0:00:03 All right, we should talk about AI, because you used to tweet about like boring, steady,
0:00:08 any cash flow businesses. And I feel like for this one, you just ripped off your clothes
0:00:11 and you’re streaking through the pool right now. So what is your take on AI?
0:00:28 Dude, I am for the first time in a while, I’m waking up greening every single morning,
0:00:32 just like stoked to get to work. I was saying to a friend, I was like, this is like somebody
0:00:37 just invented a new internet or something. Like that’s how big this feels. And he goes,
0:00:43 no, no, no, no, someone just invented fire. Right. Like it’s, it’s freaking crazy. And the
0:00:49 best quote that I’ve heard on this is it’s like we’ve discovered a new continent with 10 billion
0:00:56 people on it and they’re all geniuses and willing to work for free. Right. And so everyone’s talking
0:01:00 about this, you know, agents, right? Agents is like a meme right now. And what we’re really
0:01:05 talking about is digital employees and digital people. And when you shift your mindset like that,
0:01:11 you go, holy crap, like not only is this an incredibly exciting time to be an entrepreneur,
0:01:16 but the implications are going to be absolutely crazy. Like, you know, that’s saying there’s
0:01:22 decades, what is it? What is it? There’s nothing happens. And then there’s days where decades
0:01:28 happen. And I feel like right now, decades are happening, you know, like crazy. And here’s,
0:01:34 here’s the way I’d put it. So if we just paused AI, we shut down all the AI labs and all we focused
0:01:42 on was just roll out of what already exists. I think probably between self-driving cars and AI
0:01:49 agents, 20% of all current jobs are gone. What can you explain examples of the way that you’re
0:01:54 using it now to save and make yourself more productive, both personally and in a business
0:01:59 level? Or is this mostly like theoretical where you’re like, it’s almost there? Well, it’s a bit
0:02:05 of everything. So I’d say like, we’re in the AOL dial-up phase right now. But like AOL dial-up,
0:02:09 it’s still pretty freaking cool if you’ve never used the internet, you know, basic stuff that
0:02:14 everybody should be doing. Like I have an agent that preps me for every meeting. So 30 minutes
0:02:20 before every meeting, I get a text and it says, “Hey, you’re meeting Sean and Sam. You were emailing
0:02:24 them about this topic. You’re going to invest in their company.” And here’s two bios on them. It
0:02:30 goes to LinkedIn and you know, just summarizes it. So something very basic that otherwise a human
0:02:35 would do. Previously, my assistant would do that. I have it monitor my stock portfolio. I have it
0:02:40 tell me about local events. For the first one, so how did you set that up? How did you, what tool
0:02:43 are you using to do the meeting prep? 30 minutes before you get a text saying, “Hey, you have this
0:02:49 meeting with these people about this.” So you guys know Zapier. It’s basically Zapier for AI agents.
0:02:56 So Zapier is freaking amazing. Andrew, you introduced me to someone who worked at Lindy
0:03:03 and her name was Lindy. And I was like, wait, you, I don’t understand this. We got 25 minutes into
0:03:07 the conversation before I realized it was just a woman named Lindy who worked at this company
0:03:11 called Lindy. Basically, they founded the company and they were like, oh, like your name is an awesome
0:03:16 name for a company. Let’s just call it that. But it’s also super confusing because she runs the
0:03:24 sales team. Yeah, I was like, wait, so you’re, are you a smarter child? Most people, it’d be like a
0:03:29 quick two second thing. It’s like, oh, Lindy, yeah, yeah, that’s cool. How funny. And then they move
0:03:35 on and get into the sales demo, whereas Sam for 25 minutes is like, hold on, are you lending or
0:03:43 this is like, would not let it go for 20 minutes. All right. So when I ran my company, the hustle,
0:03:47 I think we had something like 2 million subscribers and we made money through advertising. We didn’t
0:03:52 actually make that much money per person reading the newsletter because advertising in general is
0:03:56 kind of a crappy business model. And so I remember sitting down and I’m like, what are all the
0:04:01 different ways that I can make money off the hustle that aren’t advertising? And so to make
0:04:05 sure that you don’t make this mistake, Sean, me and the HubSpot team, we went and looked at a
0:04:12 bunch of different ways to monetize your business. And we put it all together in a really cool document
0:04:17 where we lay it all out along with our research. And we call it very appropriately, we call it the
0:04:22 business monetization playbook. Go to the description of this episode and you’re going to see a link
0:04:26 to that business monetization playbook. It’s completely free. You just click the link and you
0:04:34 can see it back to the episode. So yeah, like, so think about, think about Lindy like Zapier,
0:04:40 except you can build AI in between. So Zapier might be, when I press this button on my phone,
0:04:45 go to the internet and do this act action or whatever, but I can’t think Lindy can add thinking
0:04:50 in between. So for example, you can make a Lindy that looks at your calendar. So the way it would
0:04:57 work in this instance is look at my calendar an hour before any events starts, go on perplexity,
0:05:02 go to LinkedIn, and I want you to write me a bio and then I want you to look at my email inbox
0:05:06 and tell me what we’re talking about. So that’s like pretty basic. And what we’re really talking
0:05:14 about at this point is removing admin work, right? So the agent that I have built is basically
0:05:18 by cobbling it together with duct tape and dental floss. It’s basically doing a lot of
0:05:24 stuff. My assistant would otherwise do. And my assistant’s amazing, but the problem with a human
0:05:27 is they, you’ll give them direction and you’ll say, I want you to send me these meeting
0:05:32 briefings or I want you to write my calendar like this. And over time, there’s drift,
0:05:36 they get busy, you give them more projects, and it just doesn’t happen. And so now all those things
0:05:42 that let’s be real are not really worthy of her time are now just being totally automated. So
0:05:47 that’s like the basic level. And I can name some of the other tools I’m using to do this,
0:05:52 but that, that’s where I’m at. So I use Lindy to build these things when it’s something I want
0:05:56 that’s custom. Are you an investor in Lindy? No, I’m not an investor in Lindy. I’m just like,
0:06:02 I’d power user and I’m obsessed with it. Like here’s an example. So there’s some restaurants in,
0:06:07 in town where they still require you to call in to make a reservation. They’re not on open table.
0:06:14 I made a Lindy bot where I can type in, Hey, make me a reso around five to seven PM tonight,
0:06:19 this restaurant, and it will call and an AI voice will talk to them, book the reso,
0:06:24 put it on my calendar and invite the people I asked it to invite. And if the person says,
0:06:29 are you AI, it’ll explain that it, yeah, it works for me. And if it asks for allergies,
0:06:34 it knows my allergies, like it’s pretty cool. And then other tools. So I use this one called
0:06:40 howie.ai. And it’s basically like a, you CC it and you just say, howie, can you find time for
0:06:49 me and Sam to go get a coffee? And it’ll, you know, book all that stuff. I use fixer, F Y X E R.
0:06:54 It’s this British startup. It basically reads every single email that comes into your Gmail.
0:07:00 It sorts it based on whether you need to respond to it or not. And then it drafts responses. So
0:07:07 let’s say Sean said, Hey, do you want to get coffee? It’ll draft a rejection or an acceptance of that
0:07:13 in my language. And it gets better over time. So those, those are like three tools that I’m using
0:07:18 that are awesome right now. And they’re really just replacing basic admin.
0:07:23 This is crazy. I’m looking at this right now. Sean, where’s this land with you? Do you use any
0:07:28 of this stuff? And does this make you fearful of somewhere? No, not really. I think everything
0:07:32 just evolves, right? Like ultimately, I’ve, you know, this diagram that’s always stuck with me is
0:07:37 like a K shaped future. So what’s a K shaped future? So if you imagine the letter K and you
0:07:41 say, All right, this is where we are now. That’s the first line. And then it’s basically like,
0:07:46 there’s a fork in the road with the K, right? If you learn how to use these tools, you’re
0:07:50 going to accelerate and go up at a way faster pace than you would have otherwise. And if you
0:07:54 don’t, if you just stick your head in the sand and pretend that these tools don’t exist and that
0:07:59 your job is always safe and that your business is going to run with all humans doing all jobs,
0:08:02 then you’re going to go down because you’re going to get out competed by companies that are going to
0:08:08 do something different. I personally use it less for my personal and more for my business. So
0:08:13 basically in my business, I have must have like a my own eval. So Sam, you know what an eval is?
0:08:18 No, it’s just like, when a new model comes out, like when llama drops their new model,
0:08:23 they always post with pride, like, Oh, here’s how we do on the benchmarking test, the evaluation
0:08:28 test. And it’s like some number, you don’t really know what it’s based off of. It’s just this like
0:08:34 abstract like, Oh, we’re a 73. And the last model from GPT was a 71. So we’re the best model now.
0:08:38 And is there a standard like as an eval, like a third party standard?
0:08:44 I have my own little eval, which is in my businesses, I basically picked a few tasks that
0:08:50 I’m like today humans in my companies do these tasks. And about every sort of three to six months,
0:08:54 we go back and we run through it again, we say take the latest model, take the latest product
0:08:59 everyone’s excited about and see like, what’s the real deal? Holy field. Can we actually automate
0:09:04 that task? Can it actually be offloaded from the human being yet? And there’s basically like a
0:09:09 spectrum. There’s like, can’t do it at all. There’s, you can do it, but you wouldn’t actually do it.
