a thousand pieces and suddenly your focus is, is everywhere, but like advancing your career. Cause you're doing diapers at 3am, but I'm, I'm kind of almost, I think like figured out how to handle kids. And so I, I almost need to set new big goals. So I'm glad we talked about this, but yeah very simple goals for you which is you know showing up for three podcasts a week it's going to build your profile uh and you have a great start to that i would set a goal for doubling your follower count or just doubling your views per tweet just whatever the views per tweet are just try and double them because that's exposed so i double your views per tweet forget about the follower count because that could be spots and other nonsense and you can game it just double the view count per tweet which means you actually look at them and then decide decipher what tweets work and what don't so you're being thoughtful about your tweets as opposed to you know how most people do their tweets which is i just thought of an idea and i tweeted um and then i would just set milestones for the number of paid subs 50 100 250 500 and then put that on a calendar and then divide it by the number of weeks and so if you said hey i gotta get to a thousand paid subs in five years and you know it's gonna you know the first 50 are the hardest the next 100 is easier the next 150s easier and the next 250s easier you can actually just sort of map that out that's how i would do it because i put that i call that brick by brick uh and that's how i built my career i i my philosophy is bird by bird and lamont shout out if you've have you read and lamont's book bird by bird no oh so we read very different books is what i'm discovering i'm a non-fiction guy and i'm mostly a fiction guy yeah so bird by bird by ann lamont um is uh the subtitle some instructions on writing and life uh published 1994 and um you know it's basically about doing things one step at a time and it was because um it comes from a story of lamont's brother who was trying to write a report on birds and the father advised him to tackle it bird by bird one small step at a time um and then you break down and her general philosophy break down larger large projects like we're talking about into smaller ones it's it's simple but profound um and then you know it's like a little bit about sort of how perfectionism i always tell people don't let perfect um you know the enemy of the progress yeah ah okay perfection is the enemy of progress is the way i've said it i know there's other ways to say it but don't don't perfection is the enemy of progress you get that first and you actually did it with bianca take because she's doing the twist newsletter and you just said you know the first hitting the publish key the first times it gets easier each time is what you said that's a very am lamont thing the first time you hit the publish key is scary and hard and each time it gets one percent easier you're in the rhythm right so cautious optimism that's substack.com is that the uh uh that that url would work yeah go ahead go to cautious optimism and uh yeah i have to are you accepting payments yet because when i first signed up you didn't you can give me money if you want i gotta decide if i want to be a foundational member or oh no no no i'm actually a foundation member for two hundy i'll just never offer free you were about to offer me for free don't ever offer free i will not comp you don't comp anyone no comps all right let's keep on i did call my mother-in-law listen a strong sales team can make all the difference for a b2b startup but if you're going to hire sharks, you need to let them hunt and you can't slow them down with compliance hurdles like SOC 2.

30:21What is SOC 2? Well, any company that stores customer data in the cloud needs to be SOC 2 compliant. If you don't have your SOC 2 tight, your sales team can't close major deals. It's that simple. But thankfully, Vanta makes it really easy to get and renew your SOC 2 compliance. On average, Vanta customers are compliant in just two to four weeks. Without Vanta, it takes three to five months. Vanta can save you hundreds of hours of work and up to 85 % on compliance costs. And Vanta does more than just SOC 2. They also automate up to 90 % compliance for GDPR, HIPAA, and more. So here's your call to action.

30:53Stop slowing your sales team down and use Vanta. Get$1 ,000 off at vanta.com slash twist. That's vanta.com slash twist for$1 ,000 off your SOC 2. Q2 numbers are coming out. If you don't know every single quarter, a lot of venture capital tracking firms, your pitch books, your crunch bases, et cetera, compile lots of data. We use this to understand what's going on in the market that we can't see directly. And there's quite a lot going on that I want to dig on. First of all, it turns out in Q2 that venture capital totals were actually up a little bit. So Crunchbase News says$79 billion invested globally in the second quarter, up 16 % compared to the first quarter of this year and up 12 % from a year ago.

