This Week in Startups - Justin Kan & Brian Chesky: The Rejections That Built Twitch and Airbnb | E2344

Episode Date: September 30, 2026

This Week In Startups is made possible by: Plaud: http://Plaud.ai/twist CLA: https://claconnect.com/tech Sentry: http://sentry.io/twist Lightfield: http://sentry.io/twist Today’s show: Google once interviewed the entire Justin.tv (now Twitch) team and then walked away from an acquisition deal, saying they weren’t good enough. Seven investors passed on buying 10% of Airbnb for $150,000, a stake pegged at roughly $9.25 billion today. Lon Harris and Alex Wilhelm revisit classic TWiST interviews with Justin Kan of Twitch and Brian Chesky of Airbnb to find out what those founders understood that the market didn't. PLUS, how did a box of cereal win over Paul Graham and get Airbnb into Y Combinator? Flashback Guests: Justin Kan on X: https://x.com/justinkan Twitch: https://www.twitch.tv/ Justin Kan (@justinkan) on X Brian Chesky on X: https://x.com/bchesky AirBnB: https://www.airbnb.com/ Relevant Links: TWiST E900: Justin Kan, co-founder of Justin.tv, Twitch & Atrium (Feb 5, 2019) → https://thisweekinstartups.com/episodes/L3DP8VCE12s TWiST E1735: Airbnb CEO Brian Chesky on early rejection, customer focus & AI (May 4, 2023) → https://www.youtube.com/watch?v=aZ-BjJZxNoA Halo 3: the game that flooded Mahalo with traffic → https://en.wikipedia.org/wiki/Halo_3 Weblogs, Inc.: Jason's blog network, home of Engadget → https://en.wikipedia.org/wiki/Weblogs,_Inc. Instinct: the invite-only personal AI agent from founder Noah Shinn → https://www.instinct.co/ AMD to acquire World Labs for $8.2B in stock; Li becomes AMD chief scientist (Bloomberg) → https://www.bloomberg.com/news/articles/2026-09-28/amd-to-buy-fei-fei-li-s-world-labs-ai-startup-for-8-2-billion Lovable: where Lon says he'd rebuild a bad app in two hours → https://lovable.dev/ Y Combinator → https://www.ycombinator.com/ Paul Graham on X → https://x.com/paulg Obama O's and Cap'n McCain's: the cereal that funded Airbnb → https://en.wikipedia.org/wiki/Airbnb#History Stripe: what Paul Graham suggested they build instead → https://stripe.com/ Timestamps: 0:00 Justin Kan on episode 900 (Feb 2019) 3:44 When Jason and Justin realized people watch gaming 5:52 Mahalo, Halo 3, and betting a company on 3% of its traffic 10:21 CLA: Get started with CLA's CPAs, consultants, and wealth advisors now at https://claconnect.com/tech at https://www.claconnect.com/withyou 11:42 The "zombie startup" years and Google's rejection 17:06 Instinct: the AI agent built for venture capitalists 19:43 Sentry: New users can get $240 in free credits when they go to https://sentry.io/twist and use the code TWIST 28:08 Kozmo.com and the era of "make number go up" 30:06 Lightfield: Name one person who's ever enjoyed updating a CRM. Exactly. Lightfield's AI agent does it for you - it even prospects and books your meetings. Used by thousands of startups. Free at https://lightfield.app/ 32:57 Why AI companies have to monetize faster 34:00 Clip: Justin Kan on knowing yourself as a founder 37:24 An early rejection of "founder mode"? 38:33 Brian Chesky on episode 1735 (May 2023) 39:27 "I hope that's not the only idea you're working on" 43:27 What 10% of Airbnb for $150K is worth today 46:05 Obama O's and the Y Combinator interview 48:15 Why investors fund unkillable, attention-hacking founders 54:37 Clip: Chesky on AI, deepfakes, and authenticity 58:13 Is AI costing jobs or just making them busier? 1:05:04 Did AI create "millions of startups"? Subscribe to our newsletters on Substack: TWiST: https://twistartups.substack.com/ TWiAI: https://www.thisweekinai.ai/ TWiVC: https://thisweekinvc.substack.com/ Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp Follow Lon: X: https://x.com/lons Follow Jason: X: https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis Check out all our partner offers: https://partners.launch.co/ Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland Follow TWiST: Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Instagram: https://www.instagram.com/thisweekinstartups TikTok: https://www.tiktok.com/@thisweekinstartups Substack: https://twistartups.substack.com

Hey, everybody. Welcome back to This Week in Startups. I am Lon Harris. With me, not, as always, on a special occasion. Former co-host, Alex Wilhelm. Thanks for being here. Oh, it's so good to come back. Friendly faces. I'm sweating in a blazer again. You know, just like old. That's my friend. This Week in Startups is brought to you by Lightfield. Name one person who's ever enjoyed updating a CRM. Exactly. Lightfield's AI agent does it for you. It even prospects and books your meetings. used by thousands of startups. Free at lightfield. dot app. Sentry.

Your team should be focused on shipping features, not chasing down bugs. New users can get $240 in free credits when they go to sentry.io slash twist and use the code Twist 26. CLA Innovation takes balance. CLA's CPAs, consultants, and wealth advisors can help you get from startup to where you want to end up.

Get started now at CLA. And like all times, you will notice I am wearing my plod pin. It is right here pinned to my jacket recording everything that everyone says. And then it will conveniently download it, a summary, a transcript, meeting notes, whatever I want, all organized. I don't have to do a thing. I just have to use, hit the button, tell everybody that they're being recorded so you don't do it serptitiously. And then it's all going to just be done and waiting for me when I am done with the show or the call. If your work depends on conversations, interviews, meetings, anything like that.

We urge you to get a note pin for yourself. You can check it out. Find out more at plod.a.ai slash twist. And be sure to use the code twist to get 10% off your order. Thanks to our friends applaud. We applaud plot. We do indeed and plod, plod. The first good AI wearable is how I think about plot.

They cracked it. They actually got something together that people need to use and makes their life better. And the first couple of pendants just didn't hit the marks. So shout of the plot for being early, frankly. Yeah, that AI pins. those people are like, why are we taking strays? Why do we have to take strays on the plot ad read? I had to watch their launch video lawn.

