0:00Andrew Feldman, welcome back to Sorcery.
0:03Thank you so much for having me. It was a fun time. We We spoke last what? In Paris.
0:07We spoke in Paris. Yeah, it was like middle of the summer. Super hot out at the Ray Summit.
0:12No AC in Paris, right?
0:14No AC, right? [laughter] But that was super fun. And that was like a a good time in the market.
0:21I think it went down after that and then went back up and then little bit and then now we're in like the I don't know super cyclone of cyber security and just trying to understand what's going on.
0:35Been a tough week for airlines. Yeah.
0:41So, let me just ask you right off the bat then. What do you make of the current state of the market? Where do we go from here?
0:48Look, I I I am no sort of market prognosticator. I [laughter] I I I focus on on what we can do at Cerebrus and I and broader what what the AI space has. I I think um the opportunities in front of us in AI are enormous. We've just begun to scratch the surface of delivering value, right?
1:12We've we've talked a lot. There's been a lot of uh deployment of gear, but in terms of delivering value, we're just at the front end of this. And there are whole domains that are changing. I I think um for example, let's take education, something that's near and dear to my heart. For for 2,000 years, we've known how to teach children. You give them a tutor.
1:34Tutor sits with a child and works with them on their topics. I mean, this is the way Alexander the Great was taught, right? That this is the way sort of the the the uber wealthy have taught their kids forever. And now Chachi PT can do it. And it can do it really well. and you can ask it, explain it in a different way, and it can give you attention in a way a teacher that that's got a class of 30 kids can't. And it's a tiny little thing, but I think
2:03these things creep into our lives step by step. And the impact is uh going to be enormous. I I think those who who suggest that that the impact is is going to be little are are way off base.
2:18Mhm. There's also a great impact of how many jobs this market is creating. Someone made this really good point and they're like data centers take away jobs are bad, but it's like no data centers enable new business creation. And we're seeing this across the board.
2:34And there was a reason that that in many states that that labor unions were in support of data centers is because they're big construction projects. I mean obviously they create construction jobs in the short term for a year, two years, three years. They create a huge number of highpaying construction jobs, but they also generate small business jobs after via the AI. It's the cloud and AI are small business job creators like yours.
3:01Right. They they enable yours. They enable you to edit and to do all these things that would have required a whole team couldn't have done it. and instead it's empowered you and and other creators of of various types uh by the tens of thousands to to to engage and participate in the economy in a way that was otherwise foreclosed.
3:21So you also have a really good perspective on this because you've seen the evolution. We talked about this in our interview in in Paris, but you've seen the evolution of data center buildouts and you have new chips being developed. you're filling racks with all different types of new things and you've seen the evolution of that over time even like the switch from copper to fiber and so let's talk about that and like what what are the growing pains within that obviously like we've seen some things happen in the news whether
3:50it is a fiber cable or whatever but what are the typical kind of things the switching costs and you know the growing pains that come with it look I I think uh a working system uh is at its weakest when you try and change it.
4:08Yeah. [laughter] Right. Everything was fine with with the air traffic control system. Everything's fine. Um was it going to be fine for Thanksgiving or the the Thanksgiving after? They obviously thought not that it was time to do an upgrade of some sort.
4:23And at that point, you take a a working thing that's under control and you tear it down.
4:28And that is when a system's at its weakest. and whether they were trying to move from copper to fiber, whether they were reconfiguring routers I is less important. It's true when you upgrade software, it's true when you upgrade equip. It's true in all of these domains that this is a very there's real risk.
4:45That that you do damage and I I think um that's what happened in in those environments. And you know, pointing fingers is is not so helpful about the contractor or the the FAA. it was probably the right thing to try and do an upgrade, right? And uh they goofed it up. [laughter] And you know, it it happens more than you think when when you're trying to upgrade a customer from from one technology to another. Um so I
5:15I think the industry is getting better at data centers. They're getting better at building them. They're getting better at deploying equipment quickly. they're getting better at operating and managing them. I think the number of failures uh is is remarkably small.
5:34Now, we have had a a failure at Microsoft recently and we had some failures in others. These are so rare um that that they upset us like, well, I can't use it. Well, I mean, we've been using it for 10 years and never had a problem, right? So, they had a once in 10 year failure. Um, so I I I think some of these things are so successful and so reliable that when they fail, we sort of lose our mind, [laughter] right?
6:04Teams is working, right? And and we we lose our mind. It's like, come on, guys. It's it's been working just millions of hours for companies everywhere and and and it's a bad day.
6:16So, you don't have a tin foil hat on?
6:18I don't have a tin foil hat on. I I don't believe the world is fallen. I I don't believe that AI will will kill humanity. I don't believe I I think um I think that AI will be vastly net positive. I I think that that our lives and our children's lives will be very different 15 to 20 years from now because of AI and better.
6:43I I think that um I think I'm really proud to be in this industry.
6:50Yeah. I remember you know we talked about this but one of the benefits and you do have a really optimistic view one of the benefits is it's curing diseases it's curing cancer it's biological side of things and we talk a lot about you know defense but we have a lot of offense as well that that's right I I think that um there's a great deal of research um across multiple fields whether it's math or whe whether it's chemistry or or or whether it's applied research and
7:17drug design which is synthesized ides in the existing knowledge, finding holes in it, looking to plug it, and AI is really good at that. And I I I think it is really good at some of the most timeconsuming parts of research. And this will sort of launch us forward across the board in in basic research. And basic research concatenates into applied research and then into industry.
7:46And and so I I think that if um when when we look at a custom like Glacos Smith Klein, you know, a leading drug maker, um their use of AI is is extremely interesting and helping them develop drugs in the future. And then they're startups that are looking for small molecules, looking for drugs in new and innovative ways that that couldn't have been done by four 24 year olds, right? And and that's what's so interesting.
8:13So are you speaking on stage here?
8:15I am. What are you going to be speaking about?
8:17Whatever whatever Gavin asks me. I mean, you don't know yet.
8:22Well, we we like to we like to keep it a little a little spontaneous. Well, um but look, what what am I do? I'm going to show for my product, of course, in the company a little bit and so for my own. [laughter] Well, um look, I I think one of the things that's that's changing in the market right now is a recognition of how important speed is.
8:42I I think a year and a half ago some people were like, "Well, why do we need it to be fast?" And now they're like, "I use it every day. It needs to be fast."
8:48Faster. That's right. The first thing is that's cool. The second thing is, "Can I get it faster?"
8:53And so, sort of the the pendulum has swung our way completely because we build the fastest AI compute on Earth. Not by a little bit, but 10 or 15x. And so, as the the market recognizes how important speed is, the demand for our products is off the chart. And and that's super fun to be in.
9:10Mh. Um and so that's sort of the some of the the best things we you know as you know we we we have OpenAI as a customer but in the last 6 months we we won AWS and Microsoft and and Oracle pretty good six months.
9:27Congrats, right? Yeah. That's we've been working a long time for that and and step by step everyone's like whoa this fast inference thing that's a big deal and we need product for it. we need solutions for it and we need to embed it in in our offerings and so it's been a good run.
9:44So you did have one of the hallmark IPOs of the year. The year is almost to a close but three's IPO was amazing and we're seeing that with you know SpaceX Anthropic will IPO soon and you know maybe open AAI next year. Um but one of the effects of that and we talked a little bit about this you know before but is the wealth generation the wealth creation coming from it. All the employees who now have this bonus they didn't expect before now they have
10:12something they need to watch every day.
10:14So how do you handle that psychologically?
10:18Well, look, I I I think uh change is stressful and the vast changes in your your personal wealth, both up and down are really stressful.
10:30I think um it's further sort of made challenging by the fact that that your wealth is fluctuating every day and sometimes intraday.
10:39Right. And I remember years ago at a different company when our stock went down and and in one day my position in my personal wealth changed more than my father [music] made in his lifetime.
10:52In a day, right? And that's not easy. Um I think what we've done with our team first, we we we made more than a thousand millionaires.
11:03That's something I'm really proud of. I think that's our job as leaders is to to to make our guys wealthy. Um uh we talked about the fact that we're not in this for the short run. That the daily fluctuations are are not what they should watch. What we should watch is our sales and our growth, whether we continue to deliver engineering products that are extraordinary and that we're proud of and that over time the market weighs those things. And so that you you
11:33try and teach people that that these fluctuations that we have no control of um are are are not can't be watched.
11:42I mean for a while my dad called me every day and and said, "Why is your stock down?" Like, "Dad, that's not helpful. [laughter] Come on, Dad. I I know it's down, but I don't know. I don't know why. It's it's trading. It's got nothing to do with [clears throat] there's no new news in in the market. that the stock is going up and down and and you have to modulate when it's up your your psyche and you got to modulate when it's down and you got to think about what [music] your job is every day which is how do we
12:11build extraordinary product how do we make our customers happy how do we hire great people those are things that that pay dividends not tomorrow's stock but years and years and that's the way I think about it so for people who are just entering and they're entering in a very volatile market what is the one piece of advice that you have for them.
12:31Look, I I I I think that um most companies now, especially ones that have been around for years, have a a way for as a private company for you to sell a little bit, right? And pay yourself, right? I I think it's there's no reason for you to have 99% of your wealth in one stock, right? Professional investors don't do it. And and Mr. employee, there's no reason for you to do it either.
12:58Yeah. Right. And I I think it's okay to sell a little bit. I I think it's not a a statement on on the company. It's a statement on the that your situation is such that it's unwise to have 99% of your wealth or 95% of your wealth in a single company. That's number one.
13:14Number two, if you get a big cash, you don't have to do anything, right?
13:19There's no rush, right? You you can put in a money market and you can breathe.
13:24Right. I think there are a lot of people who will come and tell you what to do and offer to to help you for a fee. Oh, of course.
13:32I I I I think you can wait and and and just get comfortable with it.
13:39I I think um for what did I do the first time we sold the company? I I paid off the house.
13:48Um and uh but I I I think it forces you to confront some of your values.
13:56Right? How much you'd like to give to charity, if any? Right? Are you the type of person that wants a super flash car?
14:03Um, do you want to help [clears throat] your brother who's doing something different with his life? Or you want to help you want to put some money away for your kids? I mean, these are all things that that um take time and real thought. And I think the the best advice is you don't have to do anything at all. number one. And number two is it's okay to sell a little.
14:28Number three is once you have some cash, it's okay to do nothing with it.
14:32Right. And just it it it's designed so that you you shouldn't be more stressed.
14:38And certainly I've seen it have that impact on uh on people when they got really really stressed. Um th those are some some observations. Um it should make your life better um rather than worse. So, we'll focus on breath work and meditation.
14:54I'm the least I'm the least newagy guy on earth. [laughter] I like what?
14:58Mindful eating. It's like totally not me. [laughter] Um, but I I think sometimes there's a a drive especially in driven people that once something happens, you got to take the next step. You got And sometimes the right thing is to think.
15:13Right. and to stop and to to to say, "All right, I'm going to make this decision over a 90-day period or over a six-month period, and that will be fine." Um, and and so I I think those are some of the the things I think some thought about your values um is really really valuable at that time. Um, for many people, paying down the house is nice. Um, getting out of
15:41college debt or what, I mean, credit card debt, all that stuff is really bad. Um, some simple things make your life better.
15:50Who is one figure, maybe a couple figures that you've learned the most from in your career?
15:58Yeah. Do you know Mark?
15:59Oh, Mark was CEO of Veraritoss for many years.
16:03Um, was extraordinary CEO. Um, and is generally wise. I mean, it wasn't just that he was a great CEO and beloved, but he he's wise and he's been a mentor of mine. Um, Pierre Lemon, who is one of the great venture capitalists in our industry.
16:19Um, now at 95 or 96, still calls me and and asks me if I'm driving hard enough, [laughter] if I'm getting it done, if my if my stuff is late or or why I'm not pushing harder. Um, I learned an enormous amount from him about how to be a CEO and how to be a board member. Um I I I think those are two people who who made me better.
16:42Um and who uh sort of encouraged me to to think and develop my own sort of way of doing it. not do it their way but but their own way and introduced me to people who had different approaches.
16:56And and sort of encouraged me to to find my own way and my own sort of philosophy on being a leader and and and sort of uh entering the public markets and negotiating through it.
17:12Do you think most founders have to learn a lot of these lessons the hard way because they want to or they just are inclined to? Well, I I think generally founders are not afraid of the unknown, right? Because this is a a you know, we the old cgraphers, they wrote here be dragons, right? In the parts of the map where they didn't know where, right? When you set out on a on on an journey like this, like a startup, I mean, there'd be dragons everywhere. I [clears throat] mean, there's a ton of
17:41stuff you have never encountered, right? The first board meeting you sit in, you're supposed to lead, right? You're right. Like, what's that?