31:33does that match what you saw in the market in the second quarter looking at the funding reports from these different places you know crunch base i think your alma mater has pretty good data because it's not unlike some databases that are more closed it's a little more open so i tend to think it goes a little bit bigger than the other ones am i correct in that yeah it does it tends to encompass a little bit more um and we have pushbook data here as well but just just the the the improvement from q1 and the improvement from i would say anecdotally i'm seeing a pulse okay so signs of life i wouldn't say we're up and running and you know breaking that speed it used to be coming out of the accelerator we would have seven companies and five of them would raise money within the six months you know during the accelerator in the next six months got it sometimes we'd have all seven like in peak zirp all seven and the the two who didn't raise money usually one of them was because they chose not to and one of them couldn't clear market so the hit rate was pretty amazing it was basically and the one who didn't raise money was typically they were strong so they were making money they're like i just want to stop and raise money i want to keep building and grow my valuation so call it six out of seven could then during when the whole market collapsed i think we had a class where only two out of seven raised during the class and the other ones were like i can't even get meetings because vcs were licking their wounds they were just circling uh their portfolio and doing triage i can say right now the my time was 80 90 triage 10 new 20 new and i think it's kind of like 50 50 now so that's what you really have to look at is the percentage of free time a founder has a gp a general partner has to dedicate to new deals you do have to place bets but because the fund is supposed to be deployed over four years is kind of the primary investment cycle that's where you're planting the seeds and then you're reaping the returns over the next six years and helping them grow and then raising the next fund people were taking that four years and they were deploying in one and a half two that upset lps because that means they were coming back twice as fast which means they're not being as thoughtful by definition right so i'm trying to tell our team hey let's slow down if we have this many applications meet with the top x percent instead of doing 120 new meetings a week which is where we peaked we're doing we did 120 meetings new meetings in one week that's like a 6 000 that's a 6 000 meeting rate for the year i said at scale i want to hit 5 000 so we kind of exceeded my goal and i said hey for the summer i want you to go down to like more like 30 meetings a week you know like drop it down and then i want to have the summer of portfolio so i have everybody right now in the firm meeting with all of our portfolio and then categorizing them into a very simple process one two or three and it's the first time i'm talking about it i like simplicity and i like challenging the team to you know do these kind of projects so i told everybody july i want this done uh we started in june i want to meet with every founder and get a candid assessment of every single portfolio company from fun one to the most recent one and then i want you to give me a capsule and we'll we'll talk about the company one is they're growing like gangbusters they're raising you know year over year two three four x or more growth number three is they've run out of money and they can't clear market and they're kind of on you know sometimes founders call it going into cockroach mode you know they got 200 thousand in revenue a million in revenue the business isn't dying but they don't they can't raise capital they're kind of trapped and you know you get a good number of companies like that and usually those companies have to be sold they need to uh maybe do some management changes get reinvigorated or shut down right liquidate and that's a process for founders because it's emotional and then there's a big product market fit triangulation going on in group two which means their revenue spiky they have you know they might have 12 or 24 months of runway so they're it's not like they're imminently gonna go out of business but they're not tripling revenue year over year if you're tripling revenue year over year you're gonna clear market with investors or you're gonna hit you're gonna hit break even so we really don't have to worry about you we just have to give you support to grow faster or get to profitability whatever it is group three you're kind of sitting like in hospice in some cases and like hey this is going to get wound down what's our plan to wind it down it's kind of it's tough yeah and i spend a lot of time on that actually even though it's most people most vcs consider that a waste of time i spend time on that because i like to be there with the founders at the end if i was there at the beginning and hold their hand and say hey it's going to be okay and let's take a year off and then or a month off and then let's talk about your next company because i do think exactly a second or third time founder man hell hath no fury like that second or third time founder because they're like chip on their shoulder or they're broken forever they're not going to start another company it usually goes uh one of those two ways from lux uh chips on shoulders but chips in pockets i think is his uh i love it endless he loves to say that good on your point yeah if you're if your company just failed and you're pissed off about it i bet you your next one's going to do quite well.

36:52But then I take it in July that everyone's working with the companies in bucket two that could help, could use some help with some product market fit or customers, etc. Okay, juggling multiple devices and apps to run your business is a mess. We all know that. OpenPhone is here to make that simple. I have an open phone number. I use it to communicate with founders and it really works for me. I have a desktop app. Boom, I can go in there. I can do voice over IP in this beautiful, elegant app right on my phone or on my desktop. I use it a lot on my desktop, I'm being totally honest, because I have my headset on.

37:21And sometimes founders want to do a call, they don't want to pop open a video conference, my sales team loves it. Why? Because they get to keep their private phone number for their private phone number. And then they have all of their business stuff track in one location. So if they need to make a phone call to somebody they talked to last week, they can see their call history, click on it, send a text message and start a phone call with that person. And the ops team uses it when we have people calling, they have questions about their investments, LPs, etc. We have a round robin phone number. So it will forward the call to two or three people on our team because we like to have really good customer support and all this is easy to do with open phone it's super affordable at just 13 a month but twist listeners get an extra 20 off of any plan for the first six months at open phone.com slash twist what if you have an existing phone number with another service no problem easy peasy lemon squeezy open phone is going to port them over at no extra cost so head over to open phone.com slash twist start your free trial get 20 off you're going to love this product it is so affordable and it's so elegant just to the product team and open phone great job i look at products all day long and yours is elegant and simple and powerful well done you can help all three groups it's just different strategies with group one we can put them into our whisper network and introduce them to top tier vcs and say this is one of our breakout companies they're not raising right now i just want you to know about them and so that's that whisper network we created inside the firm in group two yeah you want to say hey what are you doing what experiments are you running and talk to them about product market fit maybe get them more focused a lot of times it's a focus issue they're working on four different projects you know one is their original idea two are like you know like projects they never should have started and then this other one is the one that's actually going to be the winner in the company so in uber's case that was uber x not uber black uber black did great but uber x became this like breakout global product and so you know then in group three i have a new idea that i've been workshopping which is i'm going to bring a group of them together i'm going to probably i'm going to start doing some retreats in austin so i'm going to do like a retreat in austin where i bring together a bunch of those folks who are struggling or can't clear market with investors but have some revenue which is a struggle in and of itself because it's kind of like a trap because you can't shut it down because it's got a million in revenue or two million and you can't raise money because it's only growing 10 a year it's like a real trap and see if we can come up with strategies for that group and the strategies for that group i think are going to be helping them find a a soft landing mna sell the assets shut down whatever it is or my other idea is what we might have four d2c brands and two of the founders are burnt out and we create a d2c holding company with the two founders who are actually inspired and say hey why don't we just have whichever is the strongest company with the best cap table buy the other three companies everybody gets some equity in those and then see if we can clean up things there and create a new opportunity if they don't want to do what's eventually going to happen which is an m a sale you know yeah i wonder if you should just put together like a special fund just for like the got to a half million one million arr and aren't going very quickly and like do like micro pe for those those companies and it's like like literally the smallest pe possible just put them all together there were people doing that like there was um uh i think it was tiny corp was doing like um andrew wilkinson's uh was uh trying to buy a bunch of companies and put them together and so yeah they have gone out and yeah here's tiny i don't know if you know tiny but he's a smart cat he's built he's good at building businesses and tiny.com he got tiny.co um so they started in 2007 um and seems like here they wanted to kind of do the war on buffett thing they have 11 companies they founded 40 that they majority own 90 they have minority investments in 900 global employees and 19 head office team so they're kind of like a holding company this is this is a brilliant idea i'm gonna look more into tiny because this is i've thought of this several times well you know laying there in bed trying to go to sleep like what would I do if X situation happened?

