Pod Insights Summation with Auren Hoffman Dave Morin, Offline Ventures - how venture studios work About this episode Dave Morin is co-founder and managing partner of Offline Ventures, a venture fund and studio backed by Apple as anchor investor. He was an early employee at Facebook, then founded the mobile social network Path. In this episode of World of DaaS, Dave and Auren discuss: Why Meta's AI ads work while Netflix, Amazon, and Spotify fail at recommendationsHow Offline runs its venture fund/studio hybridParasocial relationships and the loneliness crisisDepression as the #1 disease by 2030 and breakthr...
Transcript 11,812 words · 2 speakers
Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:02 Auren Hoffman? Hello, fellow data nerds. Welcome to World of DAS. I'm your host, Orin Hoffman, CEO of Incubate and GP of Flex Capital. Discover more episodes, get weekly data as a service news, original content, articles on data and more at worldofdas.com. That's world of DAS, D-A-A-S, dot com. Hello, fellow data nerds. My guest today is Dave Moore. Dave is the co-founder and managing partner of Offline Ventures, a venture fund backed by Apple as the anchor investor. He was also an early employee in Facebook, and then he founded the mobile social network Path. He also co-founded Slow Ventures. Dave, welcome to World of DAS.
0:43 Speaker 01 Thanks, Orin. It's so good to be here. Always love talking to you.
0:47 Auren Hoffman? By the way, been a long time coming. You and I have been longtime friends, and I have so many questions for you that I'm like, I love the fact that I get to ask them here. You kind of helped build the Facebook platform, Connect, in the mid 2000s, which is like 20 years ago now, which is kind of crazy.
1:02 Speaker 01 So crazy.
1:04 Auren Hoffman? Do you think there was ever like a realistic off-ramp from advertising as the dominant business model for these social networks?
1:11 Speaker 01 That's a very interesting question. I actually was just debating this with my co-founder from Path this morning. Watching ChatGPT and OpenAI launch basically almost the identical thing that we launched for Facebook platform yesterday. It's like, oh, the internet just repeats itself over and over again.
1:29 Auren Hoffman? Totally. It just keeps rhyming. And everything, by the way, keeps becoming an ad tech company. Like, it seems like even Uber is an ad tech company nowadays.
1:37 Speaker 01 I mean, it's a huge business. You look at Facebook today, Meta today, it's a $1.7 trillion company as of this morning. They have 75,000 employees. It's an enormous business. It's almost entirely an advertising business. There's no other real driver there. And at the end of the day, a lot of these consumer internet companies are content companies. They create a marketplace that incentivizes users to create content. And then, you know, it's always been like 70, 30, 30% of the users create the content that the other 70 consume. And so you just kind of are in that incentive loop forever.
2:17 Speaker 01 And it's very hard to monetize otherwise. I think that over time, you know, I saw also last week that the EU is starting to, I think there's actually regulatory change over there that's requiring that Facebook offer a subscription version of the product without ads. And I think they pegged the price at $2 or it's like $2.50. I think per month is the number.
2:44 Auren Hoffman? Or 30 bucks a year.
2:45 Speaker 01 Yeah. And so like, that's what you're worth.
2:47 Auren Hoffman? Yeah, I pay for YouTube premium or whatever it is where I don't see the ads. And that's great because for me, the ads in a YouTube video, like interstitial ad in the middle of a YouTube video is super annoying. I want to watch the video, but the ads that I see on Facebook or Instagram, at least for me personally, are usually really great content that I want to see. So I actually really like the ads and they do it in a way that it doesn't really mess up the experience.
3:15 Speaker 01 That was always the dream. I mean, even back in the 2000s, when I was working there, that was always the dream. Like, Google had this incredible cash machine where people literally just told them what ad to show. And so the thing at Facebook was always, can we infer what you're most interested in based on the content that you're looking at and the people that you're connected to, what to show you? And for a long time, it was really difficult to do that by just looking at social graphs and data graphs until, I guess it was about five or six years ago, that it looks like what Meta decided was, I think it was largely because Apple made it very difficult for
3:56 Speaker 01 them to discern who users were based on the cookie thing on mobile. They just decided to point GPUs at it. And it turns out GPUs are like incredibly good at predicting the next piece of content you should look at.
4:12 Auren Hoffman? And they're really the only one that have done a good job of that. When I go to Netflix, I don't see good recommendations at all. When I go to Amazon, Amazon has history for me for basically 25 years or more of tons of purchases. I don't see any good recommendations at all of things I want to buy. When I go to Spotify, I don't even see good music suggestions of things I want to listen to. So those three companies which have incredibly talented engineering teams, which have tons of data, all are terrible at recommendations.
4:43 Speaker 01 I think that they didn't have the forcing function that Meta had.
4:48 Auren Hoffman? They don't really make money from their recommendations.
4:50 Speaker 01 Yeah, and Meta, I think it was like they lost 30% or more overnight when this Apple change happened. They were in a really intense business predicament. Necessity is the mother of all invention. And so they just said like, well, why don't we take all these GPUs that we were buying to do other things and point them at ads? And see if we can actually do better at recommending content and ads to users that way. And we're seeing it work with TikTok, so why wouldn't it work for us? And it works and it turns out it worked wildly. And you could argue that they're actually the greatest beneficiary of AI technology right now. They're the only ones of the business that AI is like just crushing.
5:35 Speaker 01 And so I don't know, that's a long and meandering answer to your question, but I don't think it was avoidable.
5:41 Auren Hoffman? I still don't understand why like the ads, except for Facebook, I would say ads everywhere are terrible. And they don't show me what I want to see. In some ways, like we pay to see ads, like you might pay the interior designer of your house to show you ads of things you did buy. We pay a real estate agent to show you ads of homes for you to take a look at. Like we pay people all the time for ads, we're actually often paying for stuff like this. If someone shows you stuff you want, that's like an amazing service. I would actually pay for that service in a weird way, right? Thinking of like a double sided marketplace.
6:17 Speaker 01 It's crazy, but it's like heuristics, algorithms. Prior to GPU based thinking just weren't good enough, right? It was just like building these huge log of trees. It's kind of the same thing that happened with Tesla where Tesla, I can't remember what the actual number is, but the autopilot code had something like 3 million lines of C code up until version 14, which I think it's been less than one year where version 14, which is fully neural network and GPU based finally shipped to the car. And the prior version was like, you know, I've driven Teslas now. I don't know since the beginning of Tesla, however long that's been. And it was kind of like, ah, I'm not sure if I trust it.
7:01 Speaker 01 It feels strange like, and now like I got home from the airport on Saturday and I put it in autopilot at SFO.
7:11 Auren Hoffman? So you went from the San Francisco airport to Marin County. It's a pretty long drive through a city through all on autopilot.
7:18 Speaker 01 Didn't touch it.
7:19 Auren Hoffman? Fantastic. That's amazing.
7:20 Speaker 01 Across the Golden Gate through toll booths, out of a parking garage.
7:24 Auren Hoffman? Oh, out of the parking garage. You did it in the parking garage?
7:27 Speaker 01 Yeah.