0:09:14 It’s not good enough. Like the person, you know, that, Oh, you wanted to use AI photography
0:09:18 for your brand, but every model has nine fingers, you know, two mouths. It’s like, All right,
0:09:23 well, I’m not actually going to put that on my website. So like cool in theory, but not in practice.
0:09:27 Then there’s, you could do it if there’s a human on top of it. So you could train a human to like
0:09:32 be in the loop to tweak it, to work with it, to make it better. And then there’s a, we don’t,
0:09:36 we just don’t need a human doing this anymore. And so what I’ve done is basically figured out,
0:09:41 all right, here’s like 10 tasks that are real tasks that we do that have a real cost and a real
0:09:46 human doing it in my companies. And I will know for myself, as we get closer to AGI, because those
0:09:52 I’ll just go from zero out of 10, which is where we started to like three out of 10, six out of 10,
0:09:55 eight out of 10, once eight out of those 10 tasks are like, yep, we just set up an agent
0:10:00 and like, we don’t need to have a person doing those anymore. That’s when I know, oh my God,
0:10:04 this is like, we’ve had like a real, real breakthrough is because it matters because
0:10:09 that’s the real test for me versus, I don’t know exactly how these evals work, but it’s like a,
0:10:13 it’s like an abstract thing, or it’ll be like, you know, they can pass the LSAT. So it’s like,
0:10:18 that’s cool, but I don’t need it to pass the LSAT. So if the opportunity is AI agents in my business,
0:10:22 then I’m going to test it on what can these AI agents do in my business. So what’s the example
0:10:26 of one of these things that you would want to automate? The simple example I gave was like
0:10:32 e-commerce, so photography. So every month we pay photographers to do shoots, real costs somewhere
0:10:38 between let’s say $5,000 to $10,000 all in for e-commerce photography. And you have basically
0:10:41 product photography, or you can have models in it, whatever, right? So there’s like a little
0:10:47 bit of a range there. So just product or models. And so I want to know like, when can I get rid of
0:10:53 that $10,000 a month? So when can I basically just show it my product on a blank and be like, cool,
0:10:57 put this on models? Or we have other ones where it’s like photos and videos. So like, you know,
0:11:01 you’ve seen these models where you give it a static image and it turns it into a video.
0:11:06 And you see the demo on Twitter and it’s amazing. It’s just a girl posing. And then she starts
0:11:10 like strutting with it. It turns into an AI video and it looks awesome. Cool. But then like,
0:11:15 you do it in practice. And yes, she starts strutting, but then her face starts morphing,
0:11:18 like she’s like being eaten by lava. It’s like, okay, I’m not going to put that on my website.
0:11:24 Like that would be a bad thing, right? To do that. Did you see the AI meme of Mark Zuckerberg
0:11:29 staring at Bezos’ wife’s boobs? He like leads it extra hard to like smell them.
0:11:36 Do you watch that video? It’s literally his eyes. They dart bat down from microsecond.
0:11:42 Yeah. And some guys made videos of him like not stopping. It’s just continually going in. It was
0:11:48 that’s pretty badass. What are we saying? Like another example would be in e-commerce,
0:11:52 you have inventory forecasting. This seems like a data problem, right? How much inventory should
0:11:56 I order? When should I order? And let’s say, let’s pretend you have five colors of a product,
0:11:59 which color should I be ordering? Well, it looks like an AI problem. You got a whole bunch of
0:12:04 historical data. You have real time sales data that’s coming in and then you should be able to
0:12:09 forecast. And like today we employ two human beings to do that forecasting. They’re not
0:12:14 perfect at it. It’s very difficult problem to solve. Can AI make them smarter and do it better?
0:12:18 Can I go from two humans to one human plus AI? Or can I replace the two humans that just have AI
0:12:22 give us more accurate demand forecasting? And this is a problem that every physical product brand in
0:12:27 the world has this problem. Whether it’s Hershey’s chocolate bars or it’s North Face jackets. Like
0:12:30 it doesn’t matter. Everybody’s trying to figure this out. How do we order the right amount of
0:12:34 inventory and not be stuck with too much inventory? It feels like we’re talking about solar power.
0:12:38 You remember like five, 10 years ago, everyone would be like, well, I mean,
0:12:44 solar is really cool and all. But I mean, like it’s still not cost competitive. But if you look
0:12:48 at the graph, it just keeps going. Cost goes down, down, down, down, down. And if you just
0:12:53 project it and you look at the rate of progress with these tools, like we’ve been exploring for
0:12:58 Aeropress doing the same thing. Cause we spent a lot of money on big expensive photo shoots. And
0:13:03 usually it’s just an Aeropress sitting there really well lit with nice coffee in it. So we’ve
0:13:08 been looking at Flair and some of these other tools. And I’ve been pretty blown away with where
0:13:14 I can get to. And I agree there’s definitely like a last mile problem, but it’ll be there in the
0:13:21 next six to 12 months, I think. And when it does, I think it, it’s really cool for your business,
0:13:27 but it also represents competition. Cause it just becomes infinitely easier to compete in all
0:13:31 businesses. And we’ll talk about this more, but like, I think that’s what’s crazy about all this
0:13:35 stuff. By the way, the most solved one of these is meeting, meeting minutes. Like, I don’t know,
0:13:38 Andrew, if you used to this, but I used to invite my assistant into our meetings and say,
0:13:42 Hey, can you take notes, jot down any action items or followups that we’re going to need to do?
0:13:46 That’s completely gone off her plate, which for her, it basically freed up, you know,
0:13:51 four hours of time for my assistant a day on average, let’s say. And if now she’s got four
0:13:56 hours back that she can spend doing the other things, because I’m using Fathom and Fathom just
0:14:00 sits in my meeting, records all the notes, perfect transcript with highlights, with a summary,
0:14:06 with action items ready to go 24 seven, never misses a beat. And not only did it replace him,
0:14:10 it’s way better than my assistant could do it. Cause she’d be trying to keep up and it’s just
0:14:14 was not going to be good enough. The level of accountability you can get to, right? Like,
0:14:18 I, I hate that feeling where it’s like, Oh, I did a board meeting with the CEO six months ago,
0:14:22 and I swear to God, I swear to God, they promised me this one thing. And then they’re like, no,
0:14:28 no, I don’t remember that. And being able to pull that up, I just did a really, a really cool thing.
0:14:33 Actually, me and my girlfriend did a relationship offsite, which sounds really lame and nerdy.
0:14:38 Patrick Campbell actually was the one that gave me the idea, but we basically would talk about,
0:14:43 Hey, what are, what are like our big strategic goals for the year, for our relationship, for
0:14:48 family, for life in general and stuff. And we recorded the entire thing with Otter, which is
0:14:54 like similar tool to Fathom. And I was blown away. Like we just chatted for three hours and it was
0:14:58 able to spit out, here’s all the action items. Here’s all the things that need to come out of
0:15:03 this. Then we’re able to put it into Claude and it built a plan for us. Very cool.
0:15:09 By the way, I was giving you a hard time. It does sound nerdy. And also I do the exact same thing.
0:15:13 We have relationship meetings as well. You, you have the local news business,
0:15:21 the newsletter. Sean and I both owned newsletters. When I did it, AI didn’t exist. And I felt like,
0:15:26 you know, hiring writers was actually the hardest part. Like who could write in my voice.
0:15:30 Now I would hate to be in the newsletter business because I see this stuff and I’m like,
0:15:35 it’s almost ready to roll, but it’s definitely ready to roll as an editor.
0:15:39 Are you using this for your newsletter now for the local news newsletter?
0:15:44 I’m not right now, but that’s what I’m working on. So like the pieces that I’m,
0:15:49 the more advanced stuff I’m doing are like, I’ve got a social media manager agent. I’ve
0:15:55 got a marketer agent. I’ve got a tax advisor and investment analyst. And I mean, these are,
0:15:59 these are roles I previously would have hired for where I didn’t, yeah,
0:16:03 either didn’t renew someone’s contract or I didn’t make a new hire for them. So it’s already
0:16:08 replacing people that I would have hired otherwise. Can you explain a couple of these?
0:16:12 Explain the tax advisor one and then explain the social media one.
0:16:18 So social media is very simple. It’s like, I post on X all the time and I’m lazy. I don’t
0:16:23 like cross posting on LinkedIn. I often forget. I’ll tweet about something and then not include it
0:16:28 in my newsletter. And so what I’ve done is I built a Lindy bot, a Lindy bot that looks at my Twitter,
0:16:34 all my other channels, and then once a week it summarizes all the stuff I’ve been talking about
0:16:39 and asks me, Hey, what do you want to include in the newsletter? And then it can give me a
0:16:45 first draft and it’s trained on my previous writing. So it can write, and it’s, you know,
0:16:50 it’s not perfect, but it’s like 80% of the way there. And if I didn’t care about my personal
0:16:54 brand, like, you know, many people just want like a quick and dirty newsletter, it would be totally
0:16:59 sendable. And it’s completely engaging if you prompt it, right? Like you can even say, like,
0:17:05 give me a subject line that will force the person to open the email, right? And it’s really good.