41:25But Jason, I do want to bring us back quickly to the Q2 numbers because there is quite a lot to talk about. So we have a chart of global AI funding through the second quarter of this year's charts in there. And as you can see, there is something going on here that is relatively indicative of some interest amongst the venture capital world for AI shares. Now, this chart that we're going to look at here is in fact influenced by the XAI round, a$6 billion series B. So that's in there. But I think that this jump in the amount of capital dispersed in the second quarter is nuts. This is just crunch-based data into AI startups.

42:05And I thought it was going to be more - A thousand deals, 20 billion. Looks like 23 billion or something. Yeah. The deal number will go up because deals tend to lag a little bit. The biggest deals with the most dollars get announced first and then the seed deals kind of fill in. So expect that to go up. But oh my gosh, that's an enormous amount of capital for one quarter for one overall sector it is it's because of the clusters like we were talking about before to dovetail with the interest in hearts one yep there's very few times where building companies requires massive and venture require massive capital outweighs we've seen it historically we saw it in the wars the competition between uber and lyft doordash and other ones those Those consumer subsidy wars, I'll call it the consumer subsidy wars, like the clone wars, that was indicative of like, let's use capital as a weapon.

42:55And here again, capital as a weapon, you need massive infrastructure. You're a big company like Apple, Amazon, Microsoft, Google, Facebook, and you can't do M &A. So what do you do? You do those two deals we've talked about where you buy teams and you skirt around LenaCon. Adept and inflection AI. thank you and then or you just build huge clusters and you you beat people on infrastructure and so you know it's it's rare that you have to have massive capital outlays and use capital as a weapon i can only really think of those two instances where capital as a weapon worked the only third time you could argue were the talent wars of the 2000s and so when google started just hiring everybody and letting them hang out hooli style on the roof invest that was capital as a weapon as well.

43:42We're just going to hire people for 250K. And you knew many of them, you know, here in the Valley. Oh yeah. I had friends that were making a lot of Google money when I was in my mid twenties. And I was like, you don't seem to be that stressed ever. How is this possible? Because my job was very stressful when I was that age in San Francisco. Now, before we move on to the twist 500, because I'm going to bring back that idea of capital as a weapon in just a second, but we have a table here that shows median US venture capital pre-money evaluations by stage i just want to touch on this jason because the numbers once again blew my mind so what this shows this is pitch book data uh their first look for the second quarter it shows us that as time has gone by pre-seed seed early stage vc and late stage vc median u.s pre-money evaluations are at all-time highs this year and we've heard a million dollars for pre-seed 2014 1.4 seed 4.6 in 2014 12 million now that tracks for me the uber and thumbtack data stacks rounds were all five under five five million or under um so i did all three of those for a 50 million dollar valuation and look like that would be the equivalent of like the seed round now now of course you know the the tech economy has grown significantly and the exits are larger.