That's why they're taking strays. They took five minutes of my life and I'm mad about it. Yeah. It's like, I don't know. It just like does that a pin is a weird form function. I think we can all. Well, anyways. The AI.

What are we here today to do? What are we doing? We're not here today to just talk about AI broaches. We're here today. We're doing another of our patented flashback episodes. We're going to look back at an episode of Twist from February 29. Folks, this is episode 900, a landmark show. And it features Justin Khan, as Jason's guest, of Justin TV, later to be known as Twitch.

The guy who once strapped a camera to his head and live streamed his entire life around the clock. That 2007 stunt, Justin TV, soon opened up to anyone who wanted to broadcast and live streamed their lives. And its gaming section later spun off as Twitch, which Amazon, of course, bought in 2014 for 900. $170 million, and you could currently still use if you want to catch up on the hottest video games or get very, very angry about politics. By the time of this chat, Justin had moved on to Atrium. That's a legal tech startup that had just raised $65 million in their series B. Atrium would shut down barely a year later, and Justin has since gone back to gaming. He had the NFT Marketplace Fractal, and now he's doing Stash, a Web Shop and Payments Platform for mobile game developers.

So let's go back to 2019, Alex, take a little trip into history and check out some of these classic clips featuring Justin Conn. Absolutely. Now, I have a number of things that I want to get too long, but where are we going to start in this relatively long? And I would say luxurious interview. Yeah, I'd like to jump. My first clip is at about five and a half minutes in. And they are talking about Justin and Jason both when they first discovered that people on the internet wanted to watch other people on the internet play. video games. Let's take a look at this startling revelation.

Gamerers were a genre unto themselves. Because I remember back in 2007, 2008, I was doing Mahalo and we started putting video game videos on the YouTube, the YouTube, and they went bonkers, like millions of views. But when did you realize video games were what streaming was about? Yeah. So how did that happen? Well, so we launched this site in October 2007. and that anyone could stream anything. And people streamed. There was a lot of people trying to make talk shows and a lot of social kind of broadcasting,

like someone would just be hanging out. And then there was this nascent category. We kind of grew that mostly focused on trying to get celebrities and things like, you know, more famous people to create content. The problem was they didn't really want to do it on an ongoing basis. Right. And so a couple of years later, this was maybe in 2010, we'd grown to a fairly big website.

It was about like the top 300 on Alexa or something. So you got over 10 million people a month, basically. Yeah, something around that. But the problem was it stopped growing. And we were like, we had no idea how to grow Justin TV, this web platform for anyone to stream. So what we decided to do was go back to the drawing board and basically sit in a room and try to brainstorm ideas.

And my co-founder Emmett came with his idea. And his idea was, let's focus on the gaming content on our site, which was only 3% of the content at the time. And so gaming was only, I think, a couple hundred thousand unique a month on our site. But his argument was this is the only type of content that I actually want to watch. StarCraft 2 had just come out. So that was driving a lot of interest. Absolutely.

And there was driving a lot of YouTube videos and he wanted to watch live streams. And mostly they were on other people's sites. They were on Ustream or live stream or but he thought we could focus on this and build a community around the gaming content. Teach people how to do the zero grush. I witnessed this person. I mean, this clip stands out to me because I was working with Jason at Mahalo in this exact era when we also made this discovery alongside Justin TV that if you're, I know it's hard to even conceive of if you are under like 35. But if you are of a certain age, the idea that people for entertainment wanted to watch other people play video games was very counterintuitive for a lot of people. It was incredibly counterintuitive.

long. I mean, my parents used to make fun of me for driving to their office, because they had faster internet that we had at home, we had dial-up. And I would watch old StarCraft Brood War Vods from South Korea and the pro circuit. And they're like, you're going to go, drive, to sit and watch someone else in the past play a game. I'm like, you bet

your ass, I'm going to do that. 100%. Yeah, now it seems so, both because it's incredibly common and everybody enjoys doing it and it is not that way to be thing. But also, if you think about like sports, like you might enjoy playing basketball yourself, but then you'd enjoy watching professionals do it very, very, very well, or college kids who are also very good.

So it isn't really that strange of an idea that people who love video games would want to watch very skilled players play video games. But at the time, I'm telling you, we did not believe it. And at Mahalo, we built Halo 3 was the first game we discovered this for. And we started building pages about how to find. all hundred skulls that are hidden in the maps or in Halo 3. And it was dwarfing all of our other traffic to the point that it was undeniable. We had to focus on video games.

And it really took a while to sink in our Gen X boomer brains that this was where things were going. Well, I think one thing that happened during this time was the mainstreamization of gaming. I mean, if you think back to when I was a kid, I was born in 89 just for context for everybody. So I grew up in the 90s. And the early 2000s, I was in my teenagers. And gaming was still very much a fringe activity. It was still something the nerdy kids did and the cool kids did not.

That is now done. FIFA and Madden and the along the games have really made gaming mainstream, Cod, et cetera. So I just love the story because I think it shows how you can catch something early and turn it into something quite large. It was 3% of their traffic, as Justin just said. That's not much. But if you can read the T-Leaves correctly,

you can build a subsidiary product that becomes so big, everyone forgets what you were originally working to build. Right. There's also a little tactical, practical, Jason founder tips sort of hidden in there, which is trusting your gut. I mean, you know, Emmett, his colleague said, this is the kind of content I would want to watch. And so even though data-wise, it was only a very tiny percentage of their overall viewership, he was like, we should still focus on this because just my instincts are telling me that other people are like me and are going to want to check out video games. Yeah, but it's just such an orthogonal bet. even given everything that you said, think about it if you're a venture back startup today,

and you're going to say, hey, we're going to just focus on 3% of our user base. Like, you would get laughed out of the room. So even like after this happened, it's still a non-traditional approach to growing a startup. And it did yield a roughly billion-dollar outcome, as you pointed out. But I'm just amazed at the courage it must have taken to devote scarce resources during a period of flat growth on a small user segment. Brilliant in retrospect, but bold and brave, I think, at the time. Yeah. And also, the last thing I would mention is in the early,

days especially now video game companies and publishers understand that people live streaming, except Nintendo, there's still, you still hear issues about Nintendo. But other video game publishers now realize we need live streamers. That's our lifeblood. They're promoting the games to new users. But back in the day, when you would try to do this, you're constantly being hit with content, copyright strikes by the publishers of the game. Like, you're not allowed to show people what happens in my game before they play.

it themselves, that's violating my copyright. So there were legal harangues to worry about as well. I mean, what they should have just said was is that we're using this for AI training purposes and therefore your IP doesn't hold any water. And we're just going to dissolve your entire business. No, I joke, but I do love seeing how things change and it become de facto over time.