0:17:07 What about the accounting thing?
0:17:14 The accounting thing is freaking like blew my mind. So basically I exported,
0:17:20 I use zero for all my accounting. And for my personal, like financials and my personal holding
0:17:26 company, it’s like a rat’s nest. Like I have like 200, 300 random investments, all sorts of weird
0:17:34 assets and stuff in there. And so I exported all, all of it as a CSV out of zero. I trained
0:17:39 a Claude project on it and I started asking it questions. Did you upload like a book,
0:17:45 like a Warren Buffett book or like what did you? No, no, this is like basically to replace like
0:17:50 right now. So I have a woman who runs my family office and I might ask her, Hey,
0:17:54 how much should we spend on software last week? It’ll be, let’s say it’s a Saturday and I’m kind
0:17:58 of bored and doing a deep dive. I’ll email her on Saturday and I’ve got to wait until Tuesday
0:18:04 to get an answer. And so I was like, Oh, I just want to be able to query Claude. And so I was like,
0:18:11 Hey, I want you to rank the top 25 things I spend money on, or I want you to look at my structure
0:18:16 and I want you to tell me if there’s any ways that I could save on tax. And so one, it was crazy.
0:18:23 I actually saved a hundred thousand dollars a year of cash taxes by moving an investment from one
0:18:30 holdco to another. Thanks to Claude, right? So it is very quickly becoming tax advisor,
0:18:35 investment analyst. I can get it to write a deal. I can get it to draft legal documents. Like
0:18:41 these, I mean, everyone’s doing this, but like it is getting, it’s getting so good that I don’t
0:18:48 hire people. Let’s put it that way. Weren’t you hiring for like an AI back office person? Like,
0:18:52 you have some role I saw that you put up there. What was that? Yes. So the woman who runs my family
0:18:59 office, she’s got a family situation. So she’s got a transition out of the role. And so me and
0:19:03 my president were like, okay, what do we, what do we really want out of this role? And when we
0:19:09 thought about it, we were like, we actually just want someone who’s like an AI nerd to automate
0:19:14 away all the accounting and bookkeeping and stuff that nobody enjoys doing. You would go, came up
0:19:20 with that? Yeah, that’s right. Exactly. I hired him right out of the office. But anyway, we, yeah,
0:19:26 we’re hiring like an AI first CFO, basically somebody where their job is basically just to
0:19:29 build automations, financial automations. You guys probably have someone who works for you,
0:19:35 who just basically manages Zapier. And like Zapier, the problem with Zapier is that it’s
0:19:39 awesome when it works. But like for some reason, shit breaks all the time, which I don’t understand
0:19:42 how like, I don’t understand enough about technology to know like, why does this shit break all the
0:19:48 time? And then the systems were so complicated that if that person left, it was like, oh,
0:19:53 fuck, what’s the stuff that we got to go uncover because they’ve automated so much shit, where it’s
0:19:58 like, we just rely solely on this one person. And it becomes actually an issue where it’s kind
0:20:01 of all automated, but it’s sort of duct tape together and the duct tape comes on done and
0:20:05 there’s so much duct tape that I got to go and pick it’s like when your Lego breaks, you got to
0:20:08 like, dude, I got to like fucking rebuild this whole shit. I don’t know where this one little
0:20:13 fucking piece goes. Totally. I’ve had the same experience like anytime I got really into no
0:20:18 code like four years ago or whatever. And you know, we build all these automations where,
0:20:23 oh, it’ll go to this one place and then a Zapier automation will move the data into air table.
0:20:27 And then you manipulate it in XYZ way. And I think the difference here is that like
0:20:34 Claude will just figure it out, right? So if you say like, hey, it’s not working. It’ll just sort
0:20:39 it out or tell you how to fix it. Whereas before I had that exact same problem and totally agree,
0:20:44 usually the automations just cause problems. How are you guys using it at Hampton right now?
0:20:50 I mean, it’s basically just like, it’s like glorified Zapier basically. So like it connects
0:20:55 air table to HubSpot, to this website, update these records in HubSpot. Like it’s basically all
0:21:00 backend operational stuff, which I think is underutilizing it. Cause like what I would do,
0:21:04 if I was you, is I would get a, um, Claude has this thing called, I forget what it’s called,
0:21:09 model context protocol or whatever. And you can basically build a database of all, all your members
0:21:14 and then just have a way to query it where you can say, Hey, we need a speaker for this event
0:21:19 in LA who would be good or who’s got a business in this space and just have like a live database
0:21:27 of people in an LLM. That’d be sweet. So, so what I’m working on though, like the next piece here
0:21:33 is actually like Sean, I’m basically lining up who are, you know, what are the roles that we can,
0:21:38 you know, not need in the future. And most of this is not firing people. It’s actually just
0:21:43 not hiring people we would otherwise would. But if you graph this out, you can see that
0:21:48 we’re all going to have virtual employees very, very soon. And that’s what’s crazy. Like I was
0:21:56 watching an interview with Dario Amade from, uh, anthropic yesterday at Davos and he basically
0:22:04 said in the next 12 months, you will have AI coworkers in Slack. Now, if you just project
0:22:08 that out right now, it’s text, right? These are not people, but pretty quick, it’ll be someone
0:22:13 on Slack that you’re messaging. Hey, can you make some new product images? Hey, do you mind
0:22:20 checking out the AdSense? Do you, can you write a report? And I think that within 24 to 36 months,
0:22:26 these will be 4K video people that are indistinguishable from a human that you’ll talk to on
0:22:31 Zoom. Now, I know that sounds crazy, but I really believe that’s what’s coming. And I think if you
0:22:37 just think about the, the implications in terms of business motes and businesses that are gonna,
0:22:43 gonna be disrupted, it’s just staggering. All right, let’s take a quick break because I gotta
0:22:47 tell you about a friend of the pod who’s got their own podcast. If you know Steph Smith,
0:22:52 she is a legend. She’s been on MFM many times and she’s got her own podcast called the A16C podcast.
0:22:57 And it’s all about technology. If you think about it, technology has evolved like crazy. I mean,
0:23:02 I grew up in the 90s. I had CDs, phones had cords. You couldn’t use the internet if your mom was on
0:23:08 the phone. And now there’s like 3D printers and there’s rockets that can go up into space, AI.
0:23:12 There’s so much crazy stuff going on. And you got to have a place that helps you stay ahead of the
0:23:18 curve. And that’s what the A16C podcast is trying to do. It’s a podcast from the VC firm, Andreson
0:23:21 Horowitz. And it’s trying to give you an inside look at the trends that are shaping our future.
0:23:25 They’ve had guests like Mark Cuban and Neil Stevenson on, and they talk about topics like
0:23:30 deep fakes or the science behind GLP ones or autonomous drones. No small boy stuff at all.
0:23:34 Steph is the host. She’s awesome. I think you’ll enjoy the podcast. So check it out.
0:23:38 It is the A16CZ podcast and I like this tagline to say it’s like eavesdropping on the future.
0:23:43 That’s pretty cool. That’s a good tagline. So check it out. The A16Z podcast, wherever you get
0:23:51 your podcast. Do your guys’ parents use it or like your 10-year-old cousins? Like how, how…
0:23:57 My mom uses it. In what way? Well, my mom’s, you know, her first language was Hindi, right?
0:24:02 So she speaks English. She speaks it well, but she, if she needs to write, it kind of stresses her
0:24:05 out. She’s like, oh, I got to write something. Got to write an email. I got to write a letter,
0:24:10 respond to something, write a note to the doctor. And she’s like kind of slow with writing and she
0:24:15 feels a little bit insecure that like her writing is going to be like well-formatted, well-granted,
0:24:21 you know, grammar’s correct and clear. And so she was early on chat GPT to just be like,
0:24:25 oh, great, I can just tell it what I wanted to write. And then it just writes a beautiful email
0:24:31 for me. And then what she does is she has to like kind of mess it up a little bit to get it back to
0:24:36 real. But my mom has been using it in that way. But that’s kind of all she used it. But I think
0:24:40 what you’re getting at is like, even though for us, a lot of these things we’re saying right now seem
0:24:43 obvious, you’re kind of right. I think if I know where you’re going with this question, which is like,
0:24:50 you know, my sister, my mom, like how much does it cross the chasm? Like my sister uses it in her
0:24:55 schools to create curriculums. So she has it do curriculums and lesson plans and gives that to
0:25:02 her teachers to to work off of, which is something like either they would do kind of either procrastinated
0:25:07 or, you know, half or she would have to do herself. And so I think that was a big win for her. But,
0:25:12 you know, it’s, it’s kind of like you find your one use case of chat GPT or, you know, you go there.
0:25:17 Another one, like my mom was giving it like her Kaiser like data test results. So like, you know,
0:25:22 you go, you get bloodwork done or you get a, you get some procedure done, you get this result in
0:25:26 your app, and you have to wait for this doctor to call you to tell you what the heck this means.
0:25:31 It’s kind of a scary gap of time. And we just upload all of the screenshots into chat GPT,
0:25:37 we’re like, what does this mean? And it gives us beautiful, perfect bedside manner explanations,
0:25:41 more patient, more thorough than the doctor, you can ask follow up questions at your own leisure.