45:00So I'm just saying that it's like when I look back in time and I realized like Microsoft went public and they'd raised 1 million in VC and so forth, like things used to be so cheap compared to where they are now. And early stage VC, the median was 12 million back in 2014. Now it's 45. And then late stage went from 30 to 68. But the thing that just shocked me is these are higher than last year and they are higher than we saw in most cases, even in 2021. And if you read twitter everyone's complaining about you know failing to make market and struggling to raise and so the thing that i just kind of wanted to ask is in in your view is this the impact of ai enthusiasm skewing the market's numbers for all startups or is it just better out there than i thought just given what i've heard from founders that i know and okay and such so remember um there's the group of cause and correlation um and survivorship bias these are the valuations of the companies that did clear market so by definition they're the strongest and so for the strongest to go from 2019 for 4 million to a 50 increase in 2024 over five years it's not exactly shocking to me it's a decent you know 50 more and so what you have to ask yourself is is the prize that much bigger i where i get concerned is early stage and late stage vc those two moments i think for those folks if they don't have product market fit and you invest it at 45 million it's incredibly hard to return your fund so i make all of our uh from the young guns to the managing directors when we make every investment i have them do a very simple calculation i'll do it with you here live on the air sure we're going to invest in a company at the six million dollar valuation let's say we put in six hundred thousand dollars sure let let's do it a five million dollar valuation five hundred thousand in a fifty million dollar fund okay put five hundred thousand in it's one percent of the fund five million dollar valuation now we own ten percent of the company so if that company becomes a unicorn we will have returned ten percent of that a hundred that billion dollars which is a hundred million we will have doubled the fund not so fast you're going to get diluted because they're going to do multiple rounds of funding so you just take that seed stage early stage investment whatever your ownership is assume you're diluted by 50 percent okay it could be as little as 30 it could be as much as 60 or 70 depending on how capital efficient the founders are let's pick the number 50 what that means is in order to return the fund if we invest one percent of the fund uh and we own 10 we're going to get diluted to five to return the 50 million we need unicorn to return the fund now let's do the same math for somebody who invested at a 50 million dollar valuation and they put in 5 million they own 10 and they get uh they have it out of a 500 million dollar fund it's a 500 million dollar fund they own 10 of the startup startup exits for a billion they own 10 it's 100 million they get diluted by half they get back 50 million yeah they've now returned 10 of their fund for hitting a unicorn this is why i was asking about andreason horowitz earlier it's so much harder to return venture style money hard yeah once the base gets large i mean entry price matters i tell this to everybody all the time entry price matters so just do that back in the envelope in order for andreason or in this fictional case of the 500 million dollar fund um which is like a series a fund would be like in that range 400 500 600 like a classic fred wilson fund at you know uh union square so you look at something like that and yeah they have to hit a five billion dollar company uh no they have to hit a 10 billion dollar company a decacorn a decacorn if they own five percent of a decacorn it's 500 million to return the fund but how many are there not that many and if you return the firm once over and your cash on cash returns are 1.0 x you're out of business you're out of business you're not gonna be able to raise a 500 million you gotta hit two to stay in business you gotta hit three to grow there you go rule of thumb and that's why nobody believes in venture right now we are in the darkest days of venture in my career uh since dot com era this is the darkest time it wasn't as dark during 2008 everybody saw that as an opportunity because that was like oh this real estate idiots did you know gave the last the last five percent of mortgages they gave were to people who should never have bought homes right it was like we all knew what happened there so they created a housing crisis they should have never given those people mortgages those people should have been renting or they should have bought smaller homes and they shouldn't have bought three of them they shouldn't have had people who were like uh cabaret dancers in vegas buying free homes on fake paperwork yeah i i uh i gotta say that was one of the funnier scenes in um the big short uh great thank you for getting the reference i got the reference uh you you twisted it slightly but i'm gonna let it pass i just i like to do cabaret dancers oh elite cabaret dancers yes special cabaret dancers you might say yes um i have read the big short so there you go business book that we both read now i have the data for you and what you just said so according to once again pitch book data we don't have this pulled up but i have it right here.

50:31US-based VCs have raised$37.4 billion so far this year. That's down from$81.5 billion last year. So it doesn't sound that bad because we're halfway through the year. It's on a$75,$76 billion run rate, down a couple. What people have to keep in mind is in 2022, VCs in the US raised$191 billion and they raised$177 billion in 2021. So right now, you're right. This is a very, very tough time compared to prior norms. It'll be interesting to see who's left standing after this particular round of venture shares stops because it does look extinction level for a lot of firms that maybe didn't hit that 2x that you mentioned and are certainly not going to hit the 3x that they would need to expand.

51:14It's going to be tough. I mean, we could take the whole number and say what the industry has to return to double. I mean, there's an interesting. so if i do that math and then i look at the exit market and i see no one going public i don't see a lot of big mna and the numbers just don't make sense to me but i've learned that the old quote about the market staying irrational longer than you can stay solvent is so true because i was worried about this back in like 2016 i'm like where are all these unicorn ipos and we're right back to it and everyone's still investing so i must be missing a trick somewhere in how this works if you say it was 37 billion was raised in one year 37 billion raised uh thus far this year last year u.s based vcs raised 81.5 so that 81 if it got deployed over four years would be 20 billion deployed a year but anyway in that vintage of 80 billion dollars getting deployed that's got a return uh let's call it two and a half 80 times two is 160 add another 40 for two and a half you get 200 billion 200 billion okay that's just uh you know uh an uber and a doordash put together seems reasonable au contraire mon frere you don't own 100 of the company you do not you own 30 of the company maybe 40 let's be generous and say the vcs own 40 of the company sure they did own 40 of the company and everybody else on 60 40 of 200 was 80 so you you know and then how many ubers and doordashes are there there aren't that many right and so it's venture math is tough this and this is why when we talk about vcs taking time off i'm always like your job sounds hard not easy you think this would take like all your time like you wouldn't have time why do they have a lack of sense of urgency is the question you have to ask it's the top well my guess is it's the time horizon that it's very hard to get graded two things time horizon and then there's a second one there's actually three i can think of after time horizon so what makes it what would lead a person in the pursuit of venture capital to not have a sense of urgency one is the time horizon you know it's like oh my god i'm not going to find out if i have winners until you're six seven or eight okay what's the second thing what makes it really cushy to have this job well you get paid management fees whether you're doing well or not that's number two is you get those uh those incredible uh management fees that you know kind of make it yeah um and i think well just because i was thinking ego just straight up yeah there's something around i was going to say entitlement um okay but there's a thing where you don't get fired from this job you just fade away you get kind of managed out okay you get a window seat you stay on your boards and such yeah nobody gets fired in this industry i mean you might get fired if you like punch somebody or did something horrible and even then it's like a golden parachute and you're spending more time with your family whatever the press release says i'm going to go farm alpacas in new mexico exactly yeah uh to change a career so like when's the last time you heard uh andreason hirowitz fired three of their partners for a lack of performance tell me the last venture you've never heard of one you've worked at crunch base and tech crunch yes i've never heard of a vc and now you work at this week in service actually that's not true i i have some friends who are investors and occasionally we speak off the record as friends and i have heard about how people get left off the next fund for example not invited to continue yes but unless you literally knew what was going on inside that firm it would be completely opaque to the outside because we're not i'm not getting the the key man contracts you know i'm not seeing that yeah when it changes it's hard to see externally so yeah it's funny though because vcs love to tell founders hire fast fire fast doesn't seem to apply are you saying vcs are hypocrites well i'm saying humans hold on i need to put my pearls on so i can clutch them oh my dear i'm not trying to single them out as hypocrites we've also discussed politics on this show so this does come up occasionally but i mean i i there's something that i've i know about that yeah i mean it's it's for something that i believe should be a much more intense i'm trying to create intensity and urgency inside my firm uh and it is counter to how the industry works remember i said i i'm not recruiting from like xvcs it's for this reason they think they can just f off they can frack off for four years make 20 bets and not be accountable to those bets i'm now looking at i gave each person on the investment team they're the primary contact for a certain number of companies and they are responsible for pulling those companies through so i found a precursor to exits the precursor to an exit is a pull through what's a pull through a pull through is when one of your companies gets funded by another venture firm at a valuation or a cap on a note that's higher than the one you invested you invested at five and like in fictional case before and then somebody else invested at 10 on a convertible note and then somebody did a series a at 30 or 40 that's something we can actually track in slow single digit years two three four years we can actually track that and i'm looking at the stats for my team i'm keeping those stats starting this year um and i remember i said before one of our big things in the in july is to do like a massive portfolio review they're giving me the capsules of every want i'm saying what's your plan for this company and how do we get this company to pull through and help the founder get to that next round of funding or help them shut the company down if they don't think it's you know if it's impossible to save and help them with their next company and so i found something pull through and i am looking at that like a hawk yeah this strikes me as the venture capital firm equivalent of the advice you gave to the venture capital analyst you were discussing earlier about working on Saturdays and Sundays.