17:52Um, you've never dealt with investment banks. What if people want to buy you?
17:55You've never dealt with a sales situation. You probably never led large teams. You you've probably never had to recruit lawyers or accountants or a finance guy or you're a technologist, right? All these things are new. And um finding your way to learn, right? What what is your learning algorithm? Um, how do you deal with uncertainty? Right. You you you interview customers and you build product, but you're not exactly sure until you get it to them.
18:23Right. Whether they're going to love it or go, "This really sucks." Right. That's your fear every day. You get it to them, they like, "Well, Andrew, it's it's it's okay." Right. And, you know, it's like, "Oh my god, they hate it." [laughter] Um, and so you you have to to to figure out how to manage that. being the CEO is really lonely.
18:45Um, and you got to manage that. You know, some of us spend talk to our wives about it. Others when we go home, we we we want to engage on something else. Um, so there I think all of these things that that you have to sort of muddle your way through and hopefully you you're introduced to some people who who've done it before and who are sort of a good fit for you and sort of personality and style and they can give
19:15you little bits of advice. Um, and often those are advice are are either how to learn or things that worked or didn't work for them. And you think about it, you turn it over and you turn it over and why why you might have made that decision or a different decision. I mean, Mark Lesley really pushed me to be extremely open with with my team and completely transparent. And um while we were private, I would answer
19:41any question, any question on uh on our all hands. I mean, I asked questions about uh our position on the US military, about the war in the Middle East. I answer questions about how much I made, what my salary was, how much stock I had. None of this was public.
19:58Oh, wow. I I I chose it because I I think there is a fear that some stuff's being hidden, that the good stuff isn't being shared.
20:08One of the things Mark Lesley told me is he said, "Most of the stuff people fear is so dead boring that if they actually learned it, Yeah.
20:14Right. they would never think about it again."
20:16Right. And so I he I thought a long time about that and I think it's true. You know, people think in your staff meeting there big things being discussed and often times it's just painfully boring and if you let anybody who wants to come in, they're like, I don't want to come back to this, right? Why would I waste my time? I mean, talking about the Christmas party, I mean, what I mean? And and so I I think by sharing you remove some of the the
20:45veil that's over a lot of stuff. And it turns out that stuff is mostly ordinary, right?
20:52And as a result, people feel like included. They feel like um that that nothing's being hidden from them. And and that's a nice feeling in a company.
21:03And so it's sort of part of the culture you choose to that that was sort of something Mark encouraged me to think about um and sort of became my own.
21:14It's good advice. I think I think um it's served us well. We have really we have really really low attrition. Um we have people who've been with with me now for 25 years. We have uh some of the the been with me over four companies, five companies. Um that's stuff to be really proud of.
21:39It is. Okay. As we wrap, I have one final question, and it's the most difficult question, of course.
21:46What is your hottest take right now?
21:49Oh, I I'm a cold take guy. [laughter] I have I I I try not to to to I think of the of the cyber and of the the garbage that happened last week. It's just this two shall pass. That that was sort of a confluence of of unlucky things. Um this will calm down. um that that the airlines are cancelling flights and and everybody had a groundhold.
22:16I'm a million mile flyer on multiple a I've never seen that before. It's a once in 30-year thing, right? And so, um I I I don't think people should think of of the tough last week as sort of the the new normal. I I I think that that that's just a tough week or two. I think AI will continue to uh to rip. I I think we'll continue to have huge demand. You'll continue to see
22:44models improve. I think the open- source models will continue to get better and better. I I think the battle between open source and closed source will continue to heat up and I think we need both.
22:55Okay, good closing. Thank you so much, Andrew.
22:59Thank you for having me. It's always fun.
23:00Of course. See you soon.
23:01Bye-bye. Now, Keller, welcome back to Sorcery.
23:04Thanks, Molly. Well, really excited to have you here today. You have so many big announcements, I think, since we recorded. It was maybe last year. You've tripled your valuation, maybe quadrupled it. You're now at $21.5 billion. You just raised a $ 1.5 billion round. I think your total funding is over three now.
23:24Jeez. Okay. Get us up to speed. What's going on effect? It was mainly I think just you coming and doing that interview was Well, I was texting with Denton. And I was like, "Fuck, I wish [laughter] I invested. Every time I do one of these tours or whatever I do, these interviews, the companies skyrocket."
23:42Yeah. I mean, I I think that um I think maybe the weird thing about Zipline is that we operated, you you know, like we started the company in 2013 and we kind of operated for 10 years sort of in obscurity. Uh and one of the main reasons was that it was illegal to do this in the US, right? like we were trying to convince investors to give us money and they're like we don't invest in illegal things.
24:03Um but you couldn't fly beyond visual site legally autonomously in the US for those for that decade. And so for that reason Zipline launched outside the US.
24:12We had to learn how to build systems that were reliable, how to get to hundreds of millions of commercial autonomous miles, how to do maintenance, data logging, fleet command, build all the different uh hardware and software that was necessary to really provide this overall automated logistics service. And it was only um two years ago that we started operating at scale in the US. And it was really I think a year into that that it suddenly became very obvious, oh wow, wait a minute, like autonomous logistics is going to be one of the largest industries in the
24:40United States. And I think that that's what's kind of led to the acceleration and a lot of the announcements that we made over the last couple months.
24:46Tell us about the announcements. Uber Eats was today.
24:50Yeah. Well, we uh so yeah, just this morning we announced this uh partnership with Uber Eatats to uh grow with them to a million deliveries a day, which is about a quarter of their overall deliveries in the US. Uh that that contract will take us to almost every metro in the US. And it's really all designed around putting every single home within like 10 minute instant delivery of anything you want across this amazing group of, you know, 1.2 million merchants who today deliver with
25:18Uber Eats. Now, all of those um all of those uh brands are going to be available to hundreds of millions of homes or you know 100 million homes across the US. So, whatever you want, you can pull out the Uber Eatats app, order um and get it delivered in a couple minutes.
25:34And what about Walmart?
25:35Yeah, so last week we announced this giant national expansion with Walmart.
25:39We've been working with Walmart for the last couple years. Uh we had grown to 22 super centers. We're now expanding to 300 Walmart super centers in the first phase. And then scaling beyond that based on performance. Uh but that is a national scale partnership to enable super centers across the US with autonomous teleportation level logistics. You can now get access to 90,000 of the 100,000 SKs in a super center delivered in just a couple minutes to your home um in a in a way that is you know autonomous. It's
26:09faster. It's less pollution for neighborhoods. It's quieter and it's a lot safer.
26:14Let's talk about product development.
26:16What comes next? What are you working on now? How does Zipline just keep on getting better and better?
26:20I think right now, you know, [clears throat] the the crazy thing about the moment we find ourselves in when you visited us and we were walking around that factory, which by the way, that factory you walked around in has now quadrupled in size. We knocked all the walls out and like yeah, it just expanded.
26:36But um we we had built our operating plan as a company to do a million deliveries a day over the course of a couple years. And Zipline now has three plus strategic partners who are each trying to buy more than a million deliveries a day from Zipline. So it's almost like our entire operating plan for the company became like the unit of sale, right? And and so we had to take the operating plan, tear it up, throw it into the garbage, remake our plan for
27:04the next three years to say, okay, like if we're going to serve all this demand with these giant partners and scale, can we scale three or four, five times faster over the next two years? is really like the main thing that the company's going through. And I think that kind of growth, you know, growing 10 to 15% week over week, seeing you just we don't really even for charging sites, we're observing [music] each of these charging sites that we build, you know, in a metro like Dallas, we zipline has about 25 charging sites that are designed to serve almost every
27:33home in the metro. even the charging sites continue to grow like week over week and month over month and we don't really know where the ceiling is. But um trying to scale to meet that demand as a hardware company is just hard cuz you're you're scaling all of your suppliers.
27:48You got this massive global supply chain that you need every supplier to scale in lock step with you. If a [music] single supplier can't scale with you then the entire thing is screwed, right? And um we and then you have to scale, you know, you have to scale logistics, you have to scale manufacturing, and then you have to scale all of operations throughout these metros. And we're launching new metros as fast as we can. Uh we just launched Houston, which sorry, just launched Austin, which is where you and I are now. Houston launched a month ago.
28:16We're now at massive scale in Dallas. Adding more and more metros of the I literally was like, we're in Austin right now.
28:22Yeah. Oh, [laughter] yeah. Yeah. Me, too. Me, too.
28:28I I know the feeling.
28:31I want to ask, did you end up putting in a live cam?
28:35We need to do a live cam. I'm telling you. So, actually the the coolest thing I think right now about walking around Zipline factory and offices is seeing the fleet command center. Remember you and I looked at it. The rock. I thought it was a rock, but you're like, "No, remote operating command center." Yeah, I think now we just call it fleet command. But it's interesting when people see that.
28:54You know, we've had a lot of people visit since you visited. Yeah. We have like Secretary Lutnik for example.
28:58You can sit there all day. It's very uh cuz I think people still think of autonomous delivery and drone delivery as kind of like sci-fi vaporware like oh yeah people are talking about it but it's not really real. I think when people stand in front of fleet command for 20 or 30 minutes and you get to see every charging site that Zipline operates and then obviously we're showing the live stream of the charging sites plus you can see the sky map of Zips going out and making deliveries.
29:24Yeah. and you see it at like 1:00 p.m.
29:26on a Friday or 4 p.m. on a Wednesday or whatever just to get a sense. People really have this sense of like sort of profoundly like wo this is actually at scale and they're seeing a new layer of infrastructure that is going to become normal for humanity and we're just seeing this first 01% of it. So I do think yeah fleet command is the thing that I really want to like put on the website and just make like available.
29:49You know, there are a lot of nerdy people all over the country who love like looking at um you know, the flights.
29:55Flights. Exactly. Or like cruise ships.
29:57You can like see cruise ships all over the world. It's very like logistics is very just satisfying to get to watch to a certain kind of like factorial mindset if you know that video game. And uh yeah, so we definitely need to we need to design it so you can just like open it up in a browser and just see like the network doing deliveries day in and day out. So, where are we on teleportation stats?
30:20Well, um, you know, Zipline is now doing more than a million deliveries a day in our busy metros. More than a million deliveries a day in each metro. And, uh, right now our goal is to, as I mentioned, is to get to a million deliveries a day and then we'll we'll be at millions of deliveries a day in order to just to serve the commercial contracts that Zipline has already signed. I think that um in you know right now our average time of delivery once food is like prepared and ready to then when it's delivered to your home is between like 2 and 5 minutes and so it
30:49is super fast and you're having like an in restaurant experience at home a lot and then the other thing that we think about in terms of teleportation is not just speed but it's also just like how safe and quiet it is. Like a lot of our customers actually don't even know that the thing was delivered. They just happen to like look outside and like oh there's the package like sitting there.
31:08it just appeared magically in your backyard. Um, so for us, you know, talking about approximating teleportation as a company, approximating teleportation means we have to make it feel so fast that it feels like the thing was teleported, but it also means it should be quiet, innocuous, just magically appears in a specific place. And keep in mind like we, you know, we can deliver to any GPS coordinate. So customers aren't just ordering at their homes. You can like be at little league practice with your kid and you can have something delivered to your side. You could be I mean, yeah, we
31:36don't do boats [laughter] quite yet, but like uh you could deliver to boats. You could be you'd be hiking to the top of a of a mountain. Like I used to always hike Piestto Peak in Phoenix, which is where I get your water at the top.
31:46Yeah. You get to the top and you just like have an order and you just deliver it right there. You know, I'll show a video later today when I'm on stage of like I just I mean we I I saw we someone shared this video with me of it's like basically a huge uh like pool party that a bunch of college students are having on the seventh floor roof deck of their dorm.
32:07Yeah. And they have zipline delivering directly next to them at the pool on the seventh floor. And so I think, you know, uh, this stuff is just super different.
32:17Like it it interacts in different ways with people's lives. And, um, I think we're just sort of scratching the surface of seeing how it sort of changes people's overall behavior. Like people change their ordering behavior when you can get whatever you need delivered directly to GPS coordinates in a couple minutes. You don't have to batch process anymore. Take a huge amount of anxiety out of your life. You don't have to do annoying errands at the last minute. Um, it's a logistic system that works around your schedule rather than vice versa. So cool. Um I want to talk about the global
32:44stage. So you started globally. You started in Rwanda which is wild learning and testing ground. Now coming domestically and serving nationally.
32:54What are you seeing on I guess like there's a lot going on in the world. So like how are you seeing the world play out in terms of software, hardware, places to play, competition?