0:25:45 I mean, that’s been another one that was like pretty, it was just like night and day. It’s like,
0:25:49 okay, I’ll never not do this anymore. This is so much better. I’ll never not do this.
0:25:54 Andrew, you had on here that software is going to be a worse business over the next 10 years,
0:25:56 what’s your like premise here of it being a worse business?
0:26:03 Let me zoom out a bit. So in a free market, the amount of competition basically defines
0:26:08 how good a business is, right? So, you know, everybody loves to shit on restaurants,
0:26:12 but if you own the only restaurant in town, let’s say that there was one restaurant license
0:26:17 given out in every major city, and you own that restaurant. So you have access to man,
0:26:22 you have no competition. That would be a phenomenal business. You could have 40, 50,
0:26:26 60% margins. Basically, as long as people don’t riot, you can charge whatever you want,
0:26:34 and the food can be bad, right? So I think that same thing applies to software businesses and
0:26:40 startups, right? So if I remember, you know, 15 years ago, I started a productivity software
0:26:46 company, and it was a terrible business because every single week, new VC back companies like
0:26:52 Asana, Monday.com, et cetera, would come out and compete with us and drive us into the ground.
0:26:59 And where there’s been a lot of wealth generated in software is in vertical markets. So what does
0:27:05 that mean? Like that’s like weird niche businesses where nobody is competing. So for example,
0:27:12 like funeral home management software, golf course management software, random like dam
0:27:18 and infrastructure software. And these are great businesses because no like amazing developer
0:27:23 wakes up and no amazing designer wakes up and is like, I’m going to go and build this software.
0:27:27 And so basically you’re the only game in town, you can charge whatever you want,
0:27:33 and nobody switches off of you. And so for the last 30 years, that’s been the best business ever.
0:27:38 You might have heard of constellation software. They built like a $70 billion behemoth public
0:27:43 company. And all they do is buy businesses like this. Don’t they buy like a hundred a year?
0:27:47 They buy a hundred a year. They increase the prices on a schedule. They hold them. They don’t
0:27:53 really, you know, do anything crazy or improve them in any particular way. And they’re buying like
0:27:59 $3 million a year software companies, three, three to 10, sometimes bigger, but they have a formula.
0:28:05 And so I believe let’s, let’s just take funeral home software, right? So let’s say that there’s
0:28:11 one developer who built funeral home software and he’s had no major competition for 10 years
0:28:18 with tools like replete, you can literally go in like, and you can reproduce basic software very,
0:28:24 very easily, right? If your software is like highly algorithmic and complicated, yeah,
0:28:29 maybe not, but something basic like funeral home management where it’s like case management
0:28:36 basically is incredibly easy to replicate in 10 or 20 minutes and replete. And so my belief is
0:28:42 basically that as more and more people don’t need to learn how to code and can build software,
0:28:47 software will become more and more commoditized. There’ll be way more competition and it’ll be
0:28:54 a way harder world for software businesses. And I think that ultimately it just comes down to time.
0:28:59 For us, all of our software businesses we’ve bought, we’ve generally bought them in a way where
0:29:05 we’ve de-risked, we’ve paid ourselves back, or we have some sort of unique competitive advantage
0:29:14 or cost advantage or something. Like for example, we own a framework, a hosting company for a framework
0:29:18 and a lot of people have standardized and built software on that framework. They’re going to
0:29:22 keep hosting with us because we support that framework and nobody else does, right? There’s,
0:29:27 there’s things like that. Personally, when I’m underwriting software deals now,
0:29:33 I am being incredibly conservative and honestly, I’m much more focused on communities and social
0:29:39 networks as a result. We still look at software when there’s like a hardware component or there’s
0:29:44 some kind of lock-in, but I’m much less enamored with software right now because of this. Do you
0:29:50 guys think I’m like out to lunch on this? Am I totally overthinking this or say the thought
0:29:56 out loud? What’s the specific thought, right? Well, I think it’s pretty wild implication that
0:30:03 we may have mass disruption in terms of jobs and mass disruption in terms of every business that we
0:30:10 know that we think of as a good business may not be a good business in, and it’s not a long term,
0:30:13 right? This is not 10 years. We’re talking about a two-year time horizon.
0:30:18 So let’s take the business side of it, right? The business side of, your argument was basically
0:30:25 software for the last 30 years has been an incredible business to be a creator of or an owner
0:30:29 of whether you’re buying or building software businesses, and you’re saying, I think that’s
0:30:36 going to change because if it becomes so easy for anybody to build software that they need,
0:30:41 then they’re not going to be buying $5,000 a year or $50,000 a year software anymore. And so maybe
0:30:48 there’s one thing, vertical software, VMS, right? Or other categories like that where it’s like
0:30:53 more of a simple CRUD database with a user interface on top of it, that those are going to
0:30:57 become less valuable. I totally agree with you that those are less appealing than they were
0:31:01 before. I don’t think it goes to zero because in reality, it’s sort of like anytime you’ve
0:31:06 actually built software, it’s the last 10% that takes, you know, you can get 90% of the way there,
0:31:11 you feel like you’re almost there, but then the last 10% will take you as much time as the first 90.
0:31:15 It’s just like a rule you see over and over and over again. And it’s the same thing with products,
0:31:20 which is that to actually make something work and solve all your problems that you need,
0:31:26 that you’re used to your current software solving, it will take time, effort, love, care,
0:31:31 and whether it’s just might just be somebody else building software that’s a competitor to you or,
0:31:35 you know, if somebody in your own company going to build that, I think it’s less likely that that
0:31:41 becomes like a major disruption to all of software. Because like you were saying, like some software,
0:31:45 like social networks or communities, the value is not the app, the value is the people and the
0:31:49 connections and the habits that they’ve formed. So you have software where there’s already habits,
0:31:52 those will be safe. Software where the value is the people or the data, those will be safe.
0:31:58 Software where the value is in the enterprise relationship, the service that’s on top of it,
0:32:04 or the like fine tuning and tailoring and the CYA, which is that if I buy enterprise software,
0:32:09 I don’t need to save 40% on this bill as much as I need to not get fired for being like, yeah,
0:32:14 we homebrewed our entire CRM and like the security wasn’t very good, right? Because we didn’t,
0:32:19 I made this in fucking Claude instead of it being software I bought from a vendor that’s
0:32:23 been doing this for 15 years and has actually like solved the security vulnerabilities.
0:32:26 And so I think that enterprise software will be safe. So I think a lot of software will be more
0:32:32 safe than software just as a category goes to zero. Yeah, I would, I would generally agree with
0:32:38 that, although I think it’s going to be a lot easier to offer incredible sport, incredible sales,
0:32:43 etc. On the other side of it, you get the, the market expands, right? Because it’s not just that,
0:32:48 oh, there’s going to be more competition in this fixed pie. It’s that now a whole bunch of software
0:32:53 that really wouldn’t have made sense to build because coders didn’t understand the problem,
0:32:57 so they weren’t working on it, or you would have need too much funding to build it. It’s not really
0:33:02 worth it. So all this new software is going to get created that was not really feasible or wasn’t
0:33:06 really easily accessible before, like I’ll give you like a stupid example. I went on Replet the other
0:33:10 day and Replet’s amazing because it’s literally chat GPT, except for when you tell it something,
0:33:15 it makes a website for you or makes an app for you. And so I had this like, you know, first
0:33:20 world selfish problem, which was that I like to log my meals. I take photos of my meals when I,
0:33:27 when I eat them to be able to send to my coach, but then my entire camera roll is just shitty pieces
0:33:30 of chicken and my kids, like beautiful pictures of my kids and then ugly pictures of chicken.
0:33:34 And I’m like, all right, this is kind of annoying that that’s there. So I just told Replet, I said,
0:33:38 make an app that when I open up the app, instantaneously, the camera is on. I don’t want to
0:33:44 like type in stuff. I just like open the camera, take the photo, show me a thumbs up or a thumbs
0:33:49 down where I can rate my own meal and then save it in a calendar view where I can see like a streak
0:33:52 of, you know, if I’ve been doing roughly good, make it green. If I’ve been doing bad, make those
0:33:57 red and couldn’t make it an app that it gets approved in the app store. So that was so immediately
0:34:02 Replet’s like, Hey, I can’t make iOS apps yet, but I can make it as a website. You could add it to
0:34:05 your home screen and you could use it on your phone, but it’ll just be a website you open up.
0:34:09 I was like, all right, deal. And then it made that up for me. And it’s like, you know, it’s not
0:34:14 perfect. Like I ran into some walls where that Replet’s like, Hey, I don’t know why it’s not
0:34:17 working. I’m like, yo, it actually is working. And it’s like, you’re, you’re just being crazy.
0:34:20 But like, you know, okay, you get through the craziness and it’s not exactly good enough. I
0:34:25 would say like, I, I’m using it every day and telling you guys how amazing this, this tool is.