57:35If you do the extra work, you will probably have the extra result. Speaking about extra work, you and I have decided that we're doing the Twist 500, which means that we now get to add a whole bunch of new companies. Let's do it. If you don't know everybody, the Twist 500 is our growing list of up to 500 private market companies that we think are the most interesting, innovative, or just fun to talk about. Mostly we're focused on quality. We'll throw a couple of weird ones in there. But today we're going to talk about robots because Jason, one of your favorite topics is remote work arbitrage. And we recently saw something that is beautiful, which is, and we have the video clip of this, a robot being used via teleoperation to stock shelves.

58:17And I'll bring this all together in a second, but John, can we bring up the video of the teleoperation robot, please? Here we have one nerd sitting in a chair with an Oculus on and a robot. You can see him moving. now jason you want to talk about this particular seven foot tall shelf stocking robot why it's a nightmarish it's a nightmare looking robot it's for some reason they made a black with a very thin waist and like a really thick chest with like pointed ears i mean it looks like something from black mirror and it is putting bottles on a shelf of soda pop and iced tea in japan it looks like um with a remote worker and this is uh you know it trended on x.com everyone was talking about it turns out this is an old clip but i thought this dovetailed with remote work arbitrage remote work arbitrage rwa okay and this is something you and i have talked about uh on and off the show for a while remote work arbitrage uh there's that reddit where people are i always call it overworked but it's over employed over slash over employed so everybody's looking at remote work from home or globalization and saying how do i optimize so here's over employed on reddit over employed tells all kinds of sordid tales some of them are true i'm sure many of them are fake of developers mainly but sometimes marketers uh working from home having three jobs and then figuring out when they have to be on a stand-up call and they're on two at the same time how to arbitrage that and how to deal with if they get called on but they have two laptops open and they're working the two jobs they're on two zooms and they know how to angle them and if they get called on simultaneously say i'm having an internet problem let me reboot they reboot one computer they finish what they're talking to the other one and they say my internet went out and boom they got to reset the router anyway they got the whole list of possible excuses oh i spilled my coffee on my keyboard i'll be right back and then they have a keyboard they're stained in coffee like they have all kinds of dashedly tricks so that's the employment side okay now let's go to the employer side do you remember i think you were co-hosting at this time uh there was a chicken shop uh making chicken sandwiches or something on the uh east side of manhattan and people went in there and there was a zoom call and they're like what what oh somebody left their computer on with the zoom call and then this woman from the philippines is like can i take your order yep we have the video of that here here is the chicken shop and here is the woman and you know it's zoom because they haven't taken away the little bar at the bottom that shows her zoom options so well it's a little it's a little bootleg you know it could be it's kind of bullet it's not refined it's it's hacked they hacked it together yes this is hacked together uh this blew my mind and then there's also another example that we were talking about before the show which is uh remote humans remote EAs, you love to talk about Athena, but we've now seen teleoperation of robots.

1:01:15We've seen people zooming in across the world to do face to face, and then also back of office help. So to me, the overall thing here is that remote unlocks quite a lot as long as you're not lying to your employer and trying to collect multiple paychecks. Let's call it what it is. Capitalism. Arbitrage. When I see overworked, overemployed, you know what I see? I see entrepreneurship. these people by you're so unhappy with them i'm like these are something great i'm not super i'm not super i mean i think it's unethical right okay okay it's obviously unethical um there's a bunch of issues with it where you could get sued and then there's you know ip issues this is a plethora of issues um and it does create a little bit of chaos and mistrust that would ruin remote work for everybody so there's that putting it all aside if you're a developer and you can do three jobs and people are happy with your output and you're as a manager too stupid to manage people correctly and say this person's working three hours and the other six hours they're working for two other companies and you're not smart enough to know that then you know i think it's on the manager right i think the problem here is there's a reason why there's i've i've never heard of a journalist having two full time jobs at the same time.