33:08I think that there is a I mean we we talked about this a little when you came and visited the factory but I think there is a huge competition playing out at a global scale which is you see so many of these economies growing fast their populations are growing incredibly fast whether it's Africa or other parts of the developing world and I think people have traditionally thought of these companies as hole countries or you know they are like char charity cases
33:35and that's quite unfortunate only what siding lowquality services like the US. It's also not very good for the countries. Like if you listen to the countries, they have been saying for a decade they want trade, not a they're sick of lowquality services provided by for free by NOS's. What they actually want are highquality jobs, you know, high-paying jobs, foreign direct investment, autonomy, uh, you know,
34:05robotics, technology, entrepreneurship.
34:07Like they want to be leaprogging their economies into the future. We're now living in this world where you you see like AI and robotics infrastructure growing incredibly fast. We're we're living during this time of acceleration.
34:18And these these countries know that they need to adopt those technologies to leaprog and to and to achieve economic independence and they're going to make decisions about who do they rely on, who's going to partner with them to help build out that infrastructure. And they basically have two options, the US and then a big, you know, global adversary of the US or, you know, at least global competitor.
34:41And so as the developing world is choosing like which technology stack do they want to build on? Do they want to build on a US tech stack? Do they want to build on on the Chinese tech stack?
34:47Like this is a huge global competition that's playing out that I think people don't really talk that much about. You know, we're really focused on like data centers and building data centers in the US, but the reality is that the vast majority of people in the world live in places that aren't the US or China. And I think there's a huge role for the US to play in terms of being economically competitive, helping these countries accelerate and and doing that in a way that would secure US economic and technology leadership for the decade to
35:18We're seeing some things happen every day now, whether it's bugs or cyber attacks or whatever you want to call it.
35:24But for a company such as yours which is vertically integrated and you've seen through the stack through and through, what do you make of legacy or new entrance that are patching things together and where those vulnerabilities sit?
35:38I don't know that I'm necessarily an expert, but I think that I think a lot of the levels of infrastructure for humanity are getting rebuilt uh in full stack and more secure ways. You know, I think that the old model of, you know, an air traffic control tower that might be buying software from like 20 different places and trying to like cluge it all together. Um, we are going to see more and more full stack solutions. I think SpaceX is an amazing example of like,
36:07you know, as full stack as it gets, raw materials in one side of the factory, finished product on the other, and then you're operating that product. You're not selling it to another customer.
36:15Um, full stack is going to be easier to secure. It's also easier to innovate because you control everything. So it's like if this part of the hardware is the bottleneck, then you just go and focus on that bottleneck and rapidly iterate and improve and increase bandwidth. And so I think that, you know, I think that we're both because it's where investor interest is focused. You know, capex used to be a four-letter word. Like
36:42investors only wanted to invest in social media companies and SAS.
36:47Yeah. for like 13 years of like zipline, you know, building a hardware company is the dumbest thing you could possibly do. Nobody was interested in robotics. Capex was like a radioactive zone for venture capital.
36:58And it's weird how fast, okay, so after 12 or 13 years of that, suddenly now venture capitalists are like, oh well AI is going to oneshot all of like, you know, the SAS companies. And so now actually the only thing that is economically defensible is hardware and infrastructure. I think maybe, you know, the thing that people aren't thinking enough about is like hardware is hard and it does take time and there's no level of like, oh, now we're just going to like accelerate from a hardware perspective and build a legendary hardware company in two years. I I think
37:27that hardware takes time. I mean, Tesla took 15 years to really achieve meaningful scale. SpaceX took 20 years to achieve meaningful scale. Zipline, we've been at it for more than a decade, you know, to kind of get to the scale where it starts to look relevant at a national level. And uh so I think that on one hand investors are right and I think a lot of the most valuable companies in the world that are going to get built over the next decade are going to be integration like complicated integrations of hardware, software and operations. But on the other hand, these companies take a long time to build and I think probably a lot of investors who
37:57are investing in these markets today don't appreciate the timelines associated with scaling up a factory or just ironing out like hardware reliability for a brand new product.
38:09It's a really good point. We did get quite used to fast, cheap capital for high margins and growth and now there's a lot more factors that's resetting what it takes to fund a business because a lot of these companies also want to be sovereign through and through. Y and so the capex is a meaningful point and discussion and how much money you're going to be raising and what you should build out yourself versus give to a third party. Yeah, a lot I I I think people, you know, there are a lot of great companies being built. Obviously,
38:38the scope and scale of ambition has increased dramatically over the last 5 years. But I think these hardware companies are going to take longer to achieve scale than people currently think cuz it just and it's all just like really unfancy stuff that honestly investors don't even know to ask about.
38:54They're just like, "Oh, but how do you make the actuator reliable? Oh, how do you, you know, handle like supplier industrialization supplier industrialization engineering for the motor that you're producing with this supplier? Like, do we have good enough end of line inspection? This stuff is like sounds boring and in the details, but it matters a lot. And I don't think that people appreciate that it's hard and it takes time.
39:17As we close out, I have to ask you the hardest question.
39:22What is your hottest take right now?
39:25Um, [laughter] I think, you know, I think it's what we already talked about. It's that like I think I think it's awesome to see so much investor interest pivoting into hardware and infrastructure.
39:36And I think that the vast majority of the the most valuable companies that are going to get built over the next decade are hardware companies or at least are going to have significant hardware components. But I also think that very few of these investors have actually any meaningful experience like setting up factories or building infrastructure or like permitting construction projects in the real world.
39:55Maybe SpaceX and Tesla investors have like some flavor. Um well obviously they do but I think that most don't and I think that you know the next three or four years are probably going to be a lot of uh real world learnings for those investors and for those companies of you know figuring out uh the pain and timelines associated with construction projects. I mean even looking at you know companies like OpenAI and Anthropic trying to build data centers you think that that would be a really at least I
40:25think intuitively I thought you know you would think of it as like oh yeah they can definitely do that like these are the best funded companies in the entire world turns out it's hard regulation permitting community acceptance and just the complexity of big supply chains where you have a lot of different components and the performance of the project is equal to the lowest performing supplier, you know, you are the you are the sum of all the parts and everything is multiplied against each other. So if there's a zero anywhere in the chain, then the overall
40:55We were talking about this before with the air traffic control centers, but it a lot of it comes down to humans.
41:01Yeah, still a lot of humans involved in these systems. Um and you know I think that that is we we are we are automating and like making these systems more efficient uh over time but in a lot of cases there are also really strong you know interest groups that are incentivized to keep things the same and so I think we're seeing that you know we're seeing that play out in a lot of different parts of the economy. I think it's it's healthy but hard.
41:27When did you know you were the power law winner? We still don't I wouldn't I wouldn't [laughter] I wouldn't I wouldn't say that that's something we know like Zipline is very you know you've been to our offices like we are we are scrappy we are frugal we are paranoid like the you know the company is working harder than ever we are not slowing down we are not patting ourselves on the back we have a lot of hard work to do and you know this is an area where the biggest companies in the world have been investing billions of dollars trying to compete directly against us to do what we do and so I
41:55think it's just Zipline has differentiated ourselves by being agile and fast and very special forces and we're not going to slow down.
42:03Keller, congratulations on all the crazy growth. I'm still kicking myself. One day I'll be an investor.
42:11And then you'll become a trillion dollar company deploying teleportation across the entire world and maybe the moon. I don't know.
42:17Yeah, we're excited. Uh we'll do all the planets. [laughter] Well, thank you so much.
42:21Cool. Thanks, Molly.
42:23Matteo, welcome to Sorcery.
42:25Yeah, thank you for having me.
42:26Absolutely. We're out here in Austin and you have one of the biggest cult brands in America. How's it going at 8?
42:34It's great. Uh so I was walking in. I always wear my hat and people uh stopped me really along the way and saying, "Oh, I have your product." It was actually a guy his his wife is sick and she was so happy to have it because, you know, she was feeling cold and so she could warm up her side of the bed. Other people were complaining they don't have the pod.
42:54They sleep here in Austin and they're missing it. Oh my gosh. So, what is the biggest type of feedback people come up to you with?
43:02That we saved their marriage.
43:04That is what they say. Yeah. All the times like 90% of the time you saved my marriage. I always fight with my partner. They have a different temperature. We were fighting but now you have fixed our problem.
43:14So, sleep saves marriages.
43:17Yeah. Yeah. [laughter] Yeah. And then others obviously they say they sleep great and they miss it when they travel and so they complain about traveling and usually they ask when are you going to start working with hotels. So [music] I will just go to that hotel and I will have my my same sleep treatment.
43:35So when are you going to start working with hotels?
43:37Sooner rather than later.
43:39Do you have any particular hotel brands you like the most? Um we want to be in premium hotels because that is aligned with the with the brand but um anything that is fivestar hotels is uh is our goal to be there similar to what kind of Starlink is doing with airlines.
43:58Uh I I see the same model for eight.
44:01So eight has been on a tear. I did mention you guys are a cult brand but you are a cult brand and you're growing really really fast. [music] People love the product. it demenstraably gives them better sleep. It makes them feel better.
44:14They live better lives. Um, what have been the biggest health hacks for you to get to this point and obviously optimize your sleep?
44:23Well, there are really two things usually is the number of hours that you sleep. I tend to sleep 7 to n hours per night. Uh, the other day I think I posted on X I slept like 8 hours and 30 minutes. That was a great night. And then consistency. So, it was really hard for me to wake up early in the early days, but now I train my body and automatically naturally I always wake up at the same time. I also go to bed at the same time.
44:48Do you like sync your circadium rhythm if you're flying in different time zones? What do you do there?
44:55Yeah, by now I almost don't have jet lag anymore.
44:58Yeah. I travel so much. I travel like 250 days a year.
45:01Oh my god. And now I'm really used whenever I go, like even today here in Noring, I still woke up at uh 5:00 a.m.
45:08So it's the equivalent of 6:00 a.m. in the East Coast. So I try whenever I can to stay on the same time zone. Even when I go to San Francisco where we have an office, but if I'm traveling to Europe or the Middle East, I try to land in the morning so I can sleep on the plane and then first thing I do, I go to the gym, I have breakfast, and then I can kick off the day.
45:28When are you going to build eight sleep [music] for planes?
45:31We were actually talking to a couple of uh airplane manufacturers and that would be pretty sick.
45:38I mean, they give you a little pad anyways. They might as well.
45:41Yeah. It could be an eight pad, right?
45:43Yeah. That would be pretty cool.
45:44Is it true that you bring one with you whenever you travel?
45:48Yes. Or in different locations. If there is one of our office offices, usually they install something for me so I can try it. The best is when I go to China or San Francisco. So, they're working on the latest generations. I always go and I try the next uh the next product.
46:03So let's talk about the scale. So how many offices do you have globally? Where are we at in terms of reach?
46:11So offices we have New York, San Francisco, Boston, uh we are in China, in Shenzhen and we are also in Milan in Europe. So those are the offices. uh we disclosed publicly that we have uh we are crossing a billion dollar in u cumulative sales with uh with our product. We sell in 37 countries launching three more by the end of Q1. So by then we will be in 40 countries. So we start serving most of the globe
46:40and you haven't even gone into the airplane market yet [laughter] and even retail everything is still direct to consumer which is pretty incredible to reach our scale just with a direct to consumer business. What have been the biggest lessons for you growing out this complex supply chain over time?
46:58That hardware is really really really hard. Uh then in the early days I mean we we we faced a lot of challenges. Now it's Caesar because we have a strong balance sheet and we have raised money.
47:09We have a larger company. Um but the complexity of hardware at scale in 37 countries is complicated right the whole supply chain warehouses where you ship when you ship is also really hard sometimes to predict like we just launched pod six we had a major spike in sales then you need to start predicting how the holidays will go and then how Q1 will go and so that part sometimes is more an art than a science and you just need to find the right balance.
47:37Let's talk about the holidays. Black Friday is coming up. We have Christmas. We have Hanukkah. We've got all of them.
47:44What should people be watching out for?
47:47Well, I think there is a a sale right now. Uh one of the best uh sales of the year, if not the best sale of the year. [laughter] Uh so they should check uh check check out everything on eight.com if they don't have one.
47:59If they have one, but it's an old generation. Now we have a tradein program and so you can trade in your old gen and we will ship you a new one at a major discount. So, it's a really good deal.
48:10What I know some people have a little bit of a hesitancy to purchase it because it can be a heavier purchase price-wise.
48:18What do you what do you see as kind of the barriers for that? And how do people kind of like make the jump to make the purchase? Like what is what typically goes on?
48:28Well, first of all, what we see is our return rate is record low for a consumer product. Like record low. Almost nobody returns our product. Our retention rating is insanely good, even better than companies like Aura and because it was the S1. So, our retention is better.