0:34:28 But nine months from now, pretty sure I’m going to be able to just do that. Right. So like,
0:34:33 I can be a software creator without knowing how to code. That’s already a big, big deal
0:34:38 because of that. Any idea I have, I can make come to fruition. So a whole bunch of new ideas
0:34:43 are going to come into the market that didn’t exist. And I could do it for like other problems
0:34:47 that weren’t going to get solved, maybe through like, you know, if you look at software today,
0:34:52 it’s like a lowest common denominator. Like I need to serve the most customers who are
0:34:57 willing to pay. So I’m going to like shave off all the edges and make some product. That’s
0:35:01 what I think most people want, even if it’s not what I personally want. Right. Or what me
0:35:06 and a cluster of people like me really want. And so I think the overall market for software gets
0:35:10 bigger because you can create way more stuff. So, so that’s the other counterpoint to like
0:35:16 software being bad. It’s phenomenal for consumers. Right. As a consumer, I’m so excited. It’s more
0:35:21 if you’re buying these businesses and you’re underwriting 10 years of returns, continued profit
0:35:28 margins, all that kind of stuff. I think it’s, it just got a lot harder because before it was like,
0:35:33 look, there’s only so many programmers out there. Right. There’s only so many competent Mac OS,
0:35:38 iOS, whatever programmers. And I’m with you in terms of like where a replete is at right now,
0:35:42 where it’s like, yeah, you got to, you got to fiddle a lot to get it to a point where it’s
0:35:47 shippable. I’ve built a lot of stuff that is 90% that I can’t put out into the world because it’s
0:35:52 not there yet. But I think it’s a little bit like stock photos. Two years ago, you know,
0:35:57 stock photos generated by AI were terrible. Now I would not want to be in the stock photo business
0:36:02 because it’s perfect and it’s only getting better. I will say one other thing, like this
0:36:07 same thing happened with media, right? So like before, only people who could make media were
0:36:13 media companies. You would, there was limited TV channels, newspapers. If you owned one of them,
0:36:17 it was amazing. If you didn’t, it wasn’t right. And you’re right that like owning those legacy
0:36:23 media businesses kind of sucks in the world of social media where anybody can create content.
0:36:28 And so those businesses went down, but then suddenly new businesses, which were social
0:36:34 media based, emerged and the set of social media based businesses are way bigger than the set of
0:36:39 legacy media businesses. And if you’re interested in whatever weird niche, you don’t have to rely
0:36:43 on the Wall Street Journal to write about it. Right. So I think it nets out way ahead. I don’t
0:36:46 know if you guys heard Larry Ellison yesterday when he was like going off about AI, like people
0:36:50 like, well, what are you excited about AI? And he like did this little rant, which I don’t know how
0:36:55 well verse Larry Ellison is in like, what’s actually like possible with AI versus like
0:36:58 Masa-san, you know, showing a picture of a golden goose laying eggs and being like,
0:37:02 that’s why I’m investing. But he was like, with AI, here’s what’s going to happen in the next
0:37:08 few years. He was like, you’re going to be able to do early detection of cancer. Then the AI is
0:37:14 going to sequence what cancer you have. It’s going to be able to design a personalized vaccine
0:37:20 for your cancer and cure you of cancer. So why am I investing $100 billion? Because I want that.
0:37:25 And I want 100 other things like that. And it was just like a very compelling like rant he did
0:37:29 outside. Did he deliver it in that type of compelling way? Well, nobody’s me, but yeah,
0:37:35 he did. He did the Larry Ellison version of it, the B plus, but he did do like he was like that,
0:37:40 like he was able to dumb it down. He was basically like, it was like he was shaking somebody and
0:37:44 be like, do you hear what I’m saying? Like, do you know what this is? Like, get this in your little
0:37:49 head. It’s not about an investment for a data center. It’s we’re going to cure cancer with this
0:37:54 shit. Okay. And it’s going to save your like your mom’s life. Like, that’s what we’re going to do
0:37:58 with this thing. Okay. That’s why we’re doing this. Let’s not get caught up in Project Stargate
0:38:03 and politics and the investment amount and the infrastructure. Like, okay, all that’s a means
0:38:07 to that end is kind of the way he was saying it. Dude, this shit’s so exciting. He also,
0:38:11 I think he’s obsessed with living forever. I think he’s one of these guys who’s like
0:38:15 into that and throws it is, you know, the 10th richest guy in the world. So he can kind of throw
0:38:18 his weight behind a lot of that stuff. And I think that’s pretty badass. And that’s probably
0:38:23 a huge motivator. Andrew, can we do one more? You had this great thing in your notes that I don’t,
0:38:29 I want, I want to hear you explain. You said AI hedge. Basically, you were looking for investments,
0:38:33 given that you think, okay, software might not be the best place to be investing.
0:38:37 Where are you investing? And what, what is all the stuff I see here down below,
0:38:42 down below your AI hedge notes? Sure. So I feel like, you know, there’s a bunch of different
0:38:48 scenarios here, including that this scaling slows down and it’s a moderate disruption.
0:38:52 We’ve got a couple of years of some people being unemployed and then they reskill,
0:38:55 they go do different things. Like you said, the economy gets bigger, whatever.
0:39:00 There’s another possibility that there’s what’s called an AI fast takeoff, which is like
0:39:07 Masa’s graph of like the singularity happens and, you know, it’s crazy. And if that happens, it’s like,
0:39:13 does capitalism matter? Does the free market matter? Does the AI basically just become this
0:39:17 like communist overlord that allocates assets and stuff? But I’ve been thinking about, you know,
0:39:22 what do you want to own in a world where that happens? So you can go out and you can try and
0:39:29 buy secondary in Claude or Anthropic or open AI. Those are very expensive. You can buy Nvidia at
0:39:35 30 times earnings or ASML at 30 times earnings, some of these or, or a TSMC, but there’s the
0:39:42 Taiwan risk, all the other stuff. And so I’ve been trying to find stocks that I can buy that are
0:39:46 actually affordable, but capture AI upside. And I found a really interesting one.
0:39:52 I love stuff that’s miscategorized and misunderstood by public market investors. Public
0:39:57 market investors are very simple in the way that they think about companies. They have a series
0:40:03 of boxes and they filter you through the boxes. And there’s a saying you want to look like a duck
0:40:08 and quack like a duck in the public market. They do not like complexity. And so if you say the wrong
0:40:13 keyword, their brain turns off and they just say, I’m not going to invest, right? And so let’s,
0:40:19 let’s, a couple of those examples are cannabis, right? They just put you in a box, biotech,
0:40:24 oh, high risk. And Bitcoin is another one. There’s all these crappy Bitcoin mining companies.
0:40:29 Dude, when we sold the company to HubSpot, one of the reasons we went into the year with millions
0:40:33 and millions of dollars booked of advertising, and we were probably going to do $20 million
0:40:38 that next year. And they were like, yeah, just cancel all those contracts and give the money away,
0:40:41 give the money back. And we’re not going to do advertising anymore. And I was like,
0:40:45 but you could scale this to like a lot to tens of millions or whatever. And they’re like,
0:40:49 yeah, we make billions of dollars a year. And also we would have just have to report that.
0:40:52 And that’s a little too complicated and people will understand it. So
0:40:59 like the decision of them to not do that was worth more than the tens of millions of dollars
0:41:01 of advertising revenue they could have made, which is wild.
0:41:06 Have you ever met one of those like really ADHD young entrepreneurs who’s pitching you on their
0:41:11 business? And they’re like, oh, we do this, but we also do this. And also I’ve got a company that
0:41:17 does this, right? And we all, the older statesmen, we always are like, uh, you know, young Padawan,
0:41:21 just focus and don’t, you know, don’t do that. Or if you’re going to do that, don’t tell me about
0:41:26 all this other stuff because it just confuses me. I feel like public market investors are like that.
0:41:32 So here’s the business I found. It’s called Iris Energy. The ticker is IREN, I-R-E-N.
0:41:38 And basically it’s these two Australian guys who were infrastructure bankers,
0:41:44 and their thesis supposedly was that compute was going to be a big deal in the next 10 to 20 years.
0:41:49 So presciently they went out, they secured a whole bunch of huge data center properties that were
0:41:55 right next to mostly renewable energies. So they’d find a power plant that’s like a hydroelectric
0:41:59 dam or something like that. And they would literally just buy a massive data center right
0:42:04 next to it and build one. So for example, where I live in BC, they have one in Prince George right
0:42:11 next to like a big hydro dam. And currently the best use for these data centers they’ve built so
0:42:16 far has been mining Bitcoin. So they’ve been doing that and they’re not Bitcoin bolts. They’re not
0:42:21 like micro strategy or someone like that. They don’t hold the Bitcoin. They just mine it. They’ve
0:42:26 got their GPUs and whatever, and they sell it immediately. And right now they’re making somewhere
0:42:34 in the range of $500 million of EBITDA annually doing that. Now the stock trades at about a $2.4
0:42:42 billion valuation. So you can buy it actually quite cheaply based on the Bitcoin thesis, right?
0:42:48 And Bitcoin would have to drop by about, I believe I’m guessing based on some assumptions,
0:42:54 but I think it would have to drop about 70% before they get to break even on mining these Bitcoin.