1:02:35I've never heard of a lot of industries. I think the reason why developers pulled this off is they have built a culture in which people will leave them alone to work because they've demanded protected time to have deep work focus. And to be clear, that makes a lot of sense. I'm totally here for it, but it also creates the ability for them to be less responsive and have it be totally normal. So there's a lot of blame to go around here. The thing that I wanted to talk about though, and the reason why I wanted to have all these different examples is to me it seems that the idea of tele-operation of robots is going to become quickly posse i still think we're going to have humans zooming into things i still think we'll have humans doing remote work from the philippines or from the u.s or from wherever yeah the tele operation robots i think is going to get uh squeezed out because it turns out there are so many companies today working on humanoid and humanoid style robots that were added into the 500 optimus is one human is one so they're studying what what what's your 100 correct so there's one issue of robotics and then the other issue and and learning and then the other issue is just arbitrage so we got these arbitrage examples athenawow.com if you want to get like a free month i think or a couple of weeks free for my friends over there i'm an investor in athena and we have two athena assistants working it's 3 000 a month 36 000 a year you can swap them in and out they're trained if you don't like the one you're working with you swap them out with another one they're unbelievably good and you know what i found is people who are taking those what i'll call operations jobs they're not just eas they're kind of like operations people when i hire operators in the u.s market they cost twice as much so they're going to cost 50 60 70 80 000 if it's in a major city it might be even more 90 100 120 in san francisco and most people are not looking to stay in that job they're looking to use it as a springboard so now you're paying two two or three times as much and then the reward for finding somebody great is they leave to go to another job so you have to replace them every 12 months or nine months or 18 months and then if you ask them to do certain things that might be above or below their pay grade in their mind americans are hard to manage because we are the one of the most successful driven countries in the world for other countries that are emerging and frontier markets the idea of getting a job with an american for 36 000 a year is just mind-blowing it's a lot of money a lot of markets i mean the athena assistance from what i understand they don't get the whole 36 but they're the 0.1 percent of knowledge workers in the philippines and wherever they're based putting all that aside um i do think uh you know we're there's something happening here where you know everybody in this global free market because of how good zoom is you know look at us here co-hosting the show from random states it used to be we'd have to be in the same city remember those days if you want to be a podcaster you had to move those days i i do miss getting paid at my job to walk places i used to like walk that's kind of fun actually coffee yeah because i would go outside and i would like walk down the street then i would sit and order a coffee talk to somebody and then i would to go back to work now i just like hop between zoom calls and my life's worse but yeah anyways anyway you're talking speaking to the socialization issue which is very real uh and it's it's having an impact on people so anyway uh this is a trend we're going to keep looking at remote work arbitrage rwa remote work arbitrage if you have examples of it email us you know give us examples any remote work arbitrage on either side of the equation we're kind of interested in it i i think i'm i'm seeing it myself i was doing um i talked about this with sunny on the ai episode this week or last week where i had um i was looking to hire people in austin for a job and i made a i asked it to give me the high low and average hourly rate of this position it was a domestic position and uh i said put it in a table and give me citations it did it now this would be a knowledge worker's job last year and a knowledge worker at 30 40 50 an hour in fact i might have used my athena assistant to do that for me right and now the athena assistant can do that as well so there's the the last piece i'll say on the rwa is what i'm seeing at uber airbnb google facebook and others which is 30 growth 20 growth year over year same number of employees i want everybody to think about that can you get can you grow a firm 30 year over year 50 year over year but keep the number of team full-time team members the same i think you can and i think it's happened for three years in a row now as part of that was these organizations were bloated so you're getting rid of all these vps and middle managers and product managers who maybe they should have never been hired again to dovetail the other story where he said people were using capital as a weapon at google to hire talent and keep them from working somewhere else.

1:07:34They literally, the co-founders of Google did that as an explicit strategy. If we hire you, you can't work for a potential search engine competitor or an advertising network competitor until Facebook came along and started outbidding them for people. Offering even more money, exactly. Even more money. But Facebook and Google are going to have the same number of employees for five years, I predict, and they're going to have their revenue go up. I think the number of employees at Google and Facebook five years from now will be the same number of employees but the revenue will have doubled. I think the way that I would phrase that bet is revenue will double, staff goes up by 30%.

1:08:11Okay, fine, sure. Yeah, but I mean, but your point stands. I'm not undercutting the gist that there is rising efficiency. And this, I presume, trickles down to startups. People now probably expect higher ARR per FTE at your average SaaS company. I'd say it starts with startups because startups are resource constrained, so they have no choice but to be efficient they that's always been their gig and so we see companies like podcast ai as but one example and they just do so much with so few people and so i think we're entering the age of it's almost like stand-up comedians uh or a solo guitar acts like bob dylan i've always been enamored by those why jerry seinfeld or um kevin hart i'm i'm friendly with kevin hart and i went to see him when i was in dubai and i was talking to him after the show and um i'd seen him before the show and i'd seen him after the show hung out with him in the green room was very nice he's such a just such a really charming guy uh and talented and i just was enamored by how he could go to dubai do three shows just him now he has friends in our entourage or whatever if he wants to have people around on stage it's just him it only requires him yeah it only requires him he could literally get off a plane private jet commercial jet go somewhere and entertain literally 10 000 people yeah at 200 a ticket on average hundred dollars a ticket whatever it is jerry seinfeld probably more i don't know who's who's who's the king of the day i mean there's these comedians who can just pack rooms and then there's you know if you know jack johnson or you know uh grimes a dj sure now a dj might need a stage show and there might be a little bit more to it but even still a dj could just show up with a thumb drive and a dope set and just crush it so i think that's kind of what startups are when compared to the big companies the big companies have infrastructure they have a campus they have all this stuff when they move they move you know they take big steps like a giant brontosaurus like some giant uh dinosaur they just lumber but when that when that foot lands it shakes the earth like aws launching a product right and startups are like the raptors you know little pack zip zip zip slaughtering what's around them and so the ai stuff uh that's where i learned about athena actually i was looking at the startup and i just saw startups talking constantly about having an athena assistant instead of having the proverbial jack of all trades or jane of all trades or they them of all trades thank you at thank you yes i'm trying to be inclusive um and in the chief of staff role this essentially replaces that i mean well what do you need that role for so the age of efficiency is here and there's a lot of remote work arbitrage and i think this is a trend for everybody to just be aware of um i don't know that it results in dramatically less employment i think what it results in is a global balancing of salaries so just oh yeah well interestingly enough the the rto movement is the biggest movement is the biggest i think piece against a harmonization of salaries around the world because then it says that you have to be in san francisco to make san francisco money versus making it available more broadly but that's neither here nor there that's return to office for those of you return to office return Return to office.