You know, talk about Uber on the show a couple times. Uber was the same thing. Like when I was growing up, it was don't talk to people on the internet, don't talk to strangers, don't get in their cars, don't go in their homes. Yeah. Well, you know, it turns out, that was entire. wrong with a little bit of technology and a little more time. One of the themes we talk about

over and over again on this week in startups is making sure you do your chores. I'm no expert on these days. I have some experience. Stephen Estes from CLA is an expert. Let's talk about being cash efficient. Tell us about efficiency and what you see in the top tier startups in your practice. We're seeing kind of an interesting trend out there where companies aren't needing to raise quite as much as they had in the past. You really have to be careful as a founder to only take on as much money as you really need. You've got to do the forecast and you've got to do the modeling and you've got to dialed in and get it right. Otherwise, you're going to end up either not raising enough capital to get to where you're going and you're going to have to go get venture debt or go back, have an extended

or you're going to give up too much of the company because you just didn't recognize how much money actually needed. Yeah, very important to get this stuff right, folks. And that's really a bummer when startups don't do things in a button-up way. I always have a great partner. A good partner to have on this adventure while things change, my friend Stephen over at CLA. So if you want a trusted advisor by your side who will navigate you through taxes, accounting and everything in between, it's time to take action. Visit c-lac connect.com slash with you. And don't forget to drop a mention that your boy, Jake, I'll send you. That's c-lac connect.com slash with you. Start today. All right. Give us your first clip. Yeah.

Yeah. So this interview blew my mind. I was a Justin TV fan. I was a early Twitch fan. Who was? Well, all of my friends. It did feel a little lonely at the time. I'm not going to lie. But one thing I really appreciated about this interview was that I didn't realize how long Justin TV team had spent in the wilderness. Lost years. Just not having the right product at the right time for the right people at the right price. and we're going to loop back to this, but compare what he's about to tell us

to the current startup era of rapid growth and what I might call instant product market fit. Let's take a listen. Working in a dark room on this, you know, programming something. Yeah. And I would think, why am I doing this? Why, what is going?

Where is this going? And it never felt like we were making progress. It never felt like there was an inflection where it was like, oh my God, we figured it out. We figured it out. Yeah. And so, or maybe we had data. like that but then it quickly referred it back to the baseline right and now a hundred

days of struggle exactly and so I you know it was it was that zombie startup feeling where we were 25 people five years in had you know weren't growing super fast and we were like how where does this go right we don't see an exit insight and you decide to call this meeting when did you decide to call that meeting what was that moment where you're just like I need to shut this thing down and and move on or I just need to challenge everybody. What was that the impetus for deciding to sit everybody down to do that brainstorming meeting? All right.

This is a good one. I don't think I've actually told this story very publicly before, but basically Google came around and was like, we're interested in acquiring Justin TV. And it wasn't a very high price, to be honest. But they basically, long story short, we talked to them, we talked to them. What was their theory? Their theory was YouTube owns basically the entire internet on the video. and this is like internet video

and this is one sliver, small sliver. A feature. Yeah, exactly. We might as well buy the feature. And they had YouTube live at the time. And so they came to us and, you know, Justin TV was a decently sized site. It just wasn't very profitable

and it wasn't like, you know, it was a small company. And so we, they came to us and they wanted to interview our entire team. They interviewed our entire team. We let them, which you should never, ever do. Never let an inquire ever interview your entire team. Because they interviewed our team and then they were like,

your team's not good enough, goodbye. Wow. Yeah, which was like the most demoralizing gut punch of 2010 or 2011, whenever I think was 2010. Thanks for letting us come in. We regret to inform you, your team sucks. Exactly. Google's people are smarter than yours.

We would never let you in the building. You would be like significantly lowering the bar for the people at Google, basically. Right. We would never let you on campus. Yeah. Like literally on a Sunday, the Justin TV team is not allowed. You have to work out of the sales office, except they're not good enough to work there, so we can't acquire your company. It's like how they treat the contract.

Exactly. It's like you could work in the cafeteria, but you have to work for another company that we contract with. Right. Yeah. You'd be like the cafeteria employees. We're going to contract you or the customer support people. We just don't want you to consider your team Googlers. Yeah.

Like you can't call your team Googlers and you have to live in Fremont. What does it like to go back to work with the team after this happens? What do you say to them when you've been left at the altar? You just say like F Google? Well, yeah, it wasn't great. Actually, the team took it very well, which is very surprising. So that was the impetus. We were like, well, we can't get acquired and we're not growing.

So we need to figure something out. And so we went to the drawing board and we came up with these ideas. And that's when we had the meeting and we started figuring it out. The thing that I find staggering about that clip lawn is how long the company had to try things, figure things out. and essentially have its flop era. One thing I love about 2006 and the AI era and startups today is that they're going from zero to 100 million in revenue,

zero to a billion dollars in revenue in record time. And frankly, as a business journalist, that's exactly what I want to cover, right? That's the type of story that I want to go dig into. But I also wonder what we're losing because I don't think today anyone would let a high-profile founder spend so many years,

essentially beating their head against the rock and trying to figure out a way through. They would tell them to go build something else or to, you know, change of what they're doing. But they had time to figure it out, to find that little audience be talked about earlier. And I almost worry that we're missing something today. I mean, how could we not be in a way? I mean, I guess the argument would be, well, you know, you would just start your next company. Like the the pace is still was always relentless. It was just like maybe you used to.

to have three years with one company to pivot five times until you found the right thing. And now you would just like start a company and it goes out and get it's done or start another company. It gets acquired. Start another company. Or you're starting three or four. Your agents are building three or four different projects at once and you're picking which ones find traction and getting rid of the other.

So I'm not really sure if things are changing that much or if it's just the tools to change the journey. Maybe. but let me let me throw this at you. So instinct, right, this well-known agent that everyone's been talking and done using. I have not gotten an invite yet, sorry, so I can't tell you how it is firsthand.