0:43:00 So here’s what’s interesting about this business. So it’s totally misunderstood by the public market
0:43:04 because everybody thinks, oh, they’re Bitcoin bulls. They’re mining Bitcoin. There’s all this
0:43:11 risk or whatever. What’s actually interesting here is they’re going to flip, I hope, or they’ve
0:43:16 kind of publicly communicated to some degree that they’re going to try and flip these data centers
0:43:23 over to compute. So all these huge companies like Amazon and Microsoft and Anthropoc and whatever,
0:43:29 they all need data centers and they need the data centers to be centralized in one place. You can’t
0:43:35 have 10 computers here or 10 GPUs here and 10 GPUs somewhere else. For training, you need them all
0:43:41 to be in a massive supercluster. So the other interesting thing is you can’t just go and build
0:43:45 these superclusters. You can’t go build these data centers anywhere. It takes a long time to get
0:43:50 permitting and approval and it takes a long time to build. And so basically there’s a finite amount
0:43:55 of this stuff to the point where you guys have probably heard like Bill Gurley talking about
0:44:02 like trying to buy nuclear power plants and all sorts of crazy stuff. So basically if they flip
0:44:09 this business over to doing inference and to working with hyperscalers on training,
0:44:14 this could be a massively, a massive boost to the business and it could, you know, 5, 10,
0:44:20 20X, I don’t know, but very significant if they signed some deals with hyperscalers,
0:44:24 which I know that they’re actively working on. They’ve communicated as much. So basically if
0:44:30 you believe in a fast takeoff, it’s a very interesting way to kind of buy something with
0:44:35 relatively low downside and very significant upside instead of buying Nvidia or Microsoft.
0:44:41 Obviously, you sound like you’re Leonardo DiCaprio. Yeah, exactly. Do your own research.
0:44:45 This is not investment advice. I do own the stock, but I think it’s interesting.
0:44:50 How much of the stock did you buy? I only bought like a million dollars, but I look at it as like
0:44:54 a small, it’s a very small position. It’s like buying fire insurance.
0:44:57 Yeah, what is this? A position for ants? Just a million dollars, Andrew?
0:45:02 Well, Andrew, I know that was not financial advice, but I do appreciate the advice that
0:45:06 you just gave me financially. The last public stock we talked about, I think, was Weight Watchers.
0:45:12 And like there was definitely a correlation from when the episode aired and when the stock,
0:45:14 I don’t know if it was causation, but there was definitely correlation.
0:45:18 I wonder if that’s going to happen here. You said the downside is like you have this
0:45:24 Bitcoin mining free cash flow priced very fairly, right? You said like 3X free cash flow,
0:45:28 basically, is the price. And then the upside is if they make this AI switch.
0:45:33 One question. You said they, it sounds like they’re not huge superclusters, so.
0:45:38 Oh no, they are huge. So they have a massive, they have a massive site in Texas that they’re
0:45:43 working on billing. Now, here’s the, probably the reason why it’s trading lower is these profits
0:45:48 do not just go to investors. They are almost all being pounded back into building out these data
0:45:55 centers. So if, for some reason, compute stops mattering or, you know, we move to distributed
0:45:59 inference or something like that, there’s definitely a downside scenario, which is why I’d say,
0:46:04 think about this cautiously, don’t put your entire portfolio in it. But I look at it as,
0:46:06 it’s like a good one to 5% position.
0:46:13 Hey, Sean here. I want to tell you this little story about Winston Churchill.
0:46:18 So Churchill once said, first, we shape our buildings, and thereafter, they shape us.
0:46:22 And I think this is true not just for the buildings we see in cities, but also for the
0:46:26 building blocks you choose in your company. For any company that I start, I use Mercury for all of
0:46:31 my banking needs. Why? Well, it was built by a YC founder, and you could tell this is built by a
0:46:35 founder who understands the needs of other founders. Second thing is this modern, it’s clean, easy to
0:46:40 use. The design is really nice. You’d never have to drive somewhere, park, put coins in the meter,
0:46:44 get out just to do one simple task. You could do everything in just a couple of clicks.
0:46:47 They got bill pay, checking account, savings account, wire transfers, everything you need,
0:46:51 they got it. I use it for not one, but actually six of my companies right now
0:46:54 and actually even have a personal account with them. It’s kind of amazing. So if you’re ready to
0:46:59 operate in the future, head over to mercury.com, apply in minutes. Disclaimer, Mercury is a
0:47:03 financial technology company out of bank banking services provided by Choice Financial Group and
0:47:08 Evolve Bank and Trust members, FDIC. Thank you to Winston Churchill for that little ad segment.
0:47:15 All right, back to this episode. I need to look into it. That’s not how I roll, obviously,
0:47:21 but it was compelling and I’m fascinated by their entrepreneurial story. I’m amazed how this was
0:47:26 just launched in 2018 and how fast they grew. How much financing did they raise? I don’t know how
0:47:30 much they’ve raised, but I do know that they have no debt, which is quite impressive for
0:47:37 a business like this. I have a lot of interest in retail, like a lot of retail buyers in the
0:47:41 public market. And so they’ve been, every time they need to expand, they do what’s called an
0:47:45 at-the-market raise. So they basically just issue stock into the market and then they raise
0:47:50 capital to do it that way. But I mean, the way they die would be big amounts of debt. So they’re
0:47:55 avoiding that, I think. Well, I forget what that online stock market club is that like Bill Ackman
0:47:58 and all these billionaires are part of where they had got to pitch their stock and they have to make
0:48:03 a compelling argument and it’s like really hard to get into. This is like, if that and Beavison
0:48:08 bought Head Head of Baby, that’s what this is right now. Have you ever been to Sound?
0:48:12 It’s a Sound conference, right? That’s what you’re talking about? No. Yeah, that’s what I’m talking
0:48:20 about. Is it online or in person? I’ve always wanted to go, but it’s an ordeal to go. Maybe this
0:48:24 was your audition tape. Yeah, there you go. What do you think of it, Sean? I mean, I don’t know
0:48:28 anything about this company. I’ve never heard of Iran until Andrew talked about it five minutes
0:48:34 ago, but I like the, I mean, the problem is Andrew is very persuasive. This is also the problem with
0:48:40 me. I can talk myself into anything. That doesn’t mean it was a good idea. And I have done that to
0:48:45 myself many times. You have to be careful. I apply basically a discount factor. It’s like a charisma
0:48:51 discount. So the more persuasive somebody is about everything, the more I have to discount
0:48:55 to some degree. That doesn’t mean I just like throw it away, right? But I have to be careful not to
0:49:00 get overly excited. And then the opposite is true in the other case. Like when somebody’s got like
0:49:04 the David Blaine deadpan delivery of everything, and then they’re saying something that sounds
0:49:09 exciting. I’m like, wait, if that sounds exciting coming from you, that means it must be 10 times
0:49:13 more exciting than you’re actually giving it. And there’s actually like an amplifier there.
0:49:17 So you got to be a little bit careful. This is why whenever I read Malcolm Gladwell books,
0:49:22 I have to be like, wait, this is just a theory. This is just a theory. Like this is not a fact
0:49:26 necessarily. That’s pretty funny. What did you call that? The charisma? The what?
0:49:30 The charisma discount. Yeah. That’s pretty awesome. Yeah, like our friend Jack Smith,
0:49:35 whenever he tells me something, he’ll be like, yeah, that’s pretty cool. I think I like this.
0:49:40 I’m like, do you love this? Yeah, I think I love it. That I’m like, all right, I’m in.
0:49:45 There’s some people who pretty good means it’s incredible. And there’s other people who are
0:49:49 like, oh, it’s the fucking best. And that just means it’s just another thing that they’re doing.
0:49:54 You know, it doesn’t mean anything, right? You have to like, you know, it’s like an audio mixer
0:49:58 has to turn some people’s levels down and turn some other people’s levels up on their microphone.
0:50:03 You know, are you guys tracking those stock stock? You guys did that like investment episode maybe
0:50:10 like six months ago. And yeah, TKO versus Ferrari TKO one, I believe, Sean, I think it’s up like
0:50:15 19%. I think since we did that episode. So you know, that one did pretty well. But I mean,
0:50:20 it was funny that we did a stock picking episode just stock stock of Palooza. And like,
0:50:24 we just didn’t just say the word Nvidia and then stop the recording because that would have been
0:50:27 that would have been sufficient. And like, it was pretty obvious and in our face. But like,
0:50:30 when you try to be smart and make good content, that doesn’t mean you’re necessarily
0:50:35 actually making the best advice is what this is my fear of all YouTube doctors, by the way.
0:50:39 And I like your women. I like it to you. I like all that I like them. But also,
0:50:43 they’re content creators now first. And so you again, you have to apply the discount factor,
0:50:48 which is that their job is every week to come up with new things to say. That doesn’t mean that
0:50:53 I’m not saying they’re wrong either. It’s just that the best advice might have been, you know,
0:50:59 five minutes long said once and never, you know, and or just repeated 15 times would that probably be
0:51:03 more helpful than listening to the three hour podcast about how like, you know,
0:51:08 getting UVB rays in your ears in the mornings is going to help you, right? Like those might not
0:51:12 be the things that actually will help anybody at this point, but that doesn’t make for good content.