1:11:39Yeah. Yes. No, that's what we're here to do is to make sure I listen deeply to what you're saying and just make, I channel the audience and I heard RTO and I'm like, oh, return to office. I wonder if they know that. So return to office is a thing we talked about last week. Yeah. And we're going to talk about that for a couple of years. Let's wrap. Yeah. Let's do it. So we have 10 companies that have raised a combined$2 billion and we have a table. I'm going to show you here to run you through the names of these firms. So I have ranked them by capital raised. These are the companies we've added so far in the robotics category.

1:12:08figure, Brian Machines, Agility, Bear, 1X, Collaborative, Sanctuary, Aptronic, Mentee, and Oversonic. And Jason, I'm going to show you a tiny clip of each one and tell you why I think they are fantastic. Shall we go? Let's do it. I love this. Figure. Let's do figure first. This is the figure bought inside of a BMW plant, picking up a piece of car. I don't know what that is, some sort of component. And it's showing how it looks like a floorboard to me. Thank you. I'm glad someone here knows what it is and there's gonna be some text coming up here that says it is a less than three centimeter tolerance that is not great but i do think that we are still seeing these robots improve very very quickly figure has raised 854 million dollars and i think it's probably the best known startup doing humanoid robots that are powered by ai of course tesla has its own and so forth but not a startup not in the same way so yeah this is figure now let's go bright machines bright machines is the second best funded company that we've added in the robots category to the twist 500 437 million dollars raised and they are doing robotic micro factories informed by ai so not humanoid robots but modular micro factories that are powered by ai i think this is one of the coolest things i've ever seen period and they're doing various different types of manufacturing with them.

1:13:33This is like the thing you see in the three-body problem television show. Like, high-end manufacturing and like, on-site. Not off in some different country or somewhere else. This is like, in your warehouse. I think this is just - These are tethered. They cannot go walk around a factory. They're locked in some portion of the factory on the assembly line. Got it. And then we have agility robotics. Now, this bad boy looks human, humanoid. Okay, cute. That's generous. I like the big eyes. Big eyes always make you look cute. That's like a classic Disney thing. But watch it. Watch how its legs function.

1:14:09So unlike a lot of these robots that are trying to be literally like humans, this is what I would call human-ish. It might be cheaper, designed for warehouse work. A lot of these robots, by the way, are designed for in-warehouse activities because we don't need humans probably carrying all those boxes. I think it's adorable. The company has raised$178 million. dollars and um as it says there are different it has horse legs the knees are backwards so for some reason horse legs and it looks like a minotaur it's a minotaur alex it it's a bottom half is horse legs and the top half is human that's isn't that a minotaur no that's a that's a centaur a centaur a minotaur legs oh yeah well a minotaur is human down and then uh bull up i think maybe anyway it does look like it is some greek mythological uh animal the cute version uh then we have bear robotics and this is another one that i'm really excited about so jason just i want you to just watch 10 seconds of this and tell me how adorable these robots are because this when i was prepping this this made my day okay okay so they're little r2d2s with four one two three four three or four trays and they're zipping around with food on them i've seen these there is a movie theater chain uh there's one in the hillsdale mall in uh in the peninsula here in the bay area they're closed at that mall and they put the food into it and it drives the food to each of the individual theaters and then there's a runner in each theater who takes the food from that there and so when you're walking to the movie theater that hallway has these type of robots um these things do seem like for room service at hotels would be epic um and when i was at disney uh i went to like the pizza and pasta place you know how kids are and we were by space mountain um and star tours and they had reconfigured the cafeteria that you because they have a disney app now that has ordering in it like toast that startup and we should probably put toast make a note about toast um and instead of like having people order food you order in your app then it tells you to go to a certain stand and your food is being assembled there so the cloud kitchens movement the toast movement because of a lack back to the labor arbitrage that's occurring they're arbitrage charging what toast does in terms of labor arbitrage is they take the labor that was the business's responsibility and they put it on the consumer so let's make a note of that type of labor arbitrage.

1:16:46And so that's what's happening here. This robotic is part of that labor arbitrage. Eliminating - There is, there's a version of this called the Butler, the Butler, B-U-T-L-R, that I saw a thousand years ago when I was at TechCrunch on my first stint, and it was designed to do room service. But the reason why I like Bear Robotics is it's raised$176 million. So it clearly has enough of a commercial footprint to raise big boy capital. So I think it's going to be in and around the world. All right, next up, 1x Technologies. Another humanoid robotics company raised$136 million. These are on wheels, and they appear to be a little bit like those wavy guys you see in like a car lot.