48:46Um, but you have a 30 days trial, so you can buy it, you try it, and if you don't like it, you can ship it back and no cost. And we just don't see many people doing it.
48:57That's awesome. Okay, so as we close out, I told you I was going to have a really difficult question. We could take this in many different angles. Maybe you have a couple answers, but what is your hottest take right now?
49:08Well, today I was on X and there was this this post about Jen now from Jensen. He and he was saying, "Look, if the border wants me back, I'm open to come back." So maybe he's coming back. That that would be pretty pretty cool.
49:22Um the other one it could be I mean wellness is a a category that is exploding from hospitality to residences. I mean everywhere in the world wellness is the number one topic and this is a trend that didn't exist when we started the company 10 years ago and now is a big tailwind for us.
49:40Are you seeing more competition or is it just very clear you have the power law?
49:46It's really power law. Yes, there are a couple of companies they are knockoffs. They try to copy technology steal technology but I mean we don't pay too much attention to them. the the key for us is to ship the best product to our customers and let them raise uh uh prices to the bottom.
50:04What's been the biggest lesson that you've learned building the business?
50:08Well, team is everything. It's something we hear over and over, but you really don't understand until when you screw up and I screw up. And then you understand when you have the best people, you can just unleash them on any task. And our best people are like incredible. That is the beauty of eight sleep because for the revenue size the the number of employees that work for us is fairly small. Even if you compare it to wearable companies um the ROI and
50:37productivity per person at leap is probably 3 to 5x higher than any other hardware company.
50:47Well, Matteo, congrats again on all the success and good luck on stage today. Thank you so much for taking the time.
50:55Brian Armstrong, welcome back to Sorcery. How are you?
50:59What's going on today? Anything major? I don't know.
51:02Well, we're at the Tomorrow X conference here in Austin and uh talking to people about Agentic Finance, you know.
51:09Yeah. Well, you know, Coinbase is really good at bringing frontier technologies to financial services like crypto, now AI, making it safe and trusted for people to use. So I think that it's really important for economic freedom actually that uh AI be decentralized kind of in the way that crypto decentralized the financial system. We don't want to live in a world where there's just one or two AI overlords.
51:29You know I think the people running those frontier labs maybe have similar thoughts and they've shared those. And so what's the antidote to that? It's we need a a decentralized marketplace of openweight models that can collaborate.
51:39And just like with human beings you know a team of people specialists can outperform a single person no matter how smart they are. And uh how do you bring teams together? How do you bring different comp agentic companies together into an economy? You need you need prices. You need mark you know free market uh economies. You need financial accounts uh for those agents. So that's what we're building now at Coinbase. Um an AI enabled financial account the payment rails so that agents can assemble into teams, hire each other, make payments in real time all over the world. And I think that'll help
52:08decentralize AI and preserve freedom.
52:10something that you've experienced quite a lot is cyber security threats and attacks um through the crypto era that was certainly one evolution of this now with AI agents and AI proliferating and whatever is going on in the world today financial services usually one of the biggest attack surfaces so can you talk through like what's going on today and what we could kind of expect for the next couple weeks yeah well I think trust matters more now
52:39than ever. And there's kind of a flight to quality in these moments. And so, you're right. Growing up in crypto, we actually became very good at this. We had to be very good at this early on.
52:46That's what's allowed us to be the most trusted crypto brand out there, storing more more crypto than anybody else out there. So, one thing I say to people a lot is that now that the attackers are using AI, we've got to be using AI on defense. And I think when we use AI correctly, we can actually it can benefit the the defenders, right? So, as an example, every line of code before it goes into production at Coinbase, it's being scanned by one of these cyber expert models. Uh, we also have designed all of our sensitive systems internally to be resilient to any individual
53:15engineer or employee having malware on their laptop. That's a common attack vector and people get tricked, you know, and so you can't um just assume that that's never going to happen. You have to design your systems to be resilient to that. And then lastly, I'd say, you know, we store 99% of all the crypto entirely offline, airgapped from the internet, totally disconnected, distributed, decentralized around the world. And uh that provides a lot of assurance to our customers as well. And that's that's what's allowed us to be, you know, providing these services to
53:43seven of the largest banks in the world, majority of the crypto ETFs, 140 government agencies, 50 of the top 100 largest hedge funds. Um, so we've they've put us through our paces on the on the cyber side and I think it's one of our real strengths for people who are concerned about this personally. What are the biggest tips you could give them for AI cyber?
54:03Well, I mean, a lot of the basics still apply. So, it's like, you know, use a password manager, right? Use two-factor authentication. Don't link it to your cell phone number because like sometimes your cell phone number can get taken over. That's a weak point. So often times people, you know, you should switch to something like a UB key ideally, but you can use um two-factor authentication on like an authenticator app. Um password manager for sure. Make sure you're installing all the updates.
54:28Um and uh you know these are good best practices for most people. And then for anybody who wants to take the next step um we offer services like uh you know in in Coinbase Prime for instance for our institutions uh in high net worth folks that are storing really really large amounts you can have multiple signers on the account. So it's like okay you'd need three of these five people to sign for withdrawal or five of these 10. You can you you can actually do a livveness
54:56check, right? Where because people want to try to use AI fakes now, right? So it's like, okay, turn left, turn right, and then it'll say, say the number 37, right? And then an AI agent would take a little a minute to generate that. Um, we also will sometimes just have people meet in person if they really want it. Like there's custom uh contracts we've designed for things like that.
55:18Awesome. Well, that's going to be super helpful. As we close out, I have to ask you, what is your hottest take right now?
55:24[laughter] my hottest take in any anything in life.
55:26Anything. We talked about bunkers.
55:29Yeah. I mean, well, I'm super excited about Agentic Finance. I think out I'll just give you one that's like outside of uh crypto maybe and AI for a minute, which is um I think that human enhancement companies are going to be a big deal um in the future. We're we're starting to see that like with the GLP1s, you know, um Eli Liy and Novo Nordis got a trillion dollars of market cap because of these these drugs and they started off treating disease, you know, but now they're being used by healthy people for enhancement. I think
55:58we're going to see other trillion dollar drugs uh come out as people sort of take control of their own evolution and biology and longevity. And it look, digital life is accelerating so fast.
56:08We've got to get biological life to grow just as fast, if not more. And so the the biotech company I co-ounded, New Limit, is doing a a lot of cool stuff in this area with longevity and I think we'll see more human enhancement companies um hit a trillion dollars in market cap.
56:22Amazing optimistic place to end it.
56:24Thank you so much, Brian.
56:25Thanks for having me.
56:28Did we get all that on camera?
56:29I don't know. Did we?
56:31Okay. We can put in chicken fried steak. Yeah. [laughter] All right. Well, Gavin Baker, welcome to Sorcery here at Tomorrowex.
56:40Thank you for having me, Molly.
56:43Well, I was going to ask you. I mean, I'm How's it going for you?
56:46It's phenomenal. We've had really great conversations.
56:49Started off with Andrew Feldman. We just had Brian Armstrong. We've had a lot of really awesome people.
56:55Yeah. So, who have you talked to on stage? You've been on stage all day.
56:59You know, um I'm a little bit worried that if I name them, I'm going to forget someone. So, I would rather not name one or two.
57:08Uh maybe the first person you talk to. you look I had a great conversation with Michael Dell, with Chendra, with Andrew Feldman, with Lynn Chow um from Fireworks. So, it's yeah, it's been awesome.
57:20What are the main themes?
57:22Oh, I obviously talked to SpaceX, too.
57:25Um just that demand is still really strong. I thought it was, you know, Catindra, the CEO of Stera, since they reported earnings two weeks ago, business has accelerated, things have gotten better. SpaceX demand for compute still very strong. Uh I thought you know the tone from Cerebras on you know the state of AI demand was encouraging and I think this is important to hear because you know
57:54stocks are trading like um you know the world is not stocks are not trading like the world's ending but stocks are trading like there is an imminent issue with demand and I just don't think that's what we've heard here today. Well, the stock market's a little emotional right now. They're not very practical. They're a little emotional.
58:13Your words, not mine.
58:16So, what should we make of the market?
58:18So, the market demand is good.
58:21I think the market is um I think the market is fundamentally doing a simple but from my perspective slightly flawed calculation. the market is is saying, you know, we were spending, let's just make the math easy, you know, whatever it is, a trillion dollars in capex. Um, let's let's say $2 trillion in capex. Um, and all we have to show for it is 80 billion in anthropic revenue.
58:54Um, plus the open AI revenue and you know, it seems like that's been reported to be around 40. So that's 120 billion and then like we'll generously say that Grock and Muse and Gemini and all these other things add up to something and you still get to something under $200 billion and that is the return on this AI investment.
59:17And I think that that calculation is flawed for the following reason. one we heard from Lynn today and she has a business that is growing really fast their top customer. is 13xed in the first uh nine or 10 months of this year and she is fundamentally serving open source models that she that Fireworks is customizing and specializing for their customers use
59:47cases and they're from my perspective they're the atrades portfolio company so I think they're the best one doing this but they're not but they're not the only one and I think if you add up this open-source ecosystem system. It is a huge amount of AI demand and because the margins are lower on these open source tokens, it drags more compute demand with it. And then I also thought Michael Dell said something really important
1:00:15today that they sell enterprises what they call an AI factory. And um in the first two years of offering AI factories, which is 24 and 25, they sold 3,200 of them. And they've sold, I believe, 3,300 in the last 9 months. So these are over 6,000 discrete enterprises who are buying GPU servers and associated equipment from Dell and running open-source AI on those servers
1:00:50and getting a return and it does not show up anywhere in that calculation that the market is doing. And so I've been investing in semis since the year 2000. Um, like I've been through many cycles and I've, you know, I've learned you, you could be surprised, but it does feel like there's a little bit of a discontinuity
1:01:18between really strong fundamentals in the present and on the ground and what stocks are discounting.
1:01:25And ultimately, you know, we're in a voting machine phase of the market to quote Warren Buffett, but ultimately the market is a weighing machine and I think as we turn the page into 27 and you hear companies give their guides for 2027 and you know the market I think the market's been very focused on second derivatives and the thought is once the growth rate peaks it's all over and that might that may be true but it may also be that this is like the railroad build
1:01:55out and it goes for 60 years and like let's just say Nvidia relative to whatever guide they're giving for the next year let's say they grow and this goes for most AI stocks let's say they grow 8% a year 5% a year 10% a year it's a cheap stock
1:02:20it's a cheap stock um so Um, I think we're in a voting machine phase and we'll see when we start the weighing. [laughter] Remains to be seen what's going to happen with Jensen.
1:02:34Sounds like he might be making a comeback. What kind of scenarios are you working through with the Nvidia news?
1:02:39Obviously, that's had some shocks. Um well um my friend Brad Gersner and I we did we sent a letter to the NVIDIA board asking that they take a very slow measured deliberate approach. There's Colette is excellent. I think I met Colette for the I think I first met Colette when she still worked at Microsoft. Colette is a superb CFO. Okay. She's been at Nvidia a long time.
1:03:08They have a strong business. There is no need to rush to hire another CEO. I the the letter we wrote essentially asked the board to take its time, be thoughtful, and we suggested um and has all the facts emerge about Jensen if appropriate to consider bringing him back as CEO. And the reason I think that's important, I've been doing this 27 years, and I've met one Jensen, and I've met one Elon.
1:03:42And you know, I think, you know, I I I remember the guy's name, Rick Patino. When Rick Patino came to coach the Celtics, he said, "Listen, Larry Bird's not walking through that door anytime soon. It's a new team." After Michael Jordan left the Bulls, they never won another championship. You do not lightly let someone like this walk away. And Jensen, there is nothing stopping him once he's in once you know he's going to testify to Congress.
1:04:11Nvidia is doing an independent investigation and he could go be the CEO of a competitor. And I think this is something that we should all think about. Um, and then fundamentally, one reason I was comfortable writing that letter is I've known Jensen. I first met Jensen in the year 2000. thousands.
1:04:31Yeah, I have known him for 27 years at this point.
1:04:35In a lot of ways, I kind of feel like I grew up with him.
1:04:39Um, I grew up with him as an investor and I know him really well. He is a friend and I would just say, you know, and I've been with Jensen all over the world, many situations.
1:04:51He the only thing I've ever heard him say about China is the absolute importance of America beating China and the only thing I've ever heard him say about Chinese hardware is the absolute importance of making sure that American AI hardware dominates and we do not let an alternative Chinese AI hardware ecosystem develop because that AI could be given to our real adversaries
1:05:18Iran, North Korea, Russia. And China's been working hard on this, but the policies that Jensen was advocating for, I think we're essentially guaranteed to keep America in the lead in AI for the foreseeable future. Now, we're going to do new policies.