7:28 Auren Hoffman? Oh, that's amazing. Okay. I, I, I do it on my Tesla all the time, but I, I don't trust it in the parking garage. I barely even trust myself in the parking garage.
7:35 Speaker 01 To me, this is the example. This is the difference between Meta's ads and the rest of the internet. They're like ads on autopilot and everybody else is still writing all of this code to try to predict what you might like. And instead Meta is just like running a neural net and guessing, and they're really, really good at it. And it turns out GPUs are great at this, just like they're great at driving a car and probably going to be great at many other things. We haven't even tried yet. I think that's largely the answer to your question.
8:02 Auren Hoffman? To me, I want somebody to send me serendipitous content that I want to see. That could be something to read. It could be a tweet or a YouTube video to look at. It could be something to buy. I want them to send something like actually that would really benefit for me. Now, this is obviously a really hard problem. I'm not something like they're, you know, they're getting the most bids or the most amount of money to send me. But the thing actually that I would benefit from, that is a service I would just freaking love. And for whatever reason, against outside of Meta, no one's really pulled that off.
8:37 Speaker 01 I think that what you've identified is the edge of AI right now. Like, the reality is, well, or the non-reality is that AI just has no understanding at all of the real world. It's very good at taking the text that you've loaded into it and then zipped up into a gigantic zip file and then put a chat interface on the front of it at predicting what the next word is or the next paragraph or the next whatever, the next pixel. But it doesn't actually have an understanding of the real world at all. Which is, by the way, why all of the videos look so strange. The thing doesn't understand gravity. Predicting what you might like requires an understanding of you,
9:23 Speaker 01 of like the internal workings of RNN that these things just don't have. Now, there was a great podcast two weeks ago by Dwar Keshe that he talks to one of the godfathers of machine learning. And this was this whole articulation of the problem right now is that everybody thinks that LLMs are the like panacea. And they're like the building block that we're going to build all the rest of the things on top of. But all of those sort of godfathers of ML are saying like, no, we need actually a totally different type of system that actually understands and can predict a world. It's like not trying to predict a word, it's trying to predict what the world is going to be like, that it's operating in next.
10:07 Auren Hoffman? I mean, obviously, there's like steps we can take. And even when I go into ChatGPT today, ChatGPT, I told ChatGPT to remember all my dietary preferences and all my family's dietary preferences. So when I say, hey, I'm in this neighborhood, pick a restaurant for me. It will say, based on your dietary preferences, think this might be better for your family or this one may be one you like, or your kids don't like spicy foods, you don't think of there. But when you're by yourself or with your wife, you'll go to the spicy food place. So it does an okay job on some of those things. It's not great. It's way better today than it was a year ago.
10:41 Speaker 01 Totally, but it's still predicting off based what it knows and what you've told it. So to achieve your grand goal of show me something truly useful, truly new, something that I can't even predict myself that I would like. We're still not there and almost nobody's doing it.
10:56 Auren Hoffman? I'm not even asking for that. I just want to be like, I just want to get it pushed. Hey, there's this new restaurant that recently opened that's within a few miles of your house. And I'm pretty confident you're gonna like this. And I saw your calendar and I saw that you and your wife didn't have a date night in the last few weeks. So I'm gonna book a night for you and your wife to go on Tuesday to this new restaurant together. Is that okay with you? That's not crazy. Like, that's not out of the realm to do, right? That would be like so cool.
11:29 Speaker 01 Yeah, well, I think we're probably pretty close to that, actually. I've seen some pitches for things like in this zone. They seem close enough. It's like two years ago, everybody was using these photo apps that made you look like a character of Game of Thrones or something. And today it's like, whoa, there's like video versions of this and it looks pretty good. Like, the difference between then and two years ago is shocking. It strikes me that the use case you just said is probably like pretty close.
11:56 Auren Hoffman? Okay, so this is what's happening. I don't know if you're having the same thing. So obviously, the capabilities of AI is getting better and better and better over time. But at the same time, my expectation of what the capability should be is growing even faster. And so there's this gap of what I expect AI to do and what it can do. And it keeps getting wider because I keep trying something out and then I keep getting frustrated that it's not working. Even though what I tried out maybe a year ago was working.
12:25 Speaker 01 I think maybe one of the things, the practices that I do to inoculate myself from that is that I try to stay really focused on what I'm doing with things right now. Like, right now, today, like, what am I using every single day? And it's a practice that I do because I'm a seed investor. Because I try to stay really grounded in the reality. The other thing that I'm doing a lot of is just vibe coding, like every day, to try to understand, like, what's possible, how crazy can you get?
12:56 Auren Hoffman? What tools are you using for vibe coding?
12:59 Speaker 01 I really like Claude Code and I've been messing around with Gemini a little bit and Codex a little bit. I try all of the command line tools because they are quite good at managing their workflows. And I'm actually surprised that I like it. I was never a terminal-based programmer. Like, I've always needed visual tools. But I think that they're good enough at structuring how they work that you can kind of talk to them like you're talking to an engineer. Like, I have easily the 10,000 hours of talking to engineers. The skillset that I've found most transferable from my life in Silicon Valley is that of talking to engineers and designers. Like, I've spent, I don't know, I could go back and catalog it, but I've definitely spent
13:43 Speaker 01 more than 10,000 hours talking to engineers and designers. And so that skillset turns out to be incredibly transferable to these command line interface code tools like Claude Code.
13:55 Auren Hoffman? To me, it seems like the main beneficiary of all these tools is the product manager. Like, obviously, everyone benefits, but the product manager is the one who can benefit the fastest.
14:05 Speaker 01 Absolutely. And I think engineers, you know, I was on a, I'm working on a prototype in our incubation studio at Offline right now. And I was on a call this morning and the lead server engineer working on the project, he just described how, the thing that he made was so shockingly awesome to me, my jaw was open on the floor. I'm like, this is so awesome. And he said, "It's so fun being a developer now because I can just write the tests that I want a piece of code to accomplish, and then it will generate the rest." My, like, way of working a developer has changed so that I'm just writing these tests.
14:41 Auren Hoffman? You're starting with the end.
14:42 Speaker 01 Yeah, I'm asking it to research, "How do I accomplish this test?" And he had this, like, really amazing set of three systems working together that probably would have taken a group of three to five, like, six months, a year ago, to do. And we had it two weeks.
14:59 Auren Hoffman? One of the things also, it's like historically I've worked with a lot of backend engineers. Mostly that's where I've worked and they're generally not that good on the front end stuff. And it's just a very different skill set. And now I've noticed so many of them now get to do the front end too, because they can use all these tools to do. So it's like they have this kind of creative side that they never had access to. And they're like way more dangerous in a good way. Like they can actually just do so much more stuff before they'd have to work like more collaboratively with others. Or now it's like the one person, so you don't have this communication problem, just can get so much more done.
15:35 Speaker 01 It's funny, I'm the opposite version of that engineer. So I've always been a product guy first. And then I started out building software in the design layer. Like I was in the nineties really into design tools and things like Dreamweaver. And you know, these sort of front end design tools.