1:17:28They just seem oddly stretched. I'm sure this is lovely. They cleaned up some coffee in this video. Just another example of a well-funded humanoid robotics startup. Moving on. So here the robot is going to sort individual things onto two different trays. And the whole shtick here, the reason why they have the clock in this shot is that they say can do it about as fast as a human. Now, clearly it's a test, it's a demo, et cetera, but I think it shows dexterity and speed and that speaks well for the whole industry. Next up, Aptronic. This is a special shot. Jason, I went through their entire YouTube channel and I found you the one in which it shows them shooting the video of the robot.

1:18:06So this is the behind the scenes version of it. I pick it because Aptronic has raised about 30 million and they have built a robot that can walk and can do things. So I wonder if the barrier or the bar to entry into this humanoid robotics space is maybe a little bit lower in dollars raised terms than I expected. And then two more, Menti Robotics. Here's Menti walking around someone's office following directions. She just says, follow me, Menti Bot, and then the bot follows. And then it takes, I think it scans to the office she's walking through to make a map of it for itself. And then finally, we have Oversonic Robotics.

1:18:43They've only raised 5 million euro as far as I can tell. I think they're Italian and it's going to, I think, get some other international kind of names on the Twitter 100. So that's 10 companies,$2 billion raised, mostly working on humanoid robotics. The thing that I took away here is AI plus robots is going to be awesome because they can think more, they can do more, they can just be more varied. And also there's a ton of different ideas about how to build the best robot electric systems or um pneumatic systems how to handle the effectors their hands so much cool stuff here this got me super stoked about like what's going to be coming the next couple years i just can't wait i i think it's a great start if you have more robotics companies for us then just you know hit us on twitter x.com slash alex x.com so jason give us your suggestions we're starting with 10 gives a really good overview what i saw in there are just general observations is um you know these things are slow they're very slow yes but they don't need a break so if they're half the speed of a human and a human works an eight hour shift and humans make mistakes and these things theoretically don't make mistakes or at scale they won't make mistakes they can beat us if they're half the speed of us they can beat us just by working 24 hours a day or 23.5 hours a day like the cafe x machines do um and the cafe x machines are faster and they are perfect so that's the thing we realized when we did cafe x eight years ago they told us like 40 of starbucks orders have a mistake in them either they're cold or they got it wrong or whatever i believe and that was like an internal starbucks number they quoted us like from some internal report like we got to get our you know we got to get orders tighter now i'm sure doing app orders again putting the work arbitrage on the consumer there it is another trend make the consumer do the work um you know is a great insight these things have to be faster they um are going to show up soon and everybody's trying slightly different footprints you have to wonder like if two arms are good why not have four arms um but they what they're trying to do is make humanoid robots i think for two main reasons one the general applicability goes way up when they have the human form factor why the world is designed for humans yes therefore any job that a human can do these would have the footprint to do them whether it's you know doing dishes or making a latte or working in a factory so okay mission accomplished there that's why they're doing it it's and i think maybe the second reason is so that they get adopted by the public the public accepts them right public acceptance i think is if these things look like us they act like us they are friendlier than us i think that's a big part of like the when they steal our jobs and they take our work we'll be like oh we used to do that work but you do it really nicely you've got really big eyes as jason yeah big eyes so you kind of look like a pixar or disney character so i'm good with it you can take my job i'll just sit home and listen i'll just stay home and collect my ubi or universal basic my ubi but this is really i think great collection i think a great start um and some companies are going it's a harder task to build a human one and it's a harder task to use an llm or general ai it's going to take longer to do those but when you do succeed it can do many things that's that's why i'm excited these are not this is not a dishwashing robot it's going to be a humanoid robot for your house that you can ask to do things and some of these um uh examples that we just saw do have the ability to chain different actions together to do a complex um i guess the robot equivalent of answering a query which is doing actions to fulfill the ask tell me what there was that famous humane uh it was like give me something to eat and it looked at a bunch of stuff in front of it and it took from a bunch of non-edible items the apple and handed to the person so it was kind of like oh yeah an llm could do that um or you know if you you could ask an out you can take a picture of what's in your fridge right now and say what can i make and chat gpt4 and claude and all these other ones can do that you can take a picture of a bunch of food and i'd be like you could make fettuccine alfredo you could make lasagna you know it'd actually tell you what you can make based on the ingredients which means robot plus llm means you could tell the robot what do we have in stock what can you make me and it's like well i can make you if that union happens on the time frame we expect and at the quality level we anticipate it's going to lead to an entirely different world for everybody so i'm hoping that that pays off and that's why by the way just in those 10 companies as i said 2 billion invested not a huge shock um before we wrap we are talking also about housing and construction startups for friday We're going to be back for more Twist 500.

1:23:32So we'll do the same format. 15 minutes, 10 companies. Yep. A little bit of commentary. We get them in there. We've already had 60 submissions. I think 60 or 70. And some of them, I'm already pulling them into a spreadsheet. Some of them have multiple votes. We're going to give extra attention to the ones that get the most community love because that often gives us a great signal for what people like in the market. But lots to say there. So I'm AlexW at launch.co. If you have a construction tech, a housing tech, or a robotics tech company, let me know. We'll talk about it. on the show. Awesome.