1:05:40And I think those policies carry enormous risks for American national security because I think the odds that China who is a formidable competitor now that they are forced to develop their own ecosystem. It's not an option.
1:05:53Before the Chinese government was trying to encourage Huawei, ironically restricting the imports of GPUs. Um but Nvidia GPUs less powerful than available in America. Jinset said America should have the first, the best, and the most were available in China and the labs preferred those. that was slowing down Huawei because to make the hardware better, it needs to be used. Now that Huawei hardware is going to be used by every Chinese AI lab. Like it might get better really fast. We don't have Jensen
1:06:20leading Nvidia. And I think as a result of misguided policies by American patriots in Washington DC who think they are doing the right thing, they may be I think they are taking grave risks with American national security. And just going back to Jensen and one reason I felt comfortable like I know a lot of CEOs. I do not know a single CEO who is more patriotic about
1:06:48America than Jensen. I simply do not believe he has it in him to betray our country. And you know he came here he was an immigrant from Taiwan. He came here when he was 9 years old. He went to reform school. It was the um most was the most expensive American school his parents could afford. But it was a reform school. He was 9 years old. He had a 17-year-old roommate with knife scars covered in tattoos. He worked as a janitor and he stayed there all summer because his parents could not afford to
1:07:16bring him home. Think about the journey he has been on from that starting Nvidia to having it become the most valuable company in the world. He's always quick to say it's only possible in America.
1:07:29And I'm so grateful to America. So I think he's a patriot. And based on everything I know about him in 27 years, I do not believe he would ever knowingly endanger our national security. I think the policies he was advocating for were pro-American national security and we may be about to learn the wisdom of them to our chagrin.
1:07:51What do you think the biggest unintended consequences are from dethroning him?
1:08:00I mean there are so many um I think it's very important that we have a vibrant American open source ecosystem. It's very important to me. I do not want to live in a world dominated by one, two or three frontier AIs because these AI systems are imbued with value systems. And one of the things that is amazing about America is the rich diversity of value systems we have here in America. And we have way more
1:08:29than two or three. And the way we reflect American diversity and all of its rich richness in this new technology is to have open-source AIs that we can something that's very important to me. I have a daughter, she's 8 months old. I want to work with fireworks and I want to take an American open source AI and I want to fine-tune it on the books, movies, and TV shows that shaped me and
1:08:58my wife Becky as humans. Star Wars, Star Trek, Dune. Maybe Becky would have some other ones. [laughter] Those really shaped you as a child at 100%. Well, we're [laughter] going to You got to start early.
1:09:13Oh my god. You got to start early. Uh I read Dune as a very young man. Lord of the Rings, The Hobbit, you know, loyalty, courage, great, perseverance.
1:09:22Yeah, we actually I've, you know, uh I see is with AI, you can kind of make baby versions of these books and and they're actually baby, but you know, Shakespeare, Homer, like the Western cannon, you know, everything that has made Western civilization and America great, you know, the, you know, Greek philosophers.
1:09:40I want a Baker family AI trained uh you know on these value systems on an American open source AI and Jensen was the biggest proponent of American open source and I think America needs that and I I hope that the industry um if you know I hope the industry is able to find its voice until Jensen finds a place whether that's back at Nvidia his CEO
1:10:10or CEO of another company where he could be an equally effective advocate. But I think open source and then I do just think the relationships that Jitson has with the Asian supply chain are unparalleled. Like if you add up everybody else's relationships, I don't think they are as good. He's never had a contract with Taiwan Semi. It's been reported that Morse Chang actually long ago asked him to be the CEO of Taiwan
1:10:38Semi. And so, um, it's going to be hard to replicate somebody who can go to Taiwan, speak the language, the native dialect, um, have a handshake and walk out of there with more volume than anyone else at lower prices. And he was, you know, everything in AI has become controversial in America. In Taiwan, Jensen like He is a living deity. They're they've
1:11:06lit up the um not a living deity to me here in America, but holy in Taiwan, I think they've lit up all of these buildings and these highways with signs saying, "Bring Jensen back."
1:11:19They're not doing that for anyone else.
1:11:22Jensen is singular. So, I just I hope the Nvidia board takes their time, lets all of the facts come out. They're going to do an investigation.
1:11:33Jit's going to testify to Congress. Jet has said he's going to cooperate with everything. Let's let all these facts come out and then I hope um that an intelligent decision is made because I will just say um you know I think it's well understood. I wrote the letter and people understand that I've known Jens in a long time and a truly vast number of people have reached out to [music] me. Nvidia employees, shareholders, partners
1:12:03saying they want to back his CEO. And ultimately, um, you know, if Jensen is cleared of these, you know, this national security innuendo, he has not been charged with anything. But if this innuendo turns out to be innuendo and he's innocent, which is what I believe, um then I hope the board considers bringing back a CEO because the shareholders that have
1:12:30called me, that is what they seem to want.
1:12:34It is quite the developing situation indeed.
1:12:38So apart from Jensen, you are very close with Elon and SpaceX. So they had a very big weekend. I know there was some drama over on the NVIDIA side, but on SpaceX's side, all green lights, amazing things, the launch, the catch, Starship, Starship, Starship 15. Yes. So, give us the latest.
1:12:56Um, I mean, it was amazing. Um, you know, uh, Walter Isacson, Elon's biographer, said he was in the room with the SpaceX engineers where Elon said, "We're going to catch this with these AR the control tower arms." and called the Mechazilla and the engineers, you know, I guess you know, Walter reported they thought it was impossible and we've done it and rapid reusability. SpaceX is the most
1:13:26important national defense contractor for America. Um it is um you know in war if you control the high ground it's a big advantage and ultimately controlling space and the orbital planes that is the ultimate advantage and America controls space because of SpaceX. So I think it is a huge moment for America. SpaceX is 85% of mass from orbit. I think the re rapid
1:13:55reusability of Starship it means we're going to see orbital compute. Brad earlier today said they're gonna put up the first orbital comput satellites in 2027. You needed Starship for that. You're gonna have Starlink direct to sell. And that is going to be amazing.
1:14:08And you know all these people who talk about Elon, if you add up all the climate activists, they've effectively, I think, been negative for the climate because they annoy people. At least [laughter] here in America. Elon's done more for the climate than all the activists combined. And then if you care about low-income people, if you care about people in the global south, what Starlink is doing for those communities is transformative.
1:14:37And for whatever reason, SpaceX doesn't really talk about it. They'll say, "We connected every school in Honduras," or, "We connected every school in, you know, the, you know, you know, the um the Amazon region of Brazil. or we saved this person who was lost at sea, but they are tangibly saving lives and making lives of the poorest people in the world better by connecting them to the modern world and all that comes with
1:15:06that education, healthcare, knowledge, everything. So, I think SpaceX is doing a lot of good things. I think the um the catching the Starship that was amazing.
1:15:17It was a huge moment for um SpaceX. I will just say having spent a lot of time at Starbase, the SpaceX engineers, you know, all of these experts online are very skeptical about orbital compute and oh the cooling and oh this um and you know to the SpaceX engineers, they have solved those problems. They're operating 11,000 satellites today. Um, Starling V3 actually consumes a significant
1:15:46call maybe 10 to 15% the power of like an orbital compute satellite. It has to be cooled. Um, the compute isn't as densely packed but you can do things different things in space. They regard the design of the orbital compute satellite that in their mind I think they've figured that out. what they're actually stressed about is Starship and um reusability and so catching the upper stage of
1:16:14Starship that's a huge step towards making Starlink direct to cell orbital compute all of that a reality and I also think Grockbot it's by far the most valuable AI tool to me I now spend more almost more time in Grockbot than I do on X which I thought would never happen Really?
1:16:34Is this like a good This is like a a supplement to being addicted to the feed. You just use Grockbot 100%. I cannot recommend it enough.
1:16:41I guarantee you it would really make every aspect of your professional and personal life better. But I think Grockbot has a lot more traction than people realize. And I think when I think the size of SpaceX's first party AI business will probably I don't know it but if I were to have an estimate based on third party data I think it's getting pretty close to 10 billion that's a very significant amount of very high margin revenue that is growing really really
1:17:12Awesome. So as we close out I have to ask you this is my most difficult question.
1:17:18What is your hottest take right now?
1:17:21Um, man, I feel like I've had a lot of hot takes over the last week. Um, you know, I would say, um, man, I can't choose a single hottest take that. I mean, Grockbot's the most useful AI tool, you know. Um, SpaceX is going to make orbital compute, Starlink direct to sell, and asteroid mining. I say asteroid mining because that hot
1:17:51take always really upsets the gold bugs really and I enjoy the commentary for the gold bugs on X.
1:17:56Um it's going to make all of that a reality. Um I think Jensen is a singular CEO. He deserves a benefit of the doubt.
1:18:05This is America. We are innocent until presumed guilty. Everybody gets due process. He is a great American and America loves nothing more than a comeback and I'm excited to see his. So, I don't know which of the I don't know if those are hot takes or which is my hottest takes, but those would be things that come to mind.
1:18:25What is your new profile picture?
1:18:28Uh, it's me dressed up as uh Jedi for um Halloween with Becky wearing Yoda ears.
1:18:34No, it's like you in the bushes.
1:18:36No, that was the old one. There's a new one.
1:18:37There's a new one. You're behind. You're behind, Molly.
1:18:41You're way behind. [laughter] Yes.
1:18:43We'll have to check that one out.
1:18:44Yeah, we're dressed out. It's a family all with lightsabers.
1:18:47My daughter has a lightsaber chronicles for your child of the Dune.
1:18:51100%. Yes. The children's Dune. It's going to [laughter] be amazing.
1:18:56Awesome. Well, thanks so much, G. Thank you so much.
1:18:59Cool. You are welcome to Sorcery.
1:19:03So, you are a world champion poker player.
1:19:06You're also an expert in game theory and now you're a founder of Just Games.
1:19:11Just Games. That's it.
1:19:12That's fine. Just Games. [laughter] So I have to ask you what is your hottest take right now?
1:19:16Yeah. So I think the discussions around export controls have gotten a little bit convoluted and nuts. Like the current thing that is being claimed is that there is basically like a six-dimensional chess argument why it is good to potentially sell advanced chips to China. And like when you're finding yourself making a very convoluted argument for this, I think it's time to question the argument a bit more. And the argument goes something roughly like, hey, we want to lock them into our
1:19:46tech stack because then they're dependent on the US and uh then we can like use that to our advantage at some later critical point. But China is not dumb like China is not so stupid to actually be locked in one and that you have to presume for it to work. And two, China already has like is performing better than we do on energy. They are bottlenecked by chips. So they can use marginal chips uh where we are in the opposite position. So they would be very
1:20:16happy to have additional chips. You're helping them additionally and I think that uh if we then to it also add so one I think the argument was already falling apart and then two now we've seen that anthropic reported that they were able to switch to AMD chips from one tax to another where it took them 6 to 12 months before to only 3 days. I mean that was part of their as well. So who knows it might be uh a little bit more optimistically phrased than it actually was. But in any case there is definitely a fundamental truth behind it that if
1:20:46you're a hardware mode was based on software of course if you believe that models keep getting better and better that will disappear. So I think in the world where uh it will be easier for companies to switch between chips, the lock in to a tech stack becomes smaller as long as it was a software one, China will want the additional chips and China will already anyway do actions that benefit China. Like they are less of a capitalistic actor than the US is always
1:21:14happen. they're more of an ideological actor and they're happy to take a short-term like um uh cost in that regard like they did with Jackmar etc.
1:21:25Uh in what they have now done is they're not taking the H20 chips because they want Huawei chips to like be sold to Deep Seek etc. But in any case some chips are getting to China that is a problem and that needs to be attacked in some way. We've seen that uh the last Chinese models were all trained at least in part uh on black wells and where that goes and how much Nvidia knew about it is unclear but uh I think that we need to yeah keep our advantage that we
1:21:55currently have around models which is fantastic. We need to have that advantage and then we also want to keep our advantage in chips rather than sell them to China and make some like multi you know the meme with a gymnastics you run around and jump in various ways.
1:22:09Yeah. So I think that's basically the meme that was going on. And then yeah paired with that the lockin happens less and Jensen said I'm not a car. I think now you're a car.
1:22:21Yeah. Now he's a car. like people can just go in, use you, step out, use another one. Like cars are very easy to switch in between and we're moving slightly into the direction of a car. He's the best car that exists and some cars have a massive advantage. It's extremely hard to build such amazing cars as he has. Uh but uh yeah, there are companies here that are also building other great chips.
1:22:46How did we get to this moment?