You know, Lon, one day we'll be at the cool kids table. Not today, but eventually, you know, down the road. Who was talking about instinct first? Venture capitalists. Who got access first? Venture capitalists. The product was venture legible from the get-go, because what does it do?

It helps you untangle your snarl family life and you solve your calendar and books your job. It's literally the venture capital toolkit. There was a post today about this exact thing. Like, it helps you do, like, what are the things that people keep talking about? It helps you do. It helps you sort your luxury items and shop around for high-end items. It helps you book your lavish vacations, and it helps you get the best reservations

at the hottest spots at town. Are those everyday consumer concerns, or are those very, very rich person concerns? I'm flying into Aspen tonight, and I don't know where I should eat. Who has the best croissant soup on the mountain? Like, that's not me. I don't need that. What I need is someone to order pouches and paper towels because I forgot. Not really the instinct use case.

But this company was not that. Like no VC makes up and goes, oh, I can't live stream my e-sports. Dying. I need that. So that's why I think that you're correct that you can build faster now. You can pivot it faster. You can try more ideas. But I wonder if we're building closer to the same bullseye versus throwing at the entire

dartboard as we might have been back in the early 2010. It's an interesting, it's an interesting question. I mean, I think it's a very, it's a very open question. And I just, it's so different. Every aspect of the founder journey was so different back then. AI has really upset the entire cart. And what we're looking at, like what investors are looking at when they're picking companies is totally different today.

I mean, I think, you know, on our venture capital show, or when you hear VCs talk, about it, it has since become reduced entirely 100% to you're making a bet on the founder. Nothing else. You can't rely on anything. Like, is there, is this moat going to exist in two weeks or is open AI going to invent this? Like, is this model going to become too expensive to run and we're going to need to use this kind of model instead? Who knows?

It depends on the next model to come out. Like, all of these things have become question marks. How big is the TAM? Like, what kind of audience exists? And so all of those questions are kind of unknowable now. And so they're making bets 100% for like, I like the cut of this guy's jib. Your team can push code faster than ever with the help of AI agents. But when something in production breaks down, you have no idea what's really going on.

So you just give the agent your error message and you hope for the best. But there is a solution and the solution is Century, which is already being used by 4 million developers. Sentry is constantly watching your app. So when something breaks, it already has all the context. And Sentry's AI debugger, Sear, can quickly determine the root cause. Now your agent isn't guessing. It's targeted on the solution. If you're running cursor, clod code, factory droid, copilot, or anything else, there's nothing to install.

Anything that can call an MCP server or a CLI can talk to Century. Sentry finds the bug and your agent fixes it. Try it free at S-E-N-T-R-Y.I-O and tell them twist. sent you, the have a free debt plan, and listeners to our show can use the code Twist 26 for $240 in Century Credits. I think also the expectation of faster growth makes more sense if we consider it from perspective of how quickly technology is changing, to your point. So think about the lifecycle of an AI model.

You spend all this money in collecting data, then you pre-trained, then you post-train, then you find turn, then you do reinforcement learning, and then you put it out there, and you have X amount of time to monetize it before it becomes essentially moot. So the cycles of tech are much faster now, so there's less time. to grow slowly, to stagnate or whatever, because by the time you find your footing, you'll be dead three times over. But back in 2019, pre-Claude code,

pre-agentic coding, pre-usible AI, pre-ChagyPT, very different worlds. Maybe that's what I'm referring to, is just that no one could do as much, so people did less and we had more leisure time, perhaps? Well, and it's also like, if I had a great idea and I built the app and I released my app,

even if my app sucked and somebody else used my app, it was like, this is bad, I could make a better version of this. that's three months, four months of like, I got to find a few other devs or a technical co-founder and we got to actually build it and test it and make sure it works right and then maybe we can launch it.

Today, if I was like using an app and it was like, this app sucks, I'm going to go into lovable and monkey around for two hours and maybe I'll make a version of it that I like better. Yeah, and I'll host it myself. I won't put the app store tax. It'll work on all of my devices,

not just iOS, et cetera, et cetera, et cetera. I just liked the story of how long they stuck with it. I just don't think even startup employees would stick around that long today to build something that ended up being iconic. So, yeah, everything develops the whiff of failure much more quickly now to the point that if it hadn't taken off after like a year, people would be like, well, I'm going to go work at that other company that took off after three months. Or if you're Jev from TypeSafe, I mean, it went from zero to trillions of tokens overnight. And then a couple weeks later, open AI at their developer day. And now it's a competing product. That might crush it.

So we'll have to find out. I hope everyone made their money in that short window of time. Before I even had a chance to understand what it does, it's already outdated. I have gone deep on Jeff and there's still about it deep inside its architecture that feels like magic to me. I don't quite understand how it works. But I want to stay on this theme of what's changed, what's different. And during this interview, Justin and Jason, who have been around the technology block a couple times, Talk about how the earlier tech was more about doing things for passion.

This is about the Web 2.0 phase. And I think it's an interesting point of between that and today. So if we could roll the tape at about the 25 minute mark, we'll take a listen. There was no money in the system. So Weblogs, Inc, when we built a gadget or movable type or dig or delicious or Flickr. These were all passion projects by small teams that just wanted to see those products exist. That's what I loved about Silicon Valley in San Francisco during that time because people were just doing it, not because they wanted to make money because it was not a popular thing to do and it was not a good career path. They were doing it because they thought it would be cool.

It was like a cool experiment. Yeah. Yeah. Let's see if we can actually get this website to exist in the world. Yeah. And like, will people do it or interact with each other on this weird website? Will people dig or read at this story? Yeah.

Will people understand what that even means? Exactly. And I love that. I love that sense of exploration. And I feel like, I know this is going to sound like the jaded, like, sad guy, an old guy, but like, I feel like there's a little bit of that that's missing now here in San Francisco because there's a lot of the culture has changed, it's matured and people are coming here for their career now.

You know, they're like, okay, I want to be a, this is a good career move, you know. You know what's interesting about that is, I think it's cyclical because that actually happened in the dot com era. Yeah. Towards the end of it, people are like, okay, I'm graduating. I went to Harvard, I got my MBA and I'm going to go into dot com. We got Pets.com, that's a good move. And everybody floated to Pets.com in every possible iteration.