15:54 Auren Hoffman? Trying to date yourself, Dave.
15:56 Speaker 01 I know. And I really always came at it through the front end. And so I've never been a good backend engineer. Like I always have relied on partnership and collaboration with people that are great at that. I think it's like how my mind works. Like I'm just not good at that type of engineering. But that's been really fun that I can do kind of the opposite. I can start with a front end idea and then now build out the server and the database really, really easily. Using these tools and just test stuff out. It's cool. It's very creative and awesome and it's kind of a beautiful thing. But that's the way that I try to stay, I guess, current. And I try to keep my mind from going to, I want this, I want this.
16:40 Speaker 01 Cause like I find that to be something that pushes me to think about the current investments that I'm doing. And perhaps ways that are not productive for me.
16:51 Auren Hoffman? World of DAS isn't just a podcast. World of DAS membership is a private invite-only community for founders and executives of important data and AI businesses. It's a place for off-the-record, high-trust conversations focused on the biggest questions data leaders are facing. The community is about peer learning, curated intros, and access to private events. Everyone in the group is building or leading something meaningful in data. If that sounds like you, apply to join at worldofdas.com. That's world of D-A-A-S dot com.
17:32 Auren Hoffman? And Offline is both a fun, but it's also a venture studio. It's like, I don't know, 80% fun and 20% studio, right? Or something?
17:40 Speaker 01 Yeah, no, that's exactly how we split it.
17:42 Auren Hoffman? So that's why you're doing all the vibe coding?
17:45 Speaker 01 Yeah, I do vibe coding, but the way we structure Offline is we've got a standard GP, group of people that make investment decisions. We're out there doing your standard pre-seed, seed investing. I've done now over 1,000. I've also got the 10,000 hours there. I've done over 1,000 pre-seed and seed investments in my career.
18:06 Auren Hoffman? Holy mackerel, that's crazy.
18:08 Speaker 01 And so we do that. We're really good at it. I think we're best-in-class practice at that. I could talk about it all day. But I also enjoy building things, and we enjoy building things. And we enjoy helping other people build things and maybe think about the way they're building them differently. And so we also have an incubation studio.
18:28 Auren Hoffman? Do you have like a fund structure and they're separate? Or are they part of the same thing?
18:33 Speaker 01 They're all part of the same fund. And the way we've structured it is 80/20. 20% of our capital goes to the incubation practices. So, you know, we run $100 million funds. And so 20% of that, minus the fees and whatnot, goes to running the studio. And then we have a team. Max Ventilla, I don't know if you know Max. He's been in the Valley forever. So Max leads our studio.
18:57 Auren Hoffman? Of course, Max is amazing.
18:59 Speaker 01 He runs that process with a group of people that many of them actually worked with him on Alt School, his big school project. Bart Mina Rasa.
19:08 Auren Hoffman? I was investor in Aardvark.
19:10 Speaker 01 Oh, cool. Yeah.
19:11 Auren Hoffman? That really dates me, but that's a great company.
19:14 Speaker 01 I mean, I tell everyone that Max is one of the greatest product guys in Silicon Valley. Like, Ardvark is one of my favorite products of all time. He came to me a few years ago and was like, I want to start an incubator. And I said, well, we're already incubating things, but we haven't actually made it official. Like, we haven't made like a really structured process to that, but we want it to be a serious endeavor here at Offline. We don't want it to be just like a hobby that each GP does.
19:40 Auren Hoffman? When you incubate something, so then it's the fund that owns the common stock in that company that you incubate? Not the GPs or whatever it is, or?
19:53 Speaker 01 No, the fund does. We start around two projects per week and we kill around two projects a week. And we run a nine-phase funnel that's kind of like DARPA, where we give each project a really structured timeline and budget.
20:09 Auren Hoffman? And the budget is basically a certain number of hours of people or something?
20:13 Speaker 01 Yeah, and then we have a team of around 70 people that work in a part-time way in our studio pool. And they will work on that timeline with that budget. They'll figure out if an idea is viable. And then if we decide that it's viable, then it moves on to being prototyped. If we think that the prototype is good and we think there's something here, then we go on and we start trying to figure out, like, what's the product that's likely going to find a market? And then we try to discover product-market fit. Once we've discovered product-market fit, that's when we actually start recruiting full-time teams to things.
20:48 Auren Hoffman? Just to like go with Nebraska tax, so obviously like, let's say, I don't know what your friend is, but let's say it's a two and 20 fund or something like that. So you've got the management fees, but you're saying separate from the management fees, you're taking actually a piece of the fund, 20% of that fund, and you are paying people, which could be people on your staff or part-time people, whatever, to create these companies. When you do that, if these companies are successful, your fund is getting equity in those companies. And that's how your LPs hopefully eventually make money back and, you know, one of them. They're going to make a massive amount of money.
21:23 Speaker 01 Yeah, and we own 3X, the amount of equity in things that come out the end of our studio funnel that we do in a typical seed investment. And so that's kind of our main pitch here is that, you know, we end up owning instead of 10%, we own 33% of everything that comes out of our studio. And so if we do win, we end up with much larger ownership, much bigger wins. And that's partially, you know, comes out of my experience doing seed investing in this market for the last, you know, I've been doing this for just over 15 years now. You know, there's a lot of competition. I mean, I know you know this, Arin, like you just started a new fund. Is it a year old, two years old?
22:04 Auren Hoffman? It's like seven, but you know, who's counting?
22:06 Speaker 01 I thought your new one was newer than that.
22:08 Auren Hoffman? It depends what you say, but yeah, I think the current fund of the seven year fund is about a two and a half year old fund.
22:14 Speaker 01 Yeah, so two and a half years.
22:16 Auren Hoffman? Yeah, Flex Capital.
22:16 Speaker 01 Yeah, Flex. So Flex, you're used to this, right? It's super hard to get the 10% that you want super early on. There's like a lot of competition, you know, you're doing, I think, amazing job marketing far better than us.
22:30 Auren Hoffman? By the way, we usually get one to 2%. We're usually investing, sometimes maybe up to 4%, we're investing 300 to 500K in a company. We're doing high volume, low dollar in a company because these are companies typically raising $3 million in a pre-seed or seed. And they don't have 2 million of allocation to give. It's very competitive to get in.
22:53 Speaker 01 Yeah, and so we're up against that same thing and we play the lead role in a lot of those rounds. And so typically in our normal practice, we're playing the seed or sorry, the lead role in those pre-seed and seed investments. And it's very competitive. And so the question of how do you get higher ownership in the things you have the most conviction in is one that I think anybody running a fund should be asking themselves. And so this is the way we've gone about solving for that. I think on the one hand, and then on the other hand, we just love building. I'm a product guy at heart, like I just love building things. And occasionally I'm out there, I'm talking to people, there's like an area where I'm Like how like there's nothing in this zone. Why not? Like, we should try to build something over there.
23:38 Speaker 01 It's nice to be able to do that. And I've always wanted to have that function in the funds that I've worked in or around. And so we just wanted to build it into the fundamental DNA of what we're doing at offline. And so that's what we've done.