1:22:50to the moment of from a chip's perspective or a the scandal and Jensen being for lack of a better word targeted in the way that here I mean I I don't have a specific inside view on it. One here is about the internal power play maybe within the admin could have led to it. Certainly Jensen had a policy position towards China chip sales that was more doubbish in terms of China chip sales for
1:23:19multiple years. I think completely independently whether Alice Huang was or wasn't a spy. I think that it's even possible that if she was that uh his desire to sell chips to China, which he spoke to to a degree, could have existed totally independently for simple shareholder increase. um benefits basically shareholder value increased benefits. I think that's very consistent thing. It's just the thing that we're
1:23:46sometimes losing sight of is we're listening to points that are being made from a desire to maximize shareholder value and take them as a kind of blue sky philosophical thinking. And this is kind of breaking the ability to make sense of who says what. I think it's very right and fair to increase shareholder value and I celebrate anybody who does that and it's very
1:24:14right and fair to create LP capital value all of that but I think uh when someone argues with motivated reasoning and pretends it's for other benefits that is when our conversation gets uh broken and uh I would love to have people that create shareholder value focus on that and those that want to see um the benefits of AI being distributed primarily to the US do that.
1:24:43How do you see the openw weight versus close competition plays out?
1:24:47Yeah, I'm not even I think that that's also a bit overframed in general. I think that you can't shut like no one's going to shut down open weight models.
1:24:56So like of course there's some push against it, but you fundamentally can't do that. Like the question is how big will the delta be fundamentally like down the line between closed weight and open weight models. And unfortunately likely it's going to grow absent uh some more funding being collected for uh an open ecosystem.
1:25:19Um because I think that over time the model companies will stop selling the tokens and offering their models and just like create um useful products and services out of them which they're already starting to do. Uh so in that world like maybe we're gonna have like more far apart close to open weight models but I don't think it's even that true that the open weight model is attacking the closed weight
1:25:49model. I think they just have fundamentally different use cases and the benefits of closed weight models will be in scientific discovery medical breakthroughs all of that stuff down the line. currently also in breaking other startups obviously uh votes and uh attempts to make money themselves.
1:26:06So as we close out I usually ask this in most of my interviews and it's more of like a personal performance question but who's one person that's really inspired you or you've learned from that's made an impact on you throughout your career?
1:26:21um two people uh and it's kind of two of the obvious names but I was fortunate to get to know them but it's uh Elon Musk and Deisabas both are like just the most extraordinary thinkers and just they speak fortunately as well often publicly right and uh just hearing how they think through problems the pure force of will and nature that Elon is in like how he pursues audacious goals and
1:26:49the ambition that he has is like massively inspiring. It can be sometimes even breaking a little bit because you're you're just not sorry maybe you're but I'm just not that I don't have the same amount of drive. It's just no it's just no one has it but it's like wow that's possible to have and you can you've motivated from both of those people to just do more be better.
1:27:10Yeah, they both do that.
1:27:12Did have you gone to a Starship launch and I just came back from one.
1:27:14Oh, you did? How was it?
1:27:16It's it's a sensory overload nearly because it's just so it's like it's mind-blowing that it's possible. The launch itself is like if you have you been?
1:27:27Okay, you have to go. It sounds like you're operating at the like limit of physics, which you actually even are, right? It's like was gravity even 10% stronger then we wouldn't have uh been able to do this.
1:27:38Um so yeah, it's it's it's very inspiring.
1:27:42Extremely. and then catching it is just the most like your mind is not [laughter] ready to see that it doesn't any time I think when people in life prior to the catches have seen a massive structure fall out of the sky at that speed it was life- ending event [laughter] and now this massive structure is getting caught delicately caught softly it's amazing it's crazy yeah I I extremely strongly recommend
1:28:12going there it's also So, like a nice warm town south of Texas. Uh or if you're in Mexico, go up north of Mexico. And uh it's open to the public. It's a great view from everywhere. It's incredible event.
1:28:28And all kids love it. It's actually a family event literally.
1:28:30I can't imagine seeing that as a kid.
1:28:33Like if you grew up seeing that, that would be that would be crazy.
1:28:37And you'd think it's normal, right? Like that's that's what we do. You're like, "Yeah, humans do this. I should be doing something else. That's also as cool.
1:28:44No, it's incredible. I can not recommend it more to everybody to go there at least once over the next few years, especially over the next few years where they still might have a failure here and there.
1:28:54Then it's because in 2 three years, it's just going to be look a successful launch and it's just like it's kind of like standard operating procedure. It's going to be as awesome but slightly less exciting.
1:29:05Damn. That's a good that's a good one. All right. Well, thank you so much.
1:29:08Thank you. And it's not because Chase got hacked. It's the third party that got hacked happened to be embedded there on the on the front end of the website. And so I think um going back to like why you have to audit the code a coding agent can tell you today, hey that's a bad choice.
1:29:24Just are you actually putting it as part of your software development pipeline?
1:29:27Are you willing to hear the the bad feedback of like hey like that's that's dumb. You should not do that. So we have the solutions we have to like as an industry broadly disseminate them. And the reality is we're dealing with massive amounts of tech debt. Our infrastructure and software systems have been built from the ground up by humans for decades. It's actually time to start rebuilding a lot of these foundations.
1:29:54We can rebuild them in memory safe programming languages. We can rebuild them in ways that are almost like provably unhackable.
1:30:01Mhm. And and I think so we're going to have some pain in the next weeks and months as we hear more about this these breaches. I think it's it's very likely they're going to happen. Um but I think on the other side as we start deploying this this cyber security AI systems we're going to end up in a very very strong position.
1:30:21I mean I think that's a good point. So do you expect more companies to build entirely new systems and then migrate over because migration doesn't take like a whole team of consultants in like multiple months anymore?
1:30:31Not anymore. Right. So obviously I'm I'm I'm CEO of our cell. So I kind of see this every day, but when when an enterprise buys our platform, we help them deploy new software that they can now produce infinitely faster because they have coding agents. And the default way that enterprises roll out Verscell is behind the firewall. So literally like the attackers can never even reach into this private networks. Now conversely the risk that exists with a
1:31:00lot of the agents is that they're going to you know throw your data out into the public internet because software like I said is now basically free. It's so easy to generate. What you need to do is make sure that it's properly governed so it doesn't accidentally, you know, kind of like uh open AI accidentally hacked hugging face. You don't want your agents to accidentally eat out your data into the internet. And so I think our solution is embrace the coding agents,
1:31:30guard rail them, deploy mostly internal software and if you do need to deploy to the public internet, get that extra set of AIS that are warning you even before the code goes out, hey, like you could be improving your security posture. So, do you think what's going on right now is less so of a tinfoil hat moment and something that is just obvious and inevitable?
1:31:56I'm I I mean, if you're if you're running critical infrastructure, your default attitude has to be like three layers of tin foil around your head like because the the gaps exist and and as you get deeper into this uh labyrinth, you realize that there's there's risk everywhere. And so the the question is how do we mitigate mitigate it as fast as possible? And the answer is we have to deploy AI. So we're all using AI in
1:32:24like creating new apps, sites, images, whatever. We now have to deploy AI towards protecting the systems. I think what we're seeing now is the models have gotten really capable at cyber security. And so the bad actors have found the meth a new methodology for hacking, but also they found a new reward for hacking.
1:32:47So the targets o of these hacks are all companies that have access to AI tokens.
1:32:54And you might have heard of this distillation campaigns and distillation attacks that are happening. So if you hold a lot of AI, you know, whether there is North Korea or what have you, they're going to try to hack you because then they can get AI which is now almost like the new currency. Tokens are very precious. They're fungeable with US dollars, right? And so there's there's black markets, not only for data and and these data bounties that exist, but also
1:33:22for tokens. Uh and so the the reward for hacking has now increased essentially.
1:33:28What do you think's going to happen with the open weight?
1:33:31I what I have seen is every company in the planet needs to use AI in order to basically think like a hacker. The best way to find vulnerabilities in your code is to audit it like a hacker would. And that's where open with models can actually be incredibly helpful because the coding agents that you use to produce the code, they miss stuff and sometimes they even have guardrails where they don't allow you to like look
1:33:59deeper into like what could go wrong.
1:34:01And so I actually like a world where all of the models are almost like pointing guns at each other. Like you almost want an agent that distrusts what the other agent produced. So if you're familiar with Verscell, anytime you you ship code, we or you you push a git commit, we build it and deploy it. And so in that pipeline, we insert an agent that uses a combination of proprietary and open models that basically takes this like very sus attitude and says like,
1:34:31yeah, I don't trust this code. Where could it go wrong? And so also when when hugging face got hacked, they actually used open models in order to make sense of the attack. They were like, okay, the the agents skip the sandbox helped me understand the the surface of attack here. And so we've seen a lot of data and anecdata of companies using open models to for defensive purposes. I think that's a really good thing.
1:34:57Should we talk about Nvidia? Are you at all like look this what I love about Nvidia is that they've always been a company that builds an ecosystem.
1:35:10They've never try to own the full PI. In fact, when they've gotten into models, they started open open sourcing models like Neotron. And so Nvidia I think is extremely spiritually aligned with the mission of Verscell which is to help customers own their intelligence and to own their agents. Of course we we collaborate with all of the labs etc.
1:35:34But I think the best outcome for the world for the free world is that we all have free access to intelligence. Free as in low cost and as in I can modify it. You know Verscell was born out of open source. So many companies run on thanks to the success of open source software. You know, famously Facebook started out by getting runtimes and frameworks and programming languages that were which were open source. And so
1:36:03the next great American companies are going to be born out of ecosystems. We need to maintain open ecosystems and um and I think Jensen was doing a phenomenal job at building an open ecosystem and I think a lot of people are are very invested in his demise because of that.
1:36:22Now obviously we can all have the moral high ground of like f Jensen. He was doing a lot of wrong things clearly but there's also I think a little bit of a fog of war. It's like why are you out to get Jensen? Is it this? Is it that? And what is your motivation? I think people need to study the the matrix a little bit more deeply.
1:36:41What should we be studying in this matrix?
1:36:44Well, I mean, what are the reasons that you would want Nvidia to fall, right? Well, one is that if you like an if you have if you're building a fully proprietary ecosystem and and Nvidia is trying to like fund the openweight alternative, I think that you have a reason to like at at the very least like be like, okay, good readance, Jensen. Sorry that happened. buy into oblivion.
1:37:06Uh obviously there's the people that are the classic people that are dogpiling after any kind of like mistake is made.
1:37:12I think he's made a lot of mistakes and like there's always a peanut gallery of people that are like ah like I'm perfect, you're not. I I everything in my personal life is 100% perfect and you are an And so I think there's an element of that. Um and then there's legitimately I mean like this is the most valuable company in the world.
1:37:31Everyone wants a piece of this. Some people are short, some people are long.
1:37:35And so I think anytime that something like this happens, it's never about just one or two stories that break out about what the CEO was doing. There's we're talking about the the most valuable company in the world. Uh and we're also talking about geopolitical uh issues, right? And so uh it's true that Nvidia was building an ecosystem that's open in in every sense of the word. They were uh uh they sell internationally. Um, you know, uh, Jensen in interviews was very clear that
1:38:05he wants a world where China depends on the US as opposed to a world in which China built something that could arguably be more formidable than the US. Maybe a a cautionary tale here is look at the success that China has had with video models. They're mogging like seedance. Even the success of Tik Tok, they're mogging.
1:38:29Yeah. And so it's good I think and healthy to be in a position where they depend on us rather than like they have firstparty everything first party data sets that could be larger than ours. If you want to create a great video model it turns out to be it's pretty helpful if you have billions of people generating video data every day in your TikTok, right? Um and so overall I think the world that Nvidia was pushing for is overall very very positive. uh and at the same time of course want America to
1:38:58to to win and and continue to be the spear of of the AI revolution. I think it's actually existential for America to be the spear as you've seen with cyber warfare. You want the best weapons and you don't want our adversaries to have the best weapons.
1:39:17Okay, I said it, but we're going to have a very hard question now. This is the final question. What is your hottest take right now?
1:39:25My hottest take perhaps is now becoming lukewarm as this conversation evolves is uh it's [laughter] there going to be a lot more high-profile hacks.
1:39:34Uh we put out a essay that I recommend everyone to read which is this hottake in essay form called everything hackable will be hacked. Everything out there that has a potential to be hacked can now actually get hacked. And so it's probably not a hot take but everyone needs to really sh up on their cyber security. And uh we're we're sharing our resources and what we're doing with the world and and we're investing in the open. We actually uh open source a cyber
1:40:01security harness called DeepSack. DeepSec allows you to like spray RAID and kill [laughter] all of the like bad stuff on your infrastructure.