And the second it got hard when the dot-com boom, they ran back to Goldman Sachs. They ran back to Madison Avenue to get an ad job. And it's, I call it tourist season. Like, we're in the middle of tourist season right now. It's high season. Yeah. And there's a lot of people here who are building companies. I don't know if their heart's really in it.

Lon, what I find really funny about this clip, well, one, history just repeats itself. Apparently, we're always going through the same cycle. But thinking about the financial sector as similar in terms of like value creation as the technology sector, like you can go to Goldman Sands or you can go to Google. Right. Today, the most valuable bank in the world is JPMorgan Chase and they're worth about 890 billion. The next one is the China Construction Bank, 430 billion, or as we now call it a series B. So like it's so funny to see how much these paramounts of prestige, these mountains of money, now seem relatively unglamorous and small. Yeah.

It's, it's, well, what's also interesting there is remember that I recall that era when like the smartest and the best and the brightest graduates from the Ivy League and the top schools were all going into finance. And there was a lot of like hand ringing like all of these people who should be our teachers and our scientists and our great thinkers and philop. They're all just becoming financiers because that's where the money is. You don't hear that today when we, the perception is that our greatest minds are now going into building. things and making things again. So perhaps it's a positive correction. It is a positive correction somewhat. It is not quite as good as taking our best and brightest and maybe putting them out into

the rural schools to really power the next generation of minds that are going to take us to the stars. But think about this. In finance, you're doing deals, you're structuring credit instruments, et cetera. You're really just playing Excel. Right. That's what you're doing. Right. And at least in technology, you're often building something that does scale to an

enormous number of people. Like there's over a billion people that use chat GPT, access to huge user base, Google searches you pay for this. So like Goldman Sachs never gave me a tool. Right. Gabe you Morgan Chase just gives me a fee.

It's impossible to argue that these big AI companies are not doing more. I mean like Grockbot saves me hours of time every day. And like there's no new financial instrument that a guy at Goldman could invent, like a new kind of derivative or whatever

that would actually make my life. Like, I'm sure there is a guy out there who is like, oh, man, I wish there was a new kind of derivative today. But like, for me, I'm definitely benefiting a lot more from people going into tech and creating amazing tools than I am from people coming up with like a new way to invest in a bet about what the market's going to do regarding banana futures or whatever. Yes, yes. But the first part of that clip, when they're talking about the web 2.0 era, which is really when I was starting to pay a lot more attention to startups over finance, I was originally, I wanted to go to

get econ degree and go into banking. That was a plan. Those companies were idealistic, but also, like, silly in a way that modern technology companies are not. I don't know if this is the American dominance theme about defense tech and beat a hurly, burly American.

But, like, delicious had a dumb URL. Do you remember how to spell that with all the dots? DEL. I-I-O-U-S, I believe. Yes. Wow. I thought I was the last person who could do that one on command. Shout out long. And like dig was spelled with two Gs just because, whereas instinct started off with instinct.com. Yeah. There was.

All right. There was, oh, I think it's instinct. dot CO actually that, not to, not to burst your bubble there. But, um, uh, but, uh, instinct.com. Oh, never mind. My, my mistake. Uh, but. And dot co.

The, oh, they got, they got both. They got both. They got both. But the other thing I was going to say is I think there was this idea that nothing mattered other than users. Like that when I for like 2007 was when I first started working it, in tech at Mahalo for a guy named Jason Calacanis. It was my very first startup.

And that was really surprising to be because you, you know, I got out college and I took a few econ classes in grad school. And the idea to my mind was always like a good company. It's like about margins. You know, it's about like how every time you sell a widget, how much are you making versus how much you spent to, put that widget together and the higher that margin, the more widgets you sell.

Like, your thumbs up. You did a good business. And then I started working in tech and nobody ever talked. I never heard one conversation about that, about like spend versus. It was just make number go up. How many people used it? Monthly active users. That was the only metric anybody was checking.

That's all the investors cared about. They'd give you basically unlimited runway if you just kept growing fast enough and adding a bigger footprint. Cosmo.com is always the one I come back to you. Remember, K-O-Z-M-O, and it was, here was the pitch for Cosmo. It was just, it was like grubhub or door dash before those existed, but the hook was there was no minimum order size.

So you could order like a lighter and one muffin, and they would send a guy from somewhere, like from Pink Dot or Circle K or something, a guy would go pick up your muffin and your lighter, and then drive it to you. And I don't even think you had to tip. You would tip like a dollar. Like you didn't even have to tip.

It was like the cheapest thing in the world because all they cared about was one more person today used Cosmo. And we can go back to the investment team and say, look, he looks. It doesn't matter how much it cost us to go bring this guy, this muffin. He used the app. Hey, everybody. I'm here with Heath Paris. He is the founder of Lightfield. They are an AI native CRM.

And it builds and it updates itself in real time based on whatever's in your email. email, your calendar, your Slack, your meeting notes, you know, all the most important stuff. That keeps getting lost. Great to be here. When people go to try Lightfield, do I need to talk to my CTO or my CEO or can I just start using it as an individual and, you know, like bottom up SaaS kind of solution? You can play single player mode or multiplayer mode. How does that work?

So you can absolutely start in a single player mode as long as your CTO or CIO will let you connect your mail and calendar. It's really sort of the backbone connector of Lightfield. So we have some folks who tried out by themselves and then eventually they move their whole company over. Lightfield is model agnostic. So if I want to use open source, if I'm super concerned or I have an on-prem system, or if I have an on-prem system, my CTO and my CIO are very militant about that.

I can switch models and use whatever I want. It's headless. That is correct. You can use any model you want. All of our providers have signed a zero data retention flaw. They're all in America. But it is truly headless. You can use the model that suits you best.

So everybody go check out. ightfield.com. And so I think that led to a lot of just like incredibly dumb nonsense because it was just purely who's going to use it, not does this make actual sense as a company? And today, everything's super frothy around AI, but I do still think people pay more attention to that. Like, does this, what's the business model? Is this ever going to make sense?

the company. I hear people talk about that with Instinct today where they're like, it's always going to be free. Like, yeah, they, yeah, that's what the Noah Chin from Instinct did that invest with the best interview, I believe Monday. And he said it's always going to be free. They're just going to take a cut of transaction. So if you buy book tickets for a plane ride trip on your instinct agent, they'll take like 2% or 3% or something.