23:53 Auren Hoffman? When you have a project, it's not like you quote unquote, fund the company. You just pay people kind of directly out of the fund. And then if the company is successful, then you quote unquote, fund it or something. Is that the way it works?
24:05 Speaker 01 It would take me a whole podcast to explain the model. But yeah, we have effectively a model that starts out as a pool of resources. And then as the ideas graduate to a level where they deserve to have a safe, then we fund a safe. And then as those safes pile up, then it turns into a traditional round. And by the time it's turning into a traditional round, like what we've incubated should be to you or anybody that's got great taste in pre seed and seed investment, it should look and feel and you should think it's a great seed investment. It should feel de-risk.
24:38 Auren Hoffman? How do you find these people? How do you recruit them? Or how do you get these other people involved?
24:44 Speaker 01 We view it as a core competency of our studio. And so we've got a great recruiting team that actually works across both our portfolio companies and our studio. Our head of talent, she's been working with Britt and I for, you know, across the various of our companies for now over 15 years. And so she's super talented.
25:05 Auren Hoffman? And there's a bit to them. If you have two founders and you guys, it's like, it's like a third, a third, a third ownership in the end, or how do you, what's the pitch to them?
25:13 Speaker 01 It's always custom. Yeah, it's just like every deal in the valley. The circumstances are different. Sometimes the IP that's coming along has more value. Sometimes it has less value. Sometimes the founder needs to be a different caliber for this type of company. It's pretty custom. And we've actually found that. As we've been building this out, we thought we were going to have one model. Of course, that was foolish. We maybe shouldn't have thought that, but. We've kind of iterated along and we iterate so much. We have so many iterations going on all of the time that we learned quite quickly.
25:46 Auren Hoffman? You're also iterating on a business model, like a fund structure. One of the things that's interesting is like basically every venture fund is essentially structured the same way. But every incubator that I've ever seen is a complete snowflake. I've never seen even two incubators that are structured even. So you're creating a whole new business model.
26:05 Speaker 01 Yeah. And I mean, that's why we look at it as like, you know, every business I've ever run, I try to take 20% of our capital that we have to, you know, allocate and allocate it to innovation. And so that's been our conversation and our goal with offline is 80% of the time we can do world-class seed investing. I mean, we will do that and we will go out and we will buy a portfolio of 20 to 25 companies. Every fund will do that, but for 20% of our capital, we do think there's an emergent new model here around studios. And I've also seen this in my career that when I'm building, I'm a much better investor.
26:49 Speaker 01 Some people just do that. They're like, I'm going to be a solo GP and that's going to be my side hobby along with the operating CEO role that I play. There's a lot of that going on. And by the way, I did that for five years. I was, you know, running path and I was also angel investing and then that turned into, I named my slow venturing or my angel investing practice, Slow Ventures, and that turned into a fund. That was all while I was building path, but you know, this time around, like this is my primary thing. This is the primary organization that I'm running and help co-running. And so I also want to build. I actually think I can create good returns and higher ownership for my investors by building. And so that's what we're doing too.
27:34 Auren Hoffman? People are now having like these personal relationships, truly personal relationships with AI. Like, how do you think that evolves over time?
27:42 Speaker 01 You know, I've done a little bit of writing about this recently. I quit most social media now almost one year ago. I've been off of all meta apps. I'm only on Twitter largely because that's where the venture capital business happens in a lot of ways. And so I-
28:00 Auren Hoffman? And LinkedIn, I presume, right?
28:02 Speaker 01 Yeah, I'm on LinkedIn. I guess you could call it pop mainstream social media. I am not on particularly the video based version. I'm just, I left.
28:10 Auren Hoffman? Yeah, those are, it's like crack cocaine.
28:13 Speaker 01 The reason why is that, back to my history working on Path, we built Path around the idea that the human brain can only maintain 150 stable human relationships at a time. And there's no like way to expand this. It's not like the brain's changing or anything like just because there's mobile phones in everyone's pockets doesn't mean all of a sudden you can have, like break this biological limit.
28:37 Auren Hoffman? Do you think that's true? Do you think like Dunbar's number, all these things, do you think they're actually true? Like-
28:42 Speaker 01 Yes.
28:43 Auren Hoffman? Do you think you personally can only track 50 relationships going on?
28:47 Speaker 01 Well, so he has several rings. So when we were doing the product, 150 is the maximum. Like famously Gore, who makes Gore-Tex, only allowed 150 people per building. The max number is 50. So 50 is the number of people that like you'd be legitimately sad that they died. Like if you were sort of throwing a birthday party, you probably invite around 50 people. And then the next ring in is 15, which is these are the people that you'd trust, that you'd go on a weekend with, that you know, you want to be around.
29:19 Auren Hoffman? That if there's something going on in their life, you literally will go fly to Dallas to deal with them.
29:24 Speaker 01 Yes. And then, you know, and that's also you're the grand master of dinner parties and small conversational parties. 12 to 15 is kind of the ultimate size. You can't really have a single threaded Jeffersonian conversation beyond 15 people. And even 15 is a lot.
29:42 Auren Hoffman? Even 15 is really pushing it, yeah.
29:44 Speaker 01 So there's a natural limit there. Five is the number where it's like, these are your ride or die, like your best friends. And then interestingly, halfway through Path, Robin Dunbar said to me, I discovered a new ring. I said, oh, like, what is it? And he goes, there's one at 1.5. I said, what? Like, does that mean it's like, you know, your wife and half of one of your friends? He's like, no, there's a gender difference. Men only have one best friend, and if they're married, it's their wife. And women have two. They have their husband and then one best girlfriend. And this is like, I found this in all my research. And so back to the question at hand, what do I think about AI relationships?
30:25 Speaker 01 Well, I think in social media, what happened was people started replacing slots in their 150 with parasocial relationships. Parasocial relationships are like the relationship everybody has with Taylor Swift. They think she's her friend. They feel like this person in social media is their friend.
30:44 Auren Hoffman? But we've had that for a while. I mean, you know, I had parasocial relationship growing up with Gandalf. You know, this like fictional character and obviously so many people read the Bible and we have parasocial relationships with biblical figures, Moses and Jesus, and this is not new.
31:01 Speaker 01 No, what social media did was like dramatically and rapidly exploded. And I think the pandemic in particular accelerated it even more dramatically. And so I think when people talk about the-
31:13 Auren Hoffman? Because you weren't seeing people, so.
31:15 Speaker 01 Yeah, I think when people are talking about the loneliness crisis, what they're actually talking about is that people have many of their 150, their slots in their head taken up by parasocial relationships. So when they go out into the real world, they're not just lonely. They're actually wondering where their friends are. And it turns out that they're not real. They're just in your head. And so when I started to realize this, I was like, oh, man, I'm out. I want all of my energy going towards the people that I'm interacting with in person. You and I keep in touch, like I'm following up with you. Let's do the podcast, like the people who you're proactively interacting with and the people who I'm seeing in my real life, having dinner with.
32:00 Speaker 01 Those are the people that should actually be filling my head with the social food that I should be eating.
32:06 Auren Hoffman? Just to kind of double click on that, it is much easier if those people are in close proximity to you geographically.