1:40:13Uh but there's going to be some pro there's likely going to be some short-term pain. And I can't tell you of what kind, but remember when like I was at an airport one day and like uh Crowd Strike had that famous like uh big downtime and everything was a blue screen.
1:40:28You know, there's probably a scenario where like stuff like that happens sadly where you know something goes wrong and a lot of critical systems go down and and we have to like really be tinfoil hatw wearing and and and fight hard and invest to to mitigate those risks.
1:40:47All right, G, thank you so much.
1:40:50It's great to see you.
1:40:51Hopefully everything goes well in the world.
1:40:54Vib, welcome back to Sorcery.
1:40:56Thanks for having me.
1:40:57Well, you're fresh off the stage. What just happened?
1:40:59Uh Joe, uh Joe Lonzel and I did a fireside chat. Joe's one of our earliest investors. Um I actually right out of college, I actually worked with Joe at HBC. Um so I've known him for a long time. So it was kind of full circle moment to do.
1:41:11What did you guys chat about? Any hot topics? Um, and really just like kind of current state of affairs with Seronic.
1:41:17Um, some of the stuff with our we had the news in June on the the the rescue of the two pilots. Talked a little bit about where some of the boats are deployed and kind of what the mission sets, not like actually, but like what they theoretically look like um and some of the challenges there. Um, we announced uh in the la in the last two months, we announced a new shipyard down in Brownsville, Port Alpha. Um, JD Vance was was down there. We did like the the the shovel [laughter] thing. It was cool. People just clowned me on Twitter because they said I wasn't holding the shovel properly. really. There's always a hater for something.
1:41:46Got the shovel haters.
1:41:48I know. Just experts. Um and then uh yeah, we talked a little bit about like kind of what uh why why we're able to move quickly especially post 2022 after we started the company. It was like all of the AI benefits um were there from the very beginning rather than like having to really pivot around okay how am I going to save money, save cost, go faster with AI. It's like kind of rebuilt in the DNA because we started the company right around the time chat GPD came out first came out. So boats chats same thing same thing
1:42:18um I think it would be good to kind of set the stage it's always a good reminder of like how far behind America is in ship building capacity and why it is so important that we open new ports and new ship building facilities.
1:42:32Yeah. Um I I'll give you actually a very relevant example. Just I think on the 13th or the 14th um uh uh Trump actually uh extended the Jones Act waiver. So the Jones Act is a 100-y old law that basically says um any ship traffic from an American port to American port has to be via an American crew and an Americanmade ship and American flagship.
1:42:53Um, and the what what the downside of that is that what actually happens very often is that you end up having to like the cheapest way to do this if especially if there's not enough capacity is you use a you know European flag ship or or whatever likely built in China or South Korea or Japan um but most likely China. China's 50% of global capacity and then you go from like LA to South America, South America to Seattle or you do u Savannah to Africa, Africa to Boston or some something weird like
1:43:22that just because uh that's just a you have capacity always and then it's cheaper. Um and so what we saw with the waiver is that especially when you first passed the waiver, I think the first 40 days of that waiver passing, we moved more um more like LG, diesel goods, etc.
1:43:38like petroleum products from the Gulf to like the Gulf of Texas, Gulf of America, Gulf of America to uh to California, like that whole route. Uh we we we moved more petroleum products in the first 40 days than we did all of last year.
1:43:51So there's latent demand and capacity needs just to move things around.
1:43:55Um and uh yeah, since since um since World War II really, we just have had zero commercial ship building base. It just it just doesn't exist. I think this federal subsidy stopped in the early 80s and since then it's basically gone to almost to zero and ceronics fleet has now expanded out to um can't say we build hundreds of Corsair. We hundreds of Corsair [laughter] deployed hundreds of Corsaires. Um we have uh Marauder 2 just came off the line. We we just posted about that. Um
1:44:25did you like whack the champagne bottle?
1:44:27Champagne bottle. I mean I didn't um but I'm sure I'm sure someone did.
1:44:31Okay. Um the uh and then we just announced on Thursday uh our ALPV product. It's a new contract that we just we just for the for the Marine Corps that we just won. Um which is basically like a 70 foot really long skinny kind of looks like an ARC boat a little bit like the um the most of the boat is actually like right there close to the water line. Uh just announced that boat on Thursday and then obviously Mirage and and uh Corsair are in are in production. So as CTO what are the most critical components or aspects of the boat?
1:44:59obviously really difficult to build, but like what are you most challenged with right now?
1:45:04Yeah, I mean I think generally speaking like in this space um your the reliability question mark is the most important thing. Um it's very easy to go to a demonstration have one boat goes and drives around in a circle comes back it's like oh great. The challenge is that what happens when that boat needs to sit out there for 5 days 10 days 15 days for the larger ships 35 45 55 days and you have nobody on board. If a cable comes loose, if an antenna falls off, these things are common in the ocean
1:45:32because, you know, you're taking like especially when the CCS gets tumultuous, you're taking slamming loads that are like 10, 15, 20 g.
1:45:39Have you seen the waves outside of San Francisco?
1:45:42Oh, it's I was watching like buoy readings and it was three stories tall out in the water.
1:45:48I mean, when you think about that and the implications that it has on the whole often times like a man ship like will they will avoid storms or they won't even go out when it's like that, right? for us, we don't have the whole point is that you don't [laughter] you don't have to. And so the the implications there from a reliability perspective are pretty challenging. So that's like everything from propulsion, the engine, uh sensor quality, right?
1:46:09Like you know a bird can come sit on the boat for a while. I think we talked about that last.
1:46:12I was like maybe some laser.
1:46:15No. Um but uh saltwater just degrades things over time. And so you know the maintenance, the repair, those are like big drivers of the total cost of ownership of ships and and and boats. Um but uh those are all things that we need to think about as an N10 system. So from the design from on the on the on the naval architecture side all the way to the component selection, what do we choose to vertically integrate? What do we buy off the shelf? Um to verification, validation, testing to actually like every boat that comes off the line, it's very difficult from a manufacturing perspective to make it
1:46:44look identical, right? So how do you know that like this one has uh you know the strakes were slightly placed in a different different position? What are the implications of that? All of those things need to be qualified. um understood really well to basically make a make a decision on like hey can this book go out today is going to complete the mission those are those are the really big challenges I was going to tell you this but we have a really hard question as our closing question what is your hottest take right now okay so different than the last one I [laughter] think I think that um you
1:47:13know Nvidia and Google I think and a couple maybe a couple other companies I think there's startups in space too but they open sourced some weather prediction models um and uh historically like meteorologists and like just humanity in general has really bad at predicting the weather.
1:47:26Um we see this when like storms come or hurricanes or things like that. Um we have some in some areas where we're quite accurate but when you actually increase that horizon we just become really bad after like 9 10 11 days like it just gets the prediction accuracy falls off the charts. Um I think that we need to invest more in better weather prediction. Like imagine the future where like you have a like 30 days you can predict weather perfectly. Like that would be amazing. And I think there would be economic implications to that.
1:47:53Like I'm sure the hedge funds would love it. Um but I think the AI models need to get better at weather prediction or the systems in general. Um and I think a lot of the data that's actually missing from that is like from the ocean. Um so hopefully I can marry my two hobbies um autonomous boats and weather prediction. Um and have a fun little side project.
1:48:10What do you think of El Nino?
1:48:13What am I supposed to think about?
1:48:14El Nino changes the weather.
1:48:16It's true. It happens. Yeah.
1:48:18So what's going to happen? I we we're so bad at printing weather I can't tell you. [laughter] Okay. [gasps] All right. Well, Vib, thank you so much. This is awesome.
1:48:29Also, great hot take. I love weather.
1:48:32I love the weather, too.
1:48:33Yeah. It's kind of hot in awesome, though. But I know I'm like I fly a lot, so I'm constantly looking the hourly weather.
1:48:41I grew up in the Bay Area. Like, you growing up in the Bay Area, you never check the weather app.
1:48:46And then you don't need to.
1:48:47Anywhere else you move, you have to check the weather app. So, same with LA. You don't have to check in.
1:48:52Yeah, but at least it's kind of hot though. I used to live in Orange County.
1:48:54It's not that hot.
1:48:55It's not as hot as Austin. That's true.
1:48:57All right. Side tangent. Okay. Thanks, babe.
1:49:00Cool. Ron, CEO of Eye Connections.
1:49:02Welcome to Sorcery.
1:49:04Hey, and welcome to Tomorrow X. Thank you for being here, Molly.
1:49:07This was like the biggest conference that ever developed in such a short time period. So, how did you do this? How did you pull this off? It was a little fast, but there's a very good explanation for this and it's at the core of what we're here to talk about today, which is AI.
1:49:23Our team used AI literally from the second we had this idea, which was around March, April, we secured the Moody Center. And you can't pull an event. This is 5,000 people in a basketball arena. No one does that in six or seven months. That's like an 18month kind of thing. But using AI, we were able to build everything we needed.
1:49:45so quickly. The website, the dashboards, the mobile app, all of the elements that go into this that, you know, you typically took months and months to build, we built them in days, and it really positioned us to then focus on recruiting speakers, marketing the event, you know, adding partners.
1:50:03If you embrace these AI tools, and I have to give Crockbot a big shout out because Grockbot is what made this whole event possible. My team set up Grockbot across the board. My COO, who normally isn't doing like in the weeds kind of stuff, had a dozen Grockot a Grockbot agents and was running literally half of the leadup to this event on her own.
1:50:30None of that was possible before these tools. So, here at this conference, we're talking about AI. And another element of it that's really important to us is how AI can improve the lives of everyday Americans. If you master these tools, I just need people to know you are valuable in the marketplace because you are so much more leverable than you could possibly be.
1:50:54Even if you're brilliant and great at what you do, this is just giving you three, four, 5x leverage on whatever your skill set is. So that's really what we're here to talk about. It's the creation of these tools. It's the infrastructure and the ecosystem that makes up this whole industry. Uh, and we want the average American to hear the positive side of the AI story. I mean, sadly, this story has had such a negative slant to it in my opinion. And we really wanted to give the industry a
1:51:23chance to tell the positive side of the story. So, what is really fulfilling for me is not only did we throw this amazing event and put it together super fast, we ate our own dog food, if you will. We use the tools that we're here to talk about to make this whole thing happen.
1:51:40All right, I'm going to ask it. Are you ready?
1:51:44What is your hottest take right now? My hottest take is this this event shows that as much as AI can do all of these amazing things, one of the things we as humans value most is meeting with each other in person. There are meetings happening. There's over a thousand meetings taking place between companies and institutional investors at this event. We have no idea how many meetings are taking place among all of the other attendees who are in all the other parts
1:52:14of this building. But that is such an important element to us as humans to be able to connect with each other, to meet each other, build relationships. These are the things that help you build your career, build your companies. So it's, you know, what we do at I Connections is run in essence the best events in the alternative investment world. And it just proves itself over and over again to me. If you can convene people and put
1:52:42them together in a room, they get more value out of that and those relationships that they can build in those one or two days that they're together than anything that they're going to do through Zoom calls, anything they're going to do online. We are still people who like to press the flesh and meet each other in person. And that's how we build trust and relationships and things really ultimately move the world forward.
1:53:05I think picking Austin was also an exceptional choice because it's a very underrated city. There's so many phenomenal founders here and investors and people that you might not normally run into because they live here and they don't usually travel to the other places. But I think this was a really good testament to exactly what you're saying. Like last night there was an event. It was like a dinner party and um you always know when the party's good when they flicker the lights on and off and they do it for like 20 minutes and
1:53:34no one leaves and then after that they all end up downstairs at the other bar and then they stay until midnight. Like that's when you know you've created a really good community of people coming to an event and curating really exceptional quality. So I hand it to you for sure. Austin is absolutely killing it. I mean, Austin, I'm a kid of the dot boom, and I remember in the '9s, everyone was headed to Silicon Valley.
1:53:59It's like if you wanted to build a career in technology, you just had to be in Silicon Valley. If you wanted to do it fast and you wanted to do it right, that is now Austin in my opinion. It is absolutely the hottest tech city in the country. There are so many amazing founders here and really from like all ranges. You've got the billionaires who've moved here, who already have made it and built great firms, and then you have a ton of startups who want to be here to be part of this ecosystem. So, picking Austin was really, really
1:54:27intentional for that reason. We knew if we could do the event here, it would make it so much easier to build that community and pull them all into this building because it was a local drive for them. But it also then drew in the best from New York, the best from the Bay Area, the best from LA. So, we we are thrilled with the outcome of this and we will absolutely be back here at the Moody Center next year.
1:54:48Well, incredible job. Congrats to you, Gavin and Antonio. This is amazing.