And they said they were even boutique who. that are offering them like a 30% cut if their agent makes a reservation. So that's, but people are already like picking this apart. Like, does this make sense? And the app is like you can't even use it yet. It's still waitlisted. It used to be years before people had that conversation. Well, you actually outlined some very good examples of why this has changed over time.

Go back to Cosmo. Think about what they're doing. That's not a pure software product. That involves real world purchasing. It involves people on bikes or cars, moving things around. There's no way to scale that without charging a lot of money. Now, contrast that to the Web 2.0 era when you have Dig and Delicious and web logs and so forth, you could add a million marginal users at the cost of some bandwidth and some CPU time, right?

Now, take that to the AI era. No longer the case. Serving customers has real cost. We are in a compute crunch, and every CPU or GPU cycle that you consume has a real cost to it. And that's probably why companies are monetizing today faster than they were in the Web 2.0. because you kind of have to, or you're going to pull a YouTube and probably not find your Google.

Yeah, if OpenAI existed during the Web 2.0 era, they would never have ads by now. They would just purely be about, we want every single human being in the planet to be using OpenAI. It's only because it costs them so much every time I'm like, what's this plant?

You know, that they have no choice. Whenever I replied, thank you to my Chi-GPD Codex instance, and I realize that I'm using Astrid, and it goes, thinking. I'm like, oh, man, He goes, you're welcome. That was like half my eye. You boiled a puddle.

How dare you? It's okay. I'm going to boil a lot of puddles before I'm dead long. I love it. It's like a shirt and a guy would be wearing on like a cruise ship. Like, I'm going to boil a lot of puddles before I die. All right, we're getting silly. One more clip before we wrap with Justin Khan.

I love this one. It's for about 57 minutes. So towards the end, he's got some great, very practical advice. founders need to recognize their own strengths, is the observation. And boy, howdy, never working for Jason, who's a wonderful manager and founder. But I have at this point now worked for a lot of different founders across a lot of different companies. And this is very applicable advice that I feel like a lot of startup people should take to heart. Let's take a look.

What have you learned now through these four companies, five projects, six projects? You've incubated stuff. Yeah. What do you think is at the core? of creating startups? Well, I mean, I think that being a founder, the most important thing, really, the kind of meta-level important thing

is to know yourself. I think most people do not spend enough time understanding themselves and understanding what are you good at and what are you not good at? And for the things that you're good at, how do I like focus on those things and do more of them?

And for the things you're not good at, how do I either improve or how do I delegate those to someone else? I think most founders hit a wall because they're doing, something, it starts working, and then they don't get out of their own way. You know, they are, they insist, they think because I created the product and it worked, which there's a heavy, a dose of survivorship bias there, I am gonna like, all my decisions

are right. So I want to micromanage kind of all these other parts of my company or I want to make all these decisions or my ego is such that I don't get out of the way of these other people I brought into the company to let them operate. And I feel like that's a very common problem that I see all the time and that's a problem that I've had in my life throughout my life that I've, you know, kind of only, you know, in the last maybe five or six years, really understood deeply. This really stood out for me so much because I do feel like you get a lot of startups where the founder becomes this kind of not just like we think of bottleneck in terms of workflow,

but they become not just a workflow bottleneck, but a creative bottleneck where it's like every decision has to go through this one person and maybe they're not great at making every kind of decision. Maybe they're great at making this handful of decisions, but they should let this hand. handful of decisions be made by the people around them who they brought in. And I just, as I went quick aside, one of other than Jason, of course, one of the other best founders I've ever worked for was Clark Benson over at Rancor, who had a very good sense of this.

He was a product guy, or he is, I shouldn't say was, he's still alive and running Ranker. He's a product guy. And so he had very good instincts about how the product should work, how the app should work, how it should be designed, UX, UI. None of us would really question him there. But then when it came to like media, especially, which is what I was doing, the editorial side, he would really let the team experiment and explore, try different things on social media, try different things on YouTube.

You guys know that landscape so well, just go. And man, it worked really well. And we ended up being able to grow that media business a lot because Clark kind of got out of our way and understood, I'm going to be over here worrying about the app and the product. You guys go worry about that stuff. This interview, and longer than just a clip we just played, I really do recommend that anyone listening to this who's enjoyed it, to go listen to the whole thing. I'm really glad that I did.

Justin talks a lot about, like, being aware of himself and understanding what happiness is and creating distance from his emotions, just comes across as a very, very mature person who's done a lot of, like, experience and a lot of learning from that experience. But in this case, what I found very interesting law on was a implicit rejection of founder mode. Now, not really, because founder mode actually came out as an idea later on than this interview was taped. But here we have an executive. who's, you know, taking a company from zero to a billion dollar exit.

So they have done the thing, the most impressive thing you can do in Silicon Valley, talking about shared responsibility and letting, you know, smart people run their own game. And that was the conventional wisdom that I grew up in in startup land. And today, very much so it's now pivoted towards founder mode, you know, one person carrying the torch, providing the direction, setting the North Star, and being involved in the details. And that's a change. And so I know all tech is bundling or unbundling or cycles and recycles. but this did surprise me as tonally different.

I think it is, it's a, it's a different approach. And it's a, it's a better, it's a better approach in a lot of ways that doesn't necessarily assume one person has to be the keeper of the flame for like everything that a startup does. They can sort of set the larger vision and then let sort of everyone, everyone go. Really good flashback coming up today. This one comes to us from May 23. It's episode 1735 of this week in startups.

and your guest is Brian Chesky, the co-founder of Airbnb. This episode dropped the same week that Airbnb rolled out Airbnb rooms, a return to the company's original, rent out your spare room concept. So that was why co-founder and CEO, Brian Chesky, had agreed to be on the show. To go back to 2008, seven investors passed on the chance to buy 10% of Airbnb for $150,000.

The founders famously kept the lights on by, selling novelty politically themed cereals. We're going to talk about that. Also in this conversation, Brian talks about early rejections, obsession with the customer, how Airbnb is organized, and where he thought AI was going to take the hospitality industry. We're going to look at some of those projections. So let's just dive in, no need for any further intro.