32:15 Speaker 01 And they should be. Oftentimes, when I explain this to people, they're like, oh, I want to keep in touch with my friend that's from college. And I'm like, well, it's fine. But like, you can text them if you want to catch up with them, right? Like, just exposing your mind to this information diet that is full of people that aren't actually in your life is not actually healthy for you. And so that's my view on AI, which is like, is this going to radically accelerate this even more dramatically? But I'm not sure. I have different worries, actually, about AI. If you look at what social networks did, they actually stopped us from socializing in person. If you talk to any college kid these days, they'll tell you, like, I don't go to the bars anymore. Your and I's experience of college is radically different.
33:00 Auren Hoffman? When I was in college, I went to bars to meet girls. That was my number one reason because that was really the best way to do. I mean, I didn't know, this is the best way I could think of to go do it. Whereas like, if I was in college today, I would use the apps. Like, that's just like a much more efficient way to meet a girl.
33:18 Speaker 01 Yeah, but it turns out that it's not. When you look at all the studies, there's tons of them. It's like Tinder optimizes for getting you to swipe, not actually connecting you to people. It turns out the apps are actually high fructose corn syrup for this. And my joke lately has been, just like after World War II, and we all stopped working out, or we stopped using our bodies and started working in offices, the gymnasium became popular, that we're probably going to need social gyms now because we've stopped socializing. And it's interesting, I actually saw a pitch last week for Duolingo for socializing. And it's coming from a kid that's in Gen Z, and he's just like, none of us were taught this. We all ended up in the pandemic. We don't know how to interact
33:58 Speaker 01 with not just girls, but like, business associates. And I'm like, wow, here we are, like, having to flex that muscle and figure it out. And so my worry about AI is actually that we're offloading thinking, and that actually, 10 years from now, it's going to be like, you're going to go To a conference and you're going to be challenged to actually use your mind to think and write an essay. And like, do the things that we're all offloading into AI. In addition to the worries that you've posed.
34:27 Auren Hoffman? I mean, it is weird that, I mean, I work out, you know, go to gym and workout. A hundred years ago, no one did that, yet they were probably like way better shape than I am in. Just because that was just like part of their day or something. And now like, we have to go to the gym and artificially like, lift these big iron things for some reason.
34:47 Speaker 01 Yeah. Meanwhile, I'm getting pitches for Duolingo for socializing. Duolingo for social skills. This is the world that we're in, right? Like behind every human invention is a atrophied human faculty. That's what we're doing, right? Like we've kind of atrophied socializing, thanks to social networks. And so are we atrophying thinking now?
35:08 Auren Hoffman? I mean, there are books, you know, there are all these fictional books about, people whispering in your ear to tell you what to say to a girl to woo her or something like that. That's been a trope for many years where we need help socializing.
35:22 Speaker 01 Yeah. And look, don't get me wrong. Like I helped build these things. I'm not anti. But I'm just kind of trying to be a realist about where are we now? AI is amazing. Like you can learn such incredible things, right? And to your point, like, it's pretty amazing to be able to just say, hey, like, I think one of the coolest use cases of AI is the one that you hear from people all the time where they're like, oh, I had this difficult relationship problem at work or in my life. And I didn't know what to say. And so I talked to AI and it helped me understand it. Oh, actually, I didn't think of it that way. And so I approached the conversation differently. And it actually made the relationship better. Like, that's awesome. Like, that use case is awesome.
36:05 Auren Hoffman? We have a, I think, a common friend who had a fifth grader who did this. And it was super helpful. I know she was having an issue with someone. She didn't know how to bring it up and she kind of game played it with chat GPT and then figured it out herself and then had the conversation and actually went really well. Cause these are hard conversations to have.
36:23 Speaker 01 Wouldn't that have been awesome to have that in fifth grade, right? Like, maybe we all would have been bullied last. I don't know.
36:29 Auren Hoffman? I wouldn't have been mature enough in fifth grade to have even thought about it. It's only a very, very mature fifth grader probably could do that.
36:36 Speaker 01 And that's cool, right? So I don't know, we may end up with a bunch of positive, but like a bunch of thinking gems in 10 years.
36:45 Auren Hoffman? I think about the hardware stuff. I mean, there's like the glasses, the pins, you know, some of these things kind of like the benefit is they kind of free us from the screen, like looking down at the screen. But there's, you know, a little bit more of augmented reality that's happening. Like, where do you think that's going?
37:03 Speaker 01 It's hard to tell. Casey Neistat had a great video about this last week where he showed the difference between Apple's thinking and Meta's thinking, like Meta's got these like plastic sunglasses with cameras in them. And they're kind of janky and cheap. And then on the other hand, Apple has this like, whatever $4,000 device that's like ski goggles. And they both kind of do the same thing. And so they're kind of going in the same direction from two wildly different angles, right? That you're going to have these cameras on you and this thing you can talk to.
37:35 Auren Hoffman? And potentially a screen in your eye that give you some extra information.
37:40 Speaker 01 And maybe that's right. When I saw that, I'm like, oh, that's cool. But like, I was joking that, do I want people to have deep fake eyes? Which is eyes that are looking at you, but not actually looking at you, you know? Like, I don't know. Is that the right way? I was an investor in Humane. I actually thought that Imran had a lot of really good thinking. Their execution path didn't end up in the right place. A lot of people panned them for it, but in my view, like at least they were trying.
38:07 Auren Hoffman? Yeah, it's really hard.
38:09 Speaker 01 Who knows? Like yesterday, OpenAI was doing a developer day and Johnny Ive was on stage and he basically said the same thing. He was like, yo, this is hard. We've got a family of things we're working on, but we're kind of not sure. And Sam Altman's like, yep, it's gonna be a while. I think that's actually true. And to your point earlier, like, we're probably just marching towards this kind of incrementally. I use the Meta Ray-Bans all the time. I love them.
38:37 Auren Hoffman? So do I. I love them too.
38:39 Speaker 01 I went on a vacation with my wife and some friends in Europe this summer. And it was so awesome, like to have those on and. It enables you to capture kind of a different part of the experience. It's not the photography that I love doing. Like, I still carry my camera to do the kind of photography that I love. But I really love that I have this new stream of videos in particular that those things can capture.
39:04 Auren Hoffman? I went on a safari and yeah, of course you have the camera, the safari or something like that. But just like the meta glasses, if you just have to like a super quick thing, you know, there's no way to get your camera up, even your phone, it would just take too long. It's so much easier to do it with the glasses. And then in some ways allowed me to be more present, too.
39:23 Speaker 01 Yeah, and credit to them. They actually solved, I think they deserve huge credit for like a breakthrough. My sense is it was kind of an accident. It's like your classic story of innovation. They put something like, you know, the Apollo project's budget into the metaverse and VR. That totally failed, but there was this one thing that was like, all the numbers were hockey sticking. And it was this little plastic Ray-Ban glasses. And so they were like, oh, and to their credit, they just like doubled down on that. And they're like, oh, that's where all of our AI metaverse stuff is going. Cause it's clearly the form factor that works.
39:58 Auren Hoffman? I love the new ones where the lenses change if you're indoors. Then I can use it instead of my AirPods, right? Too. So now if I'm indoor or outdoor, I don't have to be like switching all the time and stuff. Or if I walk into a store from outdoors, I can keep my glasses on and it's great. I mean, it's a great product. Speaking of advertising, we are in advertisement right now for this thing.
40:19 Speaker 01 Cause it's a good product. Great products deserve to be talked about.
40:23 Auren Hoffman? You and I are not getting paid for this, but we're in advertising like this is the type of advertisement is like, is a good advertisement, right?
40:28 Speaker 01 Yeah, it's like focus on making a great product and they sell themselves. And in this case, I think it's also the answer to your question, which is the famous William Gibson quote, "The future is here, it's just not evenly distributed yet." That fact that all of the technologists are using these glasses and loving it, I think actually says like, oh, this is the breakthrough use case. And yeah, not everybody's doing it yet, but I think you should expect to see like a wild Cambrian explosion.
40:57 Auren Hoffman? The lightness is really important to me because I don't want this heavy thing on my nose, but it's light enough. If that new on-screen display in your right eye, I don't, I don't know if it's going to work that well, but assuming that it will get there over time, like to me, that would be great. It gives me, it gives me a little heads up. It's a little map. You know, if I'm biking right now, often I'm biking and it's hard to look at my phone where I'm going. If that could tell me where to go better, these would be real beneficial things. I know you're a big skier, so if it just like gave you a little bit better map of the trail of where to go and stuff like that.
41:32 Speaker 01 There's great versions of this, actually. There've been lots of goggles over the years that have had electronics embedded in them. And almost all of them, you're like, oh yeah, this is pretty great. But then they have bad battery life and the displays don't work very good. So yeah, I mean, if you could have marginally new information, like altitude, speed, like all kinds of things like that, what directions your friends are on the mountains, like, yeah, this is all like great stuff.
41:56 Auren Hoffman? Andorl's building like these types of things. They're building helmets for the military. Essentially this, there's the helmets that everyone in the military needs a helmet anyway. So you can pack a little bit more, they're wearing armor, so you can put the battery packs in the armor, right? So you have a little bit more ways of moving it there. And then of course you've got the visor and that gives you a little display of letting know what's going on. And of course in your ears, the comms, so you can talk to other people. So it makes sense to have like an integrated system.
42:25 Speaker 01 I wouldn't be surprised if there were also sort of voice version of all of this, like some kind of thing you carry around, like the other like clearly extremely valuable AI use case is that people are getting a lot of value out of note taking and summarization. What was one of the hardest things to do in every organization, startup company anyone's ever worked in? It's, I just had a meeting, we talked about all of this. I'm trying to re-download it to you that this person that wasn't there. And like, that is now just solved. And so I think what you're also seeing is like a lot of, I'm an investor in this company called Rewinds. You know, this is Dan, the founder of Optimizy.
43:06 Auren Hoffman? I tried to invest, but not able to get in, so.
43:09 Speaker 01 Yeah, you know, he's done a lot of iteration on this, like, how do you create the perfect memory? And they've created a actual hardware device that's good for this. And is the form factor perfect? I don't know, but there's like a lot of versions of this right now. And here's the other thing I know, and you drive a Tesla. We talked about it earlier. It's kind of awesome. Like, my kids love talking to Grok in the car. And they especially love talking to the version of Grok that's like a surfer bro. And I've actually found myself sometimes just talking to Grok about like the local news in my hometown, for example, in Montana, which is like very hard to get, actually.
43:43 Auren Hoffman? Somehow I'm still talking, even though I'm in my Tesla, I'm talking to ChatGPT. It's just easier. You like talking to Grok like right there because it's just embedded in it.
43:53 Speaker 01 Yeah, I'm surprised, right? And so that's all voice.
43:56 Auren Hoffman? I like having the ChatGPT conversation. So yeah, I'm going to start doing the Grok ones, too.
44:02 Speaker 01 There might be a device that like fits into that zone that looks more like AirPods. You know, like, everybody carries this little AirPods thing in their pocket now, right? Like, maybe there's something like that. Like, I could see a whole family of things in this. This is why it's not surprising to me that ChatGPT and OpenAI are saying we're doing a family of things. There's like a few of these kind of use cases that seem clear that are going to be the focus for this.
44:28 Auren Hoffman? Switching topics, I know you're, or maybe even the same topic, I don't know. I know you're focused on like, really focused on curing depression, which is like a massively ambitious goal. Like, where do you think we are along that path?
44:41 Speaker 01 Yeah, I started working in mental health in 2015, right after I sold Path. And I've put a lot of philanthropic work over the last decade into trying to move the ball forward on depression, mental health. I use a pretty aggressive phrasing of trying to find the cure to try to just like move through the world. I think that 10 years later, we have a much deeper understanding of the types of people that are affected by depression in its various forms. We understand lifestyle interventions that are deeply helpful. You know, there are some people that need to exercise more. There are some people that if you tell them to get more exercise, it makes them more depressed.
45:26 Speaker 01 We're getting a lot better at helping people identify what type of person they are. And then what kind of interventions that they need. We also understand much more deeply. The social interventions that we need. So some of the biggest studies in the world. Related to depression, at least in the United States, are ones where they look at large populations. Of people that consistently get depression. So for example, doctors. Doctors in training consistently get depression. And it turns out there's a linear relationship with the number of hours that you work and depression. The more you work, the more likely it is that it's going to happen. And so. You have these things where it's like, it's become very popular in Silicon Valley, do like nine nine six, we're gonna work ourselves into the ground.
46:11 Speaker 01 Anybody who's on a professional sports team knows that if you push it too hard, you're not going to be able to perform on game day. And so there's like a lot of things like this where we're starting to understand the.
46:22 Auren Hoffman? I'm super skeptical of that.
46:24 Speaker 01 I'll send you the study.
46:25 Auren Hoffman? All right, okay. I mean, if you're working on things you don't like, sure, you have a terrible boss and you know, whatever, but. If you're working on things with like good people. And you're working on hard problems, like I feel like that's almost like the antidote to depression. It's like a shared mission that you have with shared colleagues. And you're in the trenches together.
46:45 Speaker 01 Not if it cuts into the things, the foundational things, which are nutrition, sleep. If you just think of it as a pure sliding scale between sleep and work. If your work cuts into sleep to the point where it damages your circadian rhythm, depression every time. That's like the simplest way to think of it. Now there's a wiggle's room, right? Between like how much work is too much work and creates too much stress that impacts your sleep.
47:10 Auren Hoffman? I mean, my guess is most people are on the extreme other side and they need to. Find more meaning in work and need to work more rather than there. I'm sure there's some people who are. You know, we might know some people that work too much and gotta cut back. But I'm skeptical that there's that many of those people. I feel like there's way more of the other people that would benefit more from working more and having a more shared mission and that would be an alleviator of stress. Yeah.