1:54:53Yes, I have the best partners. These guys are just such rock stars. All of these amazing CEOs are here because of those two. They have incredible reputations. They're incredible humans, and that's why we had this incredible lineup of speakers. So, I'm very lucky to have them as my partners.
1:55:11Congrats. Thank you so much, Ron.
1:55:14Harper Carol, welcome to Sorcery.
1:55:16Hey, Molly. This is the first time we're recording and I know I've asked you a couple times like maybe we'll do a walk and talk around Stamford since you are a professor.
1:55:24I wouldn't I wouldn't say I'm a professor. I do like to guest lecture there regularly, but I wouldn't say not going to claim that title. So, yes, but yes, we've we've seen each other a lot and for somehow somehow never have done anything together. I know. At least not it's changing today.
1:55:42Yes. So, for the ex audience, Harper has made her debut and so you should definitely follow her if you haven't cuz the content game is picking up. Oh, thank you. I appreciate it. I've been So, I actually started on X with fine-tuning guides. So, if you don't know who I am, I'm a computer scientist specializing in AI with my background at Stanford and Meta and then as founding engineer and then head of AI at brev.dev, which was acquired by Nvidia
1:56:12And when I was there at BEV, I started making these fine-tuning guides. We're like, how do we get people onto Brev?
1:56:18Like, why would anyone actually want to log in? What it was is a remote cloud development environments for GPUs. So basically like accessing cloud compute from AWS, Lambda, GCP, pulling all of that and making it really easy to log in with one click and like setting up Jupyter notebooks and CUDA and everything for you which like if you were an engineer back then it was kind of a nightmare. But we were like okay there's things like replicate you know API calls where you can use open source models just with an API call like why
1:56:46would you actually want to log in to a remote machine? And so we realized like, oh, to fine-tune open source models. But at the time, no one, this was 2023, no one really knew how to do that. And so I made these guides. I made one guide. I had 50 followers on X. [laughter] I made one guide. I just started my X account.
1:57:03I had no Instagram. And I made this guide on how to fine-tune withdrawal on my on your own data. And I did it on my journal entries, which was like an added like twist, you know. Um, and yeah, I released that and it got 50,000 views because I think everyone wanted to know how to fine-tune these guides. And we got this like influx of users onto Grev.
1:57:23And so I've had like hundreds of people come up to me at conferences saying, "Oh my gosh, your guide is the reason why, you know, I we have a fine-tuned model at my company. Our company runs off of your guides basically." Um, and so yeah, they were the number one most viewed guides on Google in 2023 and kept like releasing open source content, coding guides. now like you don't really need those anymore but they they were just ma basically Jupyter notebook guides and video YouTube videos things like that
1:57:50and on X and then in January of 2024 I decided to make an Instagram and just talk about like general AI concepts and that's what I've been doing the rest is history super mega star oh it's I'm a huge mega star everyone knows me um no I'm exaggerating no but but it's been really fun. I like I love it. It's I have always loved teaching like my whole life. Ever since I was a
1:58:18kid, I've always loved performing and you know acting and just like everything kind of coming together.
1:58:27So yeah and I think I think my skill is explaining complicated things in a way that people get like you know I studied AI Stanford for eight years and for my undergrad my masters while I was working on machine learning systems at meta on integrity intelligence and newsfeed ranking and it took me like years to really put it together in my mind you know it was very hard for a long time and I feel like I feel like now I like get it And I have
1:58:57a really strong understanding of of the landscape and everything and like how everything fits together. And I have a really strong intuition for it. And I I try to translate that in a way that everyone can understand like any kind of concept about AI. Um and so yeah, I've gotten really good feedback on it and that's like my specialty is like teaching how AI works under the hood basically.
1:59:18And so what are the different formats that you use to teach and where the places people can follow you?
1:59:22Yeah. Um mainly I'm on Instagram right now and X I crossost to X. Um YouTube I'm starting to pick up again. I used to do these this like long form series on the t days of AI basics where I went into depth on like you know AI versus ML was day one and then data day two. Day three is model architecture talking about neural networks and like what the architectures look like and what the different model layers look like and how they work. Um and like those were about 40 minutes. They actually still
1:59:52I would say almost everything in those videos still applies to today which is cool like when you learn about how AI works under the hood. We haven't had any massive breakthroughs so it's pretty evergreen content and I think it's something that a lot of people still really gra like struggle to grasp. So you can check those out. I'm coming out with like a series on actually how AI works under the hood. Oh my god. Is this an exclusive?
2:00:14Oh my god. Actually I have a girl.
2:00:17This is crazy. Breaking news. [laughter] Yes. Yes. I'm coming out with a series on it and I'm excited about it. I've It's been like the number one I have a free guide. I have a guide going through all my videos on on YouTube and Instagram in a cohesive like guide cuz I have so much like really good educational content.
2:00:34You need to make your own like AI bot like should be like Harper GPT where you could like talk to it and ask it questions. I think that's super helpful. I would totally do that and I think the way that it explains it should come exactly from the way that I do it. I think that's a great idea and like having it lectured to you. That's a great idea.
2:00:50Yeah, you should just go build that right now.
2:00:53Okay. Yeah, [laughter] just like BRB like just going to do that. We can really sit alongside the this the classes that I'm coming out with. But yeah, basically it's like a a like trove of how AI works under the hood, how AI models, large language models, transformers, everything about postraining, like what kinds of post- training happens. um open source models, closed source models, getting set up with open source models, fine-tuning them, all that kind of thing. So, so how can brands and companies I know
2:01:21you do some work with brands and companies. How can they work with you?
2:01:25Yeah. Um you can reach out to my team. I have an email on my page, so just check that out. Um but yeah, I've worked with like kind of the all the big ones like OpenAI, Anthropic. I've seen your socials. You're at every big release.
2:01:40Thank you. Thanks. [laughter] I saw you at your biohub. Yeah, that was that was fun. We had a good time at Biohub. Um yeah, I I mean it's it's it's fun to see you everywhere.
2:01:50Like I love that that we get to hang at all the all the things. Um although you get invited to a lot of things that I'm not at. So I see you at some of my things. You see me at some of yours. But yeah, I I yeah, you can reach out and and I basically like explain I'll promote I'll evaluate what you're doing, see if I like it, use the tech, and then kind of explain either how it works in a way that is like wow, why is this information so important? It's like how
2:02:19can people understand it? Why does it matter? Like the biohub thing, you know, that's like very technical scientific content. like most people like most people have not been exposed to that kind of thing. And so like being able to kind of explain it like learn about it in real time and then kind of explain it in a way that's digestible for people like kind of translate it is kind of what I do. So if you've got anything like that, she's coming. I'm coming. All right. So
2:02:47as we close out, this is my difficult question. I know you're so nervous for this. You couldn't sleep. [laughter] I actually did not sleep very well last night, so please excuse excuse me and excuse what's about to come out. So, yes.
2:02:59What is your hottest take right now?
2:03:02Okay, I've like debated this. I feel like a lot of them are kind of obvious, like you know, everyone's going to be a solopreneur, founder, like not everyone, but most people. Um, things like that.
2:03:13And I think maybe one of the ones that are less touched on right now is more of a hot take is that um I think the way that we consume media is going to change drastically. And I thought about this like six months ago, eight months ago when a friend of mine was like, "Oh, I'm building this kind of central interface for AI." And one of the things it does is it pulls from social media and gives you like a curated list of your like five of your friends posts and like five social media posts that you would like like to learn from that you're interested in based on like what you actually want to learn from. And so it's
2:03:42filtering through a lot of the slop, a lot of the like brain I feel like we all know that these algorithms have been optimized to suck us in and to like control our brains in a way. And it's not necessarily intentional just for to optimize for usage. And sometimes that's not it's not it's not really aligned with how we want to be in the world. And so I think with agents like it'll serve as an interface in between us and those feeds. And we actually saw this recently then I saw this recently and was reminded of it a few weeks ago with Andre Karpathy when he came out with how
2:04:11he had made his own like hacker news YouTube and he went on YouTube he saw it was a bunch of recommendations he didn't want he didn't like that it was like slop and so he made his own and so that's like kind of like you could also like you don't even have to have your own interface it could just ping you but like just having an AI in between the content like the feed I think will become less valuable and we'll and and then in the second effects of that are like people will start start to create content that is less like oh capture
2:04:40your attention oh like what are the million things you can do to capture people's attention you know that are like brain rotty to actually creating valuable content because that's what's going to be recommended you know so damn hopefully that comes true hotter take we'll just start reading books and like newspapers again you know I'm like about to throw my phone out the window and all these cyber attacks I'm like, I don't need to know this much.
2:05:07Yeah. I mean, honestly, it's funny. I was just saying this. I have started reading fiction books. Like, I never did that. I was always like a computer head. I never really cared to read fiction books. And now I'm like reading fiction books. So, I don't know what it is.
2:05:20Like blood. Give us some Rex.
2:05:21Oh my gosh. What did I just read? I read uh Dark [music] Shore. Dark Windy Shore.
2:05:26I always forget the names of these books. Oh, I read Trad recently. like these are like just really like fiction fun books [music] like instead of watching a movie you know um gosh I forget the it's on Zero Dark Shore it's like something um but I [music] just finished Tradife and I'm like interested in those kinds of like not trashy but like engaging and fun and kinds of kinds of books a good relief.
2:05:50Yeah, it's Yeah, exactly. [music] But it's been fun. Do you read any any books?
2:05:55I carry around like 400 500 page books and I read 20 pages and then I move to the next one. Yeah. Right now I'm not reading. I'm carrying with me because this is what I do. I carry the book around with me and I don't read it.
2:06:11But it's like the 200 million years of uh like female bodies, which is really cool cuz like you like learn about evolution over time, like from the [music] tiniest specimen to like today. And there's different chapters on many things including like clinical studies.
2:06:30There's like none that have ever been on women because you can't. Our bodies just change every month. Like there's too many variables because you can't Oh wow. You can't. I mean it's TOO BIG OF A PROBLEM.
2:06:40I I'M part of the problem. [laughter] Our bodies change every month.
2:06:47I mean mapping out the entire world for Google Maps like that's fine, you know, but like a woman's body changing every month, you know.
2:06:53No, that's a little too [laughter] much.
2:06:56I I got to get past it. I know. Okay. Well, anyways, as we close out, I'll also ask you who's the one person who's like really impacted you in your career over time? I know you've like studied under many different people. It could be like anybody.
2:07:12Michael Cookanderfer. Okay.
2:07:14Professor Michael Cookanderfer.
2:07:16He was at Stanford. [laughter] Yeah, that's his name is Professor Michael Cookanderfer from Stanford. He's amazing. He like you know when I started in AI at Stanford I I didn't see a lot of people who looked like me doing machine learning in computer science at Stanford and a lot of people doubted me and that was hard and I it was I knew that I wanted to do it and that I I could that I thought like that and that my brain was like that. my brain is worked in the same way as these people
2:07:43in a lot of the same ways. But I it felt weird cuz I didn't see anyone who looked like me and he was one of the first people to like he like saw my test score. I met with him for something.
2:07:54[music] It was I I emailed him on the day of an exam. We had a a test due that day and or a project due that [music] day and I found out I was receiving a surprise birthday that night. I was such a procrastinator. I was gonna finish it that night and then I was like, "Oh, no. I can't because I have the surprise party." And I emailed him. I was honest.
2:08:14I said, "I have the surprise party. Is there any chance I can get an extension?" And he said, "Yes, but please come see me in my office for time management skills." [laughter] Oh my gosh.
2:08:24And I was like, "Savage, but like honestly understandable." And so then I met with him and um and then he's like, "Oh, so like how'd you do on the test?"
2:08:31And you know, and I told him and he's like, "That was really that's really good." I was like, "Really?" He was like, "Yeah." and we like talked more about things and and I realized that I was like better than I thought.
2:08:42And that I like really did hold a candle to these other people. And so yeah, he's just been a mentor to me. I TA some he teaches like a bunch of core AI AI PhD classes at Stanford. I TA those and um he's a mentor to me even to this day.
2:08:58You know, when I post something on social media and I get dogpiled and a bunch of people tell me I'm wrong and that I'm incorrect and I call him and he's like, "No, that's literally right." like you are right and it's just nice to have like a mentor like that um who I can really rely on.
2:09:12That's amazing. Well, Harper, you are a genius.
2:09:15Okay. They um genius over [laughter] here. Thank you.
2:09:19And if you haven't followed her yet, follow her across all socials so you can also become a genius.
2:09:24Um thanks for the plug, babe.
2:09:25Of course. Well, thank you so much.
2:09:28Thanks, Molly. Wow.