The first thing, Alex, I wanted to talk about right away in the interview, Brian starts telling the story about how investors. did not like the idea for Airbnb. Now we accept Airbnb as like, obviously that's a great business. I use Airbnb when I went to Rome earlier this summer. We stayed in an Airbnb. It was a wonderful experience.

But back in the aughts, when Brian was first pitching this around, people were a lot less receptive. Let's take a look at that clip. Everybody, when they see you saying, oh my God, I love Airbnb, how do you reconcile that as an entrepreneur? We came up, you know,

obviously I know Joe told the founding story, But I'll tell a little bit of the there's a founding story and then Jason is a story of all the rejection The first time we came with the idea it was October 2007 It was for a design conference that was coming to San Francisco There was an after party at the Fairmont hotel we went to the fair my hotel the first person I told about the idea He looks at me of the stray face. He said Brian I said yes He goes I hope that's not the only idea you're working on That's what he said to me and this is like a design luminary in our industry and

And that was kind of the general sentiment. I remember in January, February, 2008, Joe and I were living in San Francisco's an apartment. And we had a roommate named Phil. And Phil worked for this company called Justin.tv. Justin. That TV was a precursor to Twitch and it was funded by this program called Wight Combinator. Now I didn't know anything about Wight Combinator. I didn't know anything about Silicon Valley.

I didn't even know what angels were. Somebody once told me there are these people called Angels. And I said, oh my God, this person believes in angels. So I was really naive. I didn't know anything about Silicon Value as a designer by training. And Michael said, I can introduce you to these angel investors. And so Michael introduced us to like 10 to 20 angel investors. Now, at the time, Jason, we were trying to raise $150,000 at a $1.5 million post money valuation.

So for $150,000, you could have owned 10% of Airbnb. And the majority of them didn't even reply to the email. I actually end up publishing a bunch of the projection emails, but many people said like, this isn't a good idea. Many people said like travel, we're not like excited about travel. I remember one investor said, we love everything but you and your idea. In other words, we don't, I'm like, well, everything else is good. I was like, so I thought of myself, wait, what else is there? There's me, there's the idea. Yeah, I guess there's the name, but they don't like the day by there.

So what they meant by they like to everything, but us and the idea was it was three founders, one, one software engineer and two designers. And they thought, well, you have like two, too many designers in your founding team. You know, like people just associated designers as like non-technical and therefore, like, maybe not adding value. And I think they, I always felt like Joe and I are ability as being designers was actually part of the secret saucy from me because it was not a pure technology problem. And also people said this is crazy.

Strangers never say other strangers. I mean, I will admit, I had, I was already working for Jason when Airbnb was new. and we talked about it, probably on a podcast episode very much like this. And that was also my reaction was like, well, that sounds dangerous. Like, how do you, I'm going to go walk into some stranger's house. I'm going to leave my keys and some strangers going to stay in my room while I'm gone. Like, it just feels crazy when you're not using it. And then when you're using it, like now, I don't even think about, when I go to an Airbnb,

I don't even think about this as somebody else's home. It's just my home for the week. No. Yeah, yeah, yeah. but I think there is something to be several generations growing up in a post-stranger danger world. I mean, think about the PSA's that you were getting when you were a kid. I mean, it was every person's going to try to offer you candy and stuff you into a van and take you off to, you know, Tuscaloosa or Timbuktu. And not really the case.

As it turns out, most people are pretty cool. Most people are fine and you can get along just okay. But for Funlon, guess what 10% of Airbnb is worth today? Oh, my God. I don't want to think about it. It's big, I guess. like $500 million. $9.25 billion.

Oh, my God. It's insane. How is Airbnb worth that much money? That's crazy. Airbnb is currently worth $92.5 billion. It's because it became a huge business. Those investors were not just wrong. They were phantasmagorically wrong because it turns out this is popular

the world around. Now, what I want to say is that Airbnb has, over time, changed a bit. If you think back to the early days, especially if you knew couch surfing from before Airbnb, it was a pretty hostile-ish experience, if you will. It was not five-star. It was not Instagramable, though we didn't have

Instagram back then. But today I do feel like Airbns are much more upscale. You went to a whole cabinet, et cetera. Lovely. But I mean, back then, it was, it was great. Yeah, it was more variable. Like, in the early days of using Airbnb, sometimes you show up at the place that it was amazing. But then other times you'd be like,

this doesn't really look exactly like the photo, it was a little bit more of a shot in the dark. Whereas I think today, they've got it basically locked down, you know from the Airbnb page, if it's got like a 4.8 or up rating and you look at the pictures and it looks nice. Like you've been pretty much confident in what kind of place you're going to be getting. For sure, but I will, though, you have to want the quirk. Like if you want a boring room with like a bed and a shower, then just go to a hotel. If you want anything else with the more personal human touch, then I think Airbnb's are great.

No hotel has ever given me. free weed gummies in Texas. But an Airbnb has. I actually, I feel like it's when you want, when you're just staying somewhere and you just want convenience and it's just like, like hotel, obviously, when you're going to be busy

or whatever. But if you're going somewhere and you want to get like a feel for the place that you're going, then you want an Airbnb because you're staying in a neighborhood and there's like a person who's going to leave you a binder that's like, go to this pizza place. It's really good. You know, like that's what you want when you're like it's like a great vacation option.

I don't know if it was like business travel. I'm just like put me at a hotel. I'm putting it on the ramp card. But for Rome, it was perfect. I would very much recommend an Airbnb. I'm so startup pill. You said Rome,

I thought R.O.A.M. Howard startup? No. Wow. All right. Rome the Eternal city, Alex. Rome the Eternal city. The place where they had the big circle they used to flood and have ship battles

inside the town. Correct. Gladiator 2. Historically, very accurate. I'm sure. So our next clip for Brian, he started. You could at the tail end of that other clip, he started to tell it. But we want to get the full story in here because it's such a good iconic startup story is when he talks about it. I tease this a little intro, how they funded the company in part by selling parody political cereals during the 2008 DNC and R&C. Here's Brian.