47:36 Speaker 01 Absolutely. There's no like, you have to have purpose. Again, that's another thing that we know. The other thing I'll say is that we also found really great treatments that are breakthroughs that were kind of hidden from the world. You know, we know that psychedelics, in particular ketamine, are incredibly powerful tools in the toolkit. They work for many, many people that don't work for everyone, but for the people that they do work for, we have scientific studies that show 80% plus efficacy. It's so far above the traditional treatments that it's not even close.
48:10 Auren Hoffman? On something like that, because it's like, it's a bit of a big intervention to go use these pharmacological things like, is it that kind of thing where, okay, that's something to look at if the depression is very severe. I assume there's some of the gradation and you've tried these five other things and it didn't really help you type of thing.
48:29 Speaker 01 Absolutely. And that's what you're typically seeing, right? You've got doctors that are out there with patients that have seen no remission in, I don't know, call it five years of psychotherapy combined with lifestyle interventions and nothing worked and they're still suicidally depressed.
48:44 Auren Hoffman? Well, how big of an epidemic is it like? And is it getting worse in the US?
48:48 Speaker 01 And it's getting worse.
48:50 Auren Hoffman? And is that a because we're getting richer? Is that because of the social media thing? Is that a?
48:54 Speaker 01 There's a lot of different theories. I mean, I think social media has perhaps accelerated it. I think we were already on our way to 2030 has always been the number where depression would be the number one contributor to the burden of disease in the United States, like above cancer.
49:11 Auren Hoffman? Whoa.
49:12 Speaker 01 It is the epidemic that it's number one. Like, we're already on the way. And these were numbers that were already, when I first started working this, this was the 2015 to 2020. That was the projection in that time zone. And then we had a pandemic, which I don't think anyone disagrees that that exacerbated the mental health crisis.
49:31 Auren Hoffman? There's some debate on things like marijuana, whether it's causing depression, it's accelerating it. There's some other people who think it's helpful. I've seen a lot of different debates on this.
49:41 Speaker 01 Cannabis is a difficult one. I've researched it, talked to many of the scientists invested around it. It's a difficult one.
49:51 Auren Hoffman? Especially like, you know, these 18 year olds, you know, maybe inducing things like schizophrenia or some of these other things.
49:57 Speaker 01 The science is 100% clear that you should not use cannabis under the age of 21, really. Like, the science is super clear. Now, is that actually a reality in the United States? Like, it seems like, no, kids just get it. But Andrew Huberman does an amazing episode on cannabis. If people want to get really into the science of cannabis and this question, I think his episode is fantastic. He's actually done a couple of episodes. It's a difficult one. I've seen it really help people, particularly with anxiety. It can have an enormously positive effect. It doesn't appear to be something that you want to be on long term. But guess what? People are also on antidepressants and anti-anxiety drugs for 10, 20, 30 years at a time.
50:41 Speaker 01 And so it's hard to look at cannabis as a part of a mental health stack of medications and be like, oh, that's that much different than like taking SSRI for a long period of time. It's all kind of like, use this tool to get yourself back to neutral. And then work on the lifestyle interventions that really take you to thriving. It's really easy to point out the things that take you to thriving. It's sleep, strength training, drinking a lot of water, healthy relationships, not drinking alcohol. Don't drink alcohol at all. If you do, drink it very carefully and absolutely do not smoke. You know, if you want to live a long, healthy life, right? And so these are like relatively simple things that have been proven by science.
51:26 Speaker 01 And, you know, you can get into Huberman's podcast on this and Brian Johnson's doing a lot of work on this right now and many of the longevity scientists.
51:33 Auren Hoffman? One of the things about SSRIs is, you know, they make your synapses fire more. And another thing that make your synapses fire more is carbs. I eat tons of donuts and it's just like makes my synapses fire. And I'm just like, you know, it's just, it causes happiness.
51:47 Speaker 01 The question you asked was really simple. Where are we on the path to the cure? We're not there. I think that everything you've mentioned and that we've talked about are all in the toolkit. But do we have a GLP one of depression? No, we don't. And so we're still out there searching and, you know, the reality is we really don't understand the brain very well. Like, we don't know if it's a computer or an antenna. We don't know what it does. We know functionally, when we look at it with an FMRI, we know a lot of things. We know a lot of things biologically, but humanity has a very loose understanding of the brain. And we still have a long way to go research wise.
52:23 Auren Hoffman? Three more quick questions. One, first, you're co-hosting the More or Less Podcast with your wife, Britt, and then Jessica and Sam Lessin. What have you learned from doing a podcast? Like, what advice and, you know, do you have as a podcaster?
52:37 Speaker 01 That I almost always am wrong about whether the episode was good or not.
52:44 Auren Hoffman? Oh, interesting. So afterwards, you're like, oh, that was a crappy episode. And then like, and then it goes through the roof.
52:50 Speaker 01 Goes viral.
52:50 Auren Hoffman? And then otherwise, oh, this was great and like, no one liked it?
52:53 Speaker 01 Never liked it.
52:54 Auren Hoffman? Oh, interesting.
52:55 Speaker 01 I have somewhat gotten better at it. I learned that I just, I don't have a good radar for it.
53:02 Auren Hoffman? And that that's just by number of views essentially, or a number of listens?
53:06 Speaker 01 Yeah, you know, you can look at it by views or the amount of feedback you get or listens or I guess, yeah. Looking at my primary metrics, which is views and time, right? Like, did they make it all the way through? The ultimate metric of a podcast is not just a view, but like, did they make it all the way through?
53:24 Auren Hoffman? Right, this is at the end of this podcast, so probably no one's listening at this point.
53:27 Speaker 01 Yeah, if you are listening, I always say, send me a DM on Twitter. You know, like, I love hearing from people.
53:34 Auren Hoffman? Yeah, DM me on Twitter or on off minute or on Twitter.
53:36 Speaker 01 If anything I said was interesting, like, let's keep the conversation going.
53:40 Auren Hoffman? Totally. My dad will be DMing me, that'll be it, yeah.
53:43 Speaker 01 Yeah, and that's like, I think one of the hardest things about podcasting is you oftentimes don't know who's out there. But I've also learned a lot about comedy and how to play different roles in groups, in particular in comedy groups. I play a role on the podcast, which is known as the straight man. I'm the Seinfeld to Sam Lessens Kramer. And so, being good at that and not competing for the bit is something that I've learned a lot about. I don't know that I'm necessarily even that good at it.
54:13 Auren Hoffman? Albert and Costello, you're the Albert and he's the Costello.
54:16 Speaker 01 Yeah, I don't know that I'm like that good at it, but I'm like getting better.
54:19 Auren Hoffman? That makes sense based on your two personalities.
54:22 Speaker 01 That's like a thing I knew nothing about before starting doing this.
54:26 Auren Hoffman? Last two questions we ask all of our guests. What is a conspiracy theory you believe?
54:30 Speaker 01 I'm really into the UFO question. Like, what are they? All throughout history, in the history of all religions, there've been things in the sky, Ezekiel's chariots. You know, in the Middle Ages, people were seeing ships in the sky. But and then at the same time, our military seems to have readings on radars of things in the sky. And so I'm really into the question. Like, I don't know whether that is an answer to your question. Do I believe in it? I believe in the phenomena. I believe people are experiencing this thing. Whether or not they are flying metal objects from another galaxy, I don't know. Are they something that we just don't understand? Is the world flat? I don't know. But that's one that I enjoy thinking about.
55:15 Auren Hoffman? All right, last question we ask all of our guests. What conventional wisdom or advice do you think is generally bad advice?
55:21 Speaker 01 I already said it. I think that 996, this notion that you should destroy yourself, live in a garage, eat ramen, not take care of your health, and have no social relationships, and you can still be a good entrepreneur.
55:35 Auren Hoffman? 996 doesn't seem that crazy. It's not that much, actually. Like, when you think about it, a lot of people, I mean, when I was that age, I was definitely doing, I was doing 11 and 11, you know, whatever, yeah, or 995.
55:47 Speaker 01 It's funny that you say that. I was gonna tweet, like, 996, like, if you Gen Z kids want to create trillion-dollar companies, you need to do what we did at Facebook, which was basically like 10, 12, 7 for several years, right? And, look, it almost killed most of those people.
56:05 Auren Hoffman? You can go out afterward. If you go to bed by 1 or 2, you get a decent sleep, right? You get the Sunday or Saturday off. And, yeah, again, if you don't have kids, like, I don't know, it doesn't seem like that crazy. If you're 24.
56:17 Speaker 01 It's not that crazy. Yeah, if you're 24 and you don't have kids, yeah, work all the time. But the reason I pointed out is, take care of yourself. Like, take care of your health. Like, don't damage your mental health to the point where you're suicidally depressed. Don't damage your body to the point where it shuts down and you can't show up to work the next day, right? Like, I like to be a moderating voice on this because I just compare it to sports or athletics. Like, even just look at the trend in how musicians are managing their careers now. You look at people like, I don't know, I think Zac Brown's a really good role model, like.
56:53 Auren Hoffman? Who's that?
56:54 Speaker 01 He's on social media, Zac Brown Band, he's a country rock star. On social media, he shows how he brings a gym trailer with him everywhere that he goes, and he treats what he's doing as a sport. And you look at Taylor Swift, for example, like, she's now really talking about how she has to do training and, you know, nutrition management, and she treats herself like an athlete.
57:16 Auren Hoffman? She is an athlete, I mean.
57:17 Speaker 01 And being a founder is athletics. You have to treat yourself like an athlete. It means high performance, intense work, you know, driving a hard culture of performance, but it doesn't mean destroying your mind and your body in service of it so that you can't do the job. I think that's like, to me, there's nuance in that advice that we should take more seriously.
57:39 Auren Hoffman? Honestly, I wish more of my founders were 996. I feel like so many of them are 9 to 5, 5 or something like that. And, yeah, I mean, maybe not the founders, but people in tech.
57:51 Speaker 01 Maybe you're right. Maybe that's where this is coming from. Maybe I'm missing something. Maybe the employees are a little bit two nine to five out there.
57:59 Auren Hoffman? I think it's pretty rare for employee at a company, even a tech company, to be working even 40 hours a week nowadays. Yeah, I think very few people at Google work 40 hours a week.
58:10 Speaker 01 Really?
58:11 Auren Hoffman? A real 40. I would say they're working 25 to 35 hours. Yeah. I would say almost, I would say it's pretty rare to find anyone at Google pushing 40.
58:21 Speaker 01 You're saying at the meg seven. I thought you meant at startups.
58:25 Auren Hoffman? I believe at Nvidia, they're working at least a real 40, but I don't think at Google, they're working a real 40. I don't think at Meta, most people are working a real 40. They might be clocking in, but they're spending, you know, when they're there, they're spending time on buying things or on social media or on TikTok or whatever.
58:42 Speaker 01 Interesting.
58:44 Auren Hoffman? And of course, if they're working from home, I think it's very unlikely that they are. So I could be wrong, but I don't have a camera on everyone, but my guess is just like productivity is not high for that reason.
58:55 Speaker 01 That's interesting. I thought initially you were saying that this was startups.
58:59 Auren Hoffman? Even in startups, my guess is like, yeah, the first four or five people are usually working pretty hard. But at some point, I feel like it just kind of goes, I think they're working harder than they are at Google. This is pretty rare to see like the average person working 60 hours a week. Like that'd be extremely rare to see that at a company.
59:18 Speaker 01 I think it depends largely on the growth path. You know, like, why was I working? I mean, I think I was easily working over 100 hours a week. Most of the time I was at Facebook.
59:28 Auren Hoffman? That's why Facebook was also successful.
59:30 Speaker 01 Yeah, but I think it pulls it out of you, right? Like you look at OpenAI right now and I guarantee you that everyone over there is like working like this.
59:38 Auren Hoffman? I think some people are, they have too many people. So I really doubt they are, I really do. Cause some of those people are just very academic and they're kind of spending time thinking about things. I know enough people at OpenAI, I don't think they're working that many hours, but there are definitely some people working 80, 90 hours a week there. And obviously that's a very special company, but it's so big, it's just too hard to track at this point.
59:58 Speaker 01 That's kind of why it's hard question to think about because you know, my experiences of you know, building a insanely big social network and then a pretty big social network and anytime that you're dealing with a high growth scenario, it sort of forces you to work. Your number of customers is multiplying so fast, right? You go from tens of thousands of customers to hundreds of thousands to millions to tens of millions to hundreds of millions in like years. And the feedback coming in is from so many users. And so it just creates this crazy cycle, but different businesses that aren't doing that.
1:00:31 Auren Hoffman? I mean, if the alternative to working is watching a movie on Netflix, what's the alternative for many people, right? Playing video games, like what are people doing when they're not working?
1:00:40 Speaker 01 TikTok, Instagram and Fortnite.
1:00:42 Auren Hoffman? Yeah, exactly. Like that's what they're doing when they're not working. So just like, what's the better alternative? For some people, the better alternative is Instagram and Fortnite. And some people, the alternative is work. Like, it's not like often they're like using that time to build deep and meaningful relationships. Some people are, but a lot of them are just like playing video games.
1:01:00 Speaker 01 Yeah.
1:01:00 Auren Hoffman? This has been amazing. Thank you, Dave Morn for joining us on World of DAS. I follow you at Dave Morn on X, which is the only social media you're on. I definitely encourage our listeners to engage with you there. This has been a ton of fun. I knew this would be great.
1:01:13 Speaker 01 Yeah, thank you so much for having me. Always enjoy talking with you and it just means a lot.
1:01:21 Auren Hoffman? Thanks for listening. If you haven't already done so, please subscribe to World of DAS wherever you consume your podcasts, YouTube, Apple, Spotify and more. And please help us get discovered by leaving a review and check out worldofdas.com. That's worldofdas, D-A-A-S dot com. And of course, connect with me on Twitter, at Auren, that's A-U-R-E-N. Would absolutely love to hear from you.
Previous The LM Brief: Why Many AI Projects Fail Next Patrick McGee - author of Apple in China – how Apple and China built each other Page updated October 31, 2026 · transcript and summary are machine-generated