0:51:15 And so whenever somebody’s primary job is content, and I say this as somebody whose
0:51:21 primary job is content, you have to like, remember that before you, before you, you know,
0:51:26 get totally hooked and listen to everything that they say. This is the thing I remember in 2013
0:51:31 or something. I had sold a business. I was sitting on like $7 or $8 million of cash and I was like,
0:51:36 okay, I need to invest this. And I read this Tony Robbins book. I’m not a big Tony Robbins person,
0:51:40 but he had that money. I think it was called money mastering the game or something. And
0:51:45 basically the whole book is just like by index fund. I have this disease and I think most entrepreneurs
0:51:50 do where they say, well, that’s the obvious thing. That’s for the normies. I’m going to go find door
0:51:56 number three. And I look back, like I read chip war, which is like all about semiconductors. I knew
0:52:02 that NVIDIA was an amazing business. TSMC was an amazing business. I saw Facebook trading at
0:52:06 five times earnings, but it’s like, I have a really hard time going through door number one.
0:52:12 And I think going forward, I want to be more, I just want to do the thing that makes sense
0:52:16 and is obvious and then forget about it and stop trying to be clever. This is actually an
0:52:20 instance of me trying to be clever. What did you want to invest in? And then you look back
0:52:26 and you’re like, ooh, dodged a bullet on that one. Oh, where was the obvious thing, the wrong answer?
0:52:30 Yeah. Because it won’t be the right answer 100% of the time. NVIDIA was a good, yeah,
0:52:35 NVIDIA Facebook. You look back and you think, I should have known better, but where was it the
0:52:42 opposite? I don’t think I’ve ever really messed up on the one that’s obvious that slaps me in the
0:52:47 face. And I don’t, like I’m actually really decisive when I see that, but it’s interesting.
0:52:53 Mostly it happens for me in private businesses. I think that’s where my decision making is usually
0:52:59 the strongest because I usually do what’s obvious and will work for sure. But in public markets,
0:53:04 it’s very easy to get excited. Have you had one, Sean, where you’re like, I almost did this and
0:53:09 God, I’m thankful I didn’t. Well, I’ve had a bunch of those, but is it specifically where it was
0:53:15 like an obvious kind of scene pretty straightforward, but it was a mistake? Oh, I have one. So my first
0:53:20 designer was this guy named Adam St. And he’s the best, the best dude. Like he was one of our
0:53:25 designers, but also like a great friend. And he went off, he left after a couple years at Metalab
0:53:30 and he started this company called Bench. And so he comes to me and he’s like, Hey, we’re doing,
0:53:38 you know, basically AI accounting 10 years early. And I loved him. I loved his co-founder. I loved
0:53:43 the idea, but I looked at it and I was like, Oh, it’s a services business for accounting and low
0:53:49 margin. And this is going to be hard. So I didn’t invest. And for the next 10 years, I felt like
0:53:55 an idiot because over and over and over again, I’d see these big series A, series B, Bain Capital
0:54:01 buys them for, you know, huge amount of money or whatever. But it turned out that when Bain invested,
0:54:07 I don’t think they got any money off table, nor would I as a shareholder. And I think a lot
0:54:13 of people saw this, but he just posted a big post on LinkedIn about how the business basically went
0:54:18 to zero and shut down overnight and was a total disaster. And that’s an example of one where I
0:54:23 was like, man, I wish I was so badly wanted to be part of that business and I would kick myself
0:54:27 for years. And then in the end, I’m like really glad that I didn’t invest in it.
0:54:31 Well, I first want to co-sign what Andrew said. I think, you know, I have meddled in the
0:54:37 overthinking Olympics many times. And I, and more than anything, the reminder is to do exactly what
0:54:42 Andrew is saying, which is like, stop trying to be clever and just simply asking, what’s the obvious
0:54:46 here? What’s the obvious move here? What’s the obvious answer here? And if there’s not an obvious
0:54:51 answer, just move on. Like you don’t have to do anything. No action is required in most situations.
0:54:56 And I’m struggling to come up with a time just like Andrew. I’m struggling to come up with a
0:55:02 time where there was a obvious thing staring at me that I did and it turned out wrong,
0:55:07 which almost just reinforces even more how hard that is. Whereas the opposite, if you said, hey,
0:55:12 tell me a time where you kind of thought you got a little cute, you tried to be clever,
0:55:17 and it didn’t work out. Your buddy texted you at 11 o’clock at night and said, hey,
0:55:23 rounds closing tomorrow and these 10 cool people are in. And it’s this really cool founder. I can’t
0:55:29 even blame anyone. It’s even simpler than that. It’s, I was a believer in crypto early. Santa was
0:55:34 talking to you about crypto early for a long time. And then one year last year, the year before
0:55:39 something like that, where crypto had a dip, I was like, I’m a tax loss harvest. So clever. You
0:55:43 know what I could do? I could sell this shit. I could buy it right back. I still own the thing.
0:55:52 And I booked this beautiful tax loss. Oh my God, I am 900 IQ and I just, I am the guy. And so then
0:55:58 I sell the thing and I booked the loss and I tell my accountant, I say, don’t worry about this one.
0:56:02 I advised myself on this. I got it. And they’re like, okay, great. We’ll make sure we have that
0:56:07 booked for the year. And I’m like, cool. And then I’m like, I should buy back in. I was like,
0:56:10 you know, but I did just read this report. I think it’s going to go a little lower first.
0:56:17 Let me just wait. Just let it go a little bit lower. And the stupid goddamn tax decision
0:56:22 ended up costing me what probably saved me maybe, I don’t know, hundreds of thousands of dollars
0:56:28 in taxes cost me millions of dollars in gains because I didn’t just do the thing. The obvious
0:56:32 thing was sell the thing, book the loss, buy it right back. You’re exactly where you were five
0:56:36 minutes ago with additional tax losses. The getting cute or clever was thinking, you know,
0:56:40 the knife still falling. Let me let it fall a little further. And then it fell a little bit
0:56:44 further. And I just patted myself. And I said, I’m so smart. I knew it was going to go down.
0:56:48 It’s going to go down a little bit more. I was actually reading this article the other day.
0:56:51 And I saw this guy on Twitter, who’s got a picture of a fucking cartoon. And, you know,
0:56:54 he said something pretty smart. Let me just hold on for a second. Let me just wait till it goes
0:56:58 down a little bit more. And then it went right back up way past where I bought it. And I ended
0:57:01 up buying in at a much higher price. I would have been better off doing nothing. It’s like,
0:57:07 you know, Ross Ulbricht was in jail and just got freed, which is I know a big day for you, Sam.
0:57:11 Big day. And so I’m trying to I’ve got to shave my legs for our first date.
0:57:15 He’s walking out of prison, by the way, holding a plant, which I love. I don’t know what he’s
0:57:19 just got a small house plant that he had, I guess. I don’t know what that is. But,
0:57:24 you know, everyone’s like, oh, my God, you know, I’m so glad he’s free. He must have been so tough,
0:57:27 blah, blah, blah. And then somebody goes, well, government did do my favor because
0:57:32 he was basically forced to not sell his Bitcoin from 2011 or whatever when he was running,
0:57:37 you know, Silk Road and Bitcoin was like, you know, in the dollar range to now it’s $100,000.
0:57:41 And it’s like, yeah, they also did give him a gift by not letting himself.
0:57:45 Wait, is that so he still he owns some stuff? I thought that Tim Draper bought it.
0:57:49 No, that’s what they seized. Presumably they didn’t seize all of the Bitcoin that this guy had,
0:57:52 right? Like he could have had Bitcoin in multiple wallets. He could have had personal
0:57:56 Bitcoin. Like they seized what they got from the Silk Road. That doesn’t mean they had all
0:58:02 of his crypto wallets necessarily. You know how people say, like buying a home is forced savings?
0:58:08 He, he just had prison was his forced savings. Yeah, 24 hour lockdown forced savings, my friend.
0:58:12 For example, I had the opportunity to invest in Coinbase, not super early, but like let’s say
0:58:16 series A or B somewhere there. And I remember doing the math and being like, okay, I believe in
0:58:21 Bitcoin, should I invest in Coinbase, the exchange or buy Bitcoin? And again, tried to outsmart
0:58:25 my answer should have been just do a little bit of both and let it ride. And in the end,
0:58:29 I did this analysis that showed me that you should just own the underlying asset Bitcoin. It’s going
0:58:33 to grow faster than this exchange, which is going to have multiple competitors that are exchanges,
0:58:37 et cetera, et cetera. And actually, I was right, Bitcoin over performed. But the thing I didn’t
0:58:42 factor in was that I would have made more money had I just a done both, but B Coinbase, you couldn’t
0:58:47 sell your shares. So like Bitcoin price went down, you couldn’t panic sell. And so the forced
0:58:53 savings of just being a Coinbase stockholder will turned out to be more beneficial than being liquid
0:58:58 with Bitcoin, where you were now at the mercy of your own brain. And there’s been, you know,
0:59:02 tons of examples like this. I have a good one. I talked about this a little bit,
0:59:09 but MetaLab back in 20, I guess it was 2012 or 2013, we designed Slack. And at the time,
0:59:15 Slack was actually a failed gaming startup. And Stuart Butterfield was basically trying to do
0:59:20 something with the remains of it, and trying to build a chat tool. And so he comes to me, he’s
0:59:26 like, Hey, I’ve only got $80,000 budget for this. And I was like, well, we normally, you know,
0:59:32 this would cost quite a bit more, but we really, really want to work with you. We’re like big fans.