So we go to White Combinator. We go to the interview. It's exactly what I expected. It's Paul. It's Jessica, it's Trevor Brackwell, and it's, I think, Robert Morris. And they're all like, basically, they all ask us questions at same time, like, all four of them. And I was like, totally bewildered. And the first question Paul Graham asked me is, people are actually doing this. And I said, yes. His second question was, what's wrong with them? And the interview at that point went downhill from there. He, midway through the interview, he basically tried to get us to create stripe. He's like, you should create this like payments company or like an online bank or something. And,

And we're like, that seems like a really good idea, but like we have an idea. And we're about to leave the interview. It doesn't seem like it's going well. And Joe takes out a box of Obama owes. And he hands it to Paul Graham. And Paul Graham's like, let's like, he just got a novelty gift. He's like, all right, thank you. Yeah.

And we go, what's the story? He goes, this how we fund the company. And he goes, what do you mean? He said, well, we told him the story, how we like made the cereal boxes. And then the way we sold the serial boxes, we mailed them to reporters. And they put them their news. room desk and everyone buy them. And he said something like, I guess if you can convince people to pay $40 for a $4 box of cereal, then maybe you can convince at least some people

to stay in each other's homes. And he ended up admitting us to Y Combinator. And I guess we've never looked back since. I also want to give credit to Jessica because I think Jessica thought, I remember Paul and Jessica said later, they thought we were like cockroaches. And I think they mean it in a good way that it was like an investment in nuclear winter. And in a nuclear Holocaust, us. The only thing that survives are cockroaches. And the only thing that would survive an investment Yuka Winter would be the founders of Airbnb. We said, we won't die no matter what happens. And so they basically funded us because we seen Brazilian, unkillable. And like, even if the fundraising market was dried up, we would just go on. So a couple things. One, I do love hearing about a startup

struggling in their Y Combinator interview because it's kind of a famous test of time, a coming of age moment. You become a man when you go through this. But it just goes to show that not everyone's perfect and people still make mistakes, even the vaunted founders of YC. But Lon, I have to say it does track with what I've learned about startups to know that the grit matters more than last week's traction. And if you do know that a founder won't give up, well, that's probably half the battle of backing someone right there. Yeah.

And I mean, I think that, you know, it's just like I was saying a little bit before about the, you know, that you're funding the founder, not the concept or the company. That has always been kind of a truism. And I think this is a real, you know, case where you can see that is, like, it's not, it is resiliency, but it's also the creativity. And it's the resourcefulness, I think, that Paul Graham was noticing from these guys that, like, they're just doing business. They're just making observations. They're seeing angles. They're trying things.

They're not afraid to, like, do something kind of goofy to try to get attention to make something happen. And I think that, seeing that in a founder has got to be really encouraging as an investor, that it's not some person who's just going to sit at their desk head down and work on their one project single-mindedly. There is like head on a swivel looking for opportunities. Well, there's something about being a founder that is just sheer persistence and not letting anything stop you. And so if you find some people that I've managed to somehow hack at the same time, the political cycle, the news, media, and their funding crisis in one fell swoop. to do it in a way that makes them look kind of cool in punk rock, I mean, seems like a pretty

obvious thing to fund if you're an early stage venture capitalist because you want people, and we've talked about this on the show, ad nauseum over the years, who are willing to look from a different angle or perhaps even a little bit outside the rules. Like, I don't think you're actually supposed to use Obama's name and reuse cereal and probably resell it and blah, blah, blah, but no one really cares, so go for it. You know, off you go. There's tremendous value, and this has always been true from the aughts to the present day. a person who has a good sense of like how to get something to go viral, like how to communicate an idea that a lot of people are going to be interested in, there is always value in being

one of those people. If you have that, like we were talking about instinct moments ago, the viral assistant app, that that's is the same thing. It's like the entire strength of it is how to create something, how to kind of have a comment, how to make a statement that is repeatable and shareable and encourages people to, you know, what they now call, like stop the scroll. Like, if you are that kind of person, that has been a marketable, useful skill

for 25 years plus now in the internet age. And I think especially as attention gets more and more fragmented across platforms and shorter and shorter form content, being able to stop the scroll is going to become, I think, like a core competency of pretty much any founder. I mean, think about how we've seen launch videos of startups go from being super dorky nerds staring at a camera. Shout out Justin Kahn,

to now the slickly produced, you know, quasi-Marvels Avengers-style films that people are not putting it at the cost of like 50 to 100K. Yeah. I mean, Airbnb was trying to raise 150 total. Yeah. I mean, we've seen it even extend beyond tech. I mean, like entertainment media now is also about like capturing those viral moments and like like the like backrooms, you know, like backrooms

was a viral moment long before it was like a cultural phenomenon. It grew in the telling because it was like originally like a person with a good sense of like this weird liminal hallway video is going to get clicks. And so I think we're seeing that across the board. That is a marketable skill in every industry in 2026. Yeah, if you can hack attention, pretty much you can work in media, PR, venture, startups, tech, law, business.

I mean, like, anything about coal mining, you're going to be fine. The back room is one. It's an interesting analogy to the line. Can I just go off script here and just ask you if you liked that film? I did. I do like the back. Okay. Okay.

Why? I will tell you this. Or to quote Paul Graham, what's wrong with you? I think as a movie, you can tell it's a first time filmmaker. It's a young guy who's kind of figuring out narrative. I think there are some like dead ends narratively in it and you could maybe make that same story a little cleaner. Having said that, I think the backrooms themselves are extremely compelling.

And I have since YouTube, like the idea of being trapped in this like, sort of surreal, repeatable, like, like quasi-realistic environment. And I think that he invented, like, the original backroom shorts are from before the AI era. But I feel like in the AI era, it feels very, like, relevant, this, like, endlessly repeating environment

that, like, degrades in realism over time and, like, the text becomes unreadable and things start bleeding in. It's like a glitch burst in some ways that we become more, accustomed to in the AI era. And I think that makes it sort of inherently just very compelling. I think it's a great setting more than a great movie. But I did enjoy the movie.

Repeating text, quick release cycle, slowly losing your sanity. Are you talking about backrooms, the movie, or AI in 2020? I mean, I think that's what's in part so potent about it, is that without it, like, the best AI metaphors are not like super on the nose. They're like, this is what interacting with AI. sort of feels like, and I feel like backrooms really captures that specifically. And so the other one, like that is Pluribus, that Apple TV. It's not about AI, but it's like Carol's frustrating conversations with the hive mind are reminiscent in some ways of what it's like to interact with AI.

I can't wait until my kids are older and