0:59:37 And so he’s like, Hey, what if we did some stock? And I’m like, no, no, no, no, no, like, you know,
0:59:42 we don’t like MetaLab would go broke if we just took stock, we don’t do that, whatever. And so I
0:59:47 said, no, and my thinking was, well, I want to get as much cash as possible so I can invest it in my
0:59:52 own businesses. And if you think about what I was investing in at that time, it was my productivity
0:59:58 software where I lost $10 million. So I instead put it into this money losing productivity software,
1:00:03 I could have made $100 million or something crazy, because it was at like a 20 million valuation at
1:00:08 that time, and it sold for $28 billion, I think sales force. That’s a good one where it’s like,
1:00:14 maybe, maybe obvious. Actually, that’s maybe not that obvious, but it’s still, it’s still a brutal
1:00:20 one. My most recent one is Justin Mayer’s true man. He was raising for it at a, I forget what it was,
1:00:25 reasonable valuation. Justin’s my friend. I think Justin’s amazing.
1:00:30 I thought he had too much going on. And I was like, no way are you going to focus on this. I pass.
1:00:33 And he just told me how things are going. And I’m like,
1:00:37 shit, like, it was so obvious that you’re a winner. I should have just said, yes, I’m in.
1:00:44 In my, my Ecom business, my, I brought in a friend into the business, gave him equity. He,
1:00:47 he knew Ecom a much better than I did. I thought, okay, my rate of learning is going to go up with
1:00:53 this guy. And I was so excited for this like value ad from this guy. And his value ad was like,
1:00:56 he’s super good at a bunch of things. One of the things he’s really great at is marketing.
1:01:00 So I was like, okay, he’s going to be so helpful with marketing. And so we closed the deal. We signed
1:01:04 the docs and I set up the first session. I’m like, oh, I can’t wait for this guy to come in and just
1:01:09 spill the beans and all the secrets that I don’t know and make, and they solve all of our marketing
1:01:13 problems. This is going to be amazing. And he comes in and he has that first, first session. And
1:01:18 instead of spilling the beans, he just opens up a Google doc and he’s like, okay, what’s the best
1:01:23 product that you sell? And we’re like, it’s this one. He’s like, cool, what’s the best photo of that
1:01:27 product? We’re like, what do you mean? He’s like, just like, what’s the best picture you have about
1:01:33 that product? And he puts it in the Google doc. And he’s like, okay, if you know, like, what do
1:01:37 people say, like, why do people buy this? And we’re like, well, I mean, the bike is great. He’s
1:01:40 like, yeah, but like, what, like, why do they buy it? Like, what do they say? And he writes down
1:01:44 the caption and then he like goes and sets up a Facebook ad of that product with that picture
1:01:48 with that line of like what people say. And he’s like, cool, let this run for like the next three
1:01:53 months. And so then he so that becomes our top performing added took him five seconds. And I’m
1:02:00 like, okay, that was cool. But like surely there’s more. And so nine months go by and I hit them up
1:02:04 and I’m like, dude, you’re right, great friend. And I was so excited for you to be in this thing.
1:02:08 But like, honestly, I feel a little disappointed. I feel like you could have done more. I feel like
1:02:12 the expectation was you’re going to do more, you know, like, he’s like, oh, I did so much. And
1:02:15 I was like, wait, what do you mean? Like, what did you do? Like, where’s the list of like, what,
1:02:17 what do you mean? When you’re saying you did so much, I’m saying you didn’t do enough. What’s the
1:02:23 gap? And he was like, oh, I stopped you from doing all the dumb things. He’s like, I didn’t do extra
1:02:27 things. He’s like, remember when you got really excited about influencer marketing? I was like,
1:02:30 oh, yeah, that was awesome. I read this article and I met this guy who was crushing with influence
1:02:33 marketing. I wanted to go all in on influencer marketing. He’s like, yeah, remember, I just told
1:02:39 you like, Hey, Facebook’s working, like just keep that going. Like, don’t have you run into a wall
1:02:44 with Facebook? No. All right, then let’s just keep doing that. Don’t get distracted. And he’s like,
1:02:48 that’s what I did. Like the second thing you wanted to expand your products and like, actually,
1:02:51 I just stopped you from doing that. And then you had this genius idea to go subscription.
1:02:56 And I told you really like, Hey, like this thing’s working. So let’s just keep doing the
1:03:00 obvious thing here and just keep it going. And basically he laid out like four or five things
1:03:06 that were all just simply like him taking the knife out of my hand as I’m a toddler and preventing
1:03:11 me from stabbing myself. And he was like, those were all super important moments that like this
1:03:14 business would have gone a different way. And I was like, I don’t know if you just yet I mind tricked
1:03:19 me into believing that you added ton of value or you actually just said something very, very
1:03:24 insightful to me. And I landed up, I think he’s actually very, very insightful and that that is
1:03:28 one of the best ways to help someone is to just simply prevent them from hurting themselves,
1:03:32 rather than solve their problems for them or give them some new tool that they’ve never seen before
1:03:36 or will solve all their problems. And this is why I love investing in private markets, right?
1:03:42 To Sam’s point about houses being forced savings with a private business, there’s so many of our
1:03:48 businesses that I’ve owned where if I could press a red button, like a sell button, just, just any
1:03:54 given day because I read something or I’m paranoid or whatever, I would have way, way less wealth
1:03:58 because I would have panicked over and over and over again. And when I look at my public market
1:04:03 track record, I’ve actually had the right call over and over and over again, but then sold way too
1:04:10 quick. Like I bought Chipotle stock when they had that health crisis. I doubled and then I just sold,
1:04:15 right? I could have made way more money. Same with meta, all these other things. So I think that
1:04:23 having a psychological lock in creates a really positive thing for people like us who are impulsive.
1:04:28 There was a guy at the event we just threw and it was like, everybody was giving their claim
1:04:32 to fame. It’s like, okay, this guy built this thing. He’s been spent 20 years building this.
1:04:36 He’s been grinding every day and he built this empire. And then this guy, he’s building this,
1:04:41 this empire. Jimmy’s building this YouTube empire and then this chocolate empire on top of it.
1:04:45 And everybody’s just like, go, go, go, go, go, go, try to figure it out. And then there was one guy
1:04:49 there and it was like, yeah, he’s one of the most successful investors in the world. He invested
1:04:54 in all five of Elon’s companies at the first round and didn’t sell. And it was like,
1:04:59 there’s one way to do it. It’s like, identify the best operator in the world or one of somebody
1:05:03 who you think is just like, absolutely brilliant and obsessed. Invest in all their things. Don’t try
1:05:09 to judge. Rockets are going to fail, but this is going to work. Just invest in the jockey. Was it,
1:05:14 was it an Antonio? No, I can’t say the name, but one of the wealthier people there in a room of
1:05:19 the wealthiest people in America, some of, and someone was like, what did you do? He goes,
1:05:24 I’m like the CEO of my kids right now. Like I dedicate my life to my children.
1:05:29 And like you guys work 16 hours a day on making money. I work 16 hours a day on making my kids
1:05:33 like happy and healthy and making sure they’re well loved and all this other stuff. And I was like,
1:05:35 you’re the greatest person I’ve ever met.
1:05:44 Yeah. And he had that like, it’s not the right way to describe it, but like a
1:05:48 simpleton mindset, like the mid-wit thing, but like where it was like, this seems simpler,
1:05:55 I’ll go that route. Whereas many other people refuse to sit still and stay the course.
1:06:02 What’s the quota Pascal? All of man’s misery arises from inability to sit quietly in a room,
1:06:08 you know? This guy did not suffer from that. Guys, this is awesome. I can keep going.
1:06:10 It was great. I just, I’m happy you’re well, Andrew.
1:06:15 Yeah. Yeah. It’s great to see you guys. I got to come back soon. It was fun.
1:06:16 All right. Thank you. That’s it. That’s the pub.
1:06:35 Hey, Sean here. I want to take a minute to tell you a David Ogilvy story.
1:06:39 One of the great ad men. He said, remember, the consumer is not a moron. She’s your wife.
1:06:44 You wouldn’t lie to your own wife. So don’t lie to mine. And I love that you guys,
1:06:48 you’re my family. You’re like my wife and I won’t lie to you either. So I’ll tell you the truth
1:06:54 for every company I own right now. Six companies. I use Mercury for all of them.
1:06:58 So I’m proud to partner with Mercury because I use it for all of my banking needs across my
1:07:02 personal account, my business accounts. And anytime I start a new company, it’s my first
1:07:05 move. I go open up a Mercury account. I’m very confident in recommending it because I actually
1:07:10 use it. I’ve used it for years. It is the best product on the market. So if you want to be like
1:07:16 me and 200,000 other ambitious founders, go to mercury.com and apply in minutes. And remember,
1:07:19 Mercury is a financial technology company, not a bank. Banking service is provided by Choice
1:07:25 Financial Group and Evolve Bank and Trust members FDIC. All right, back to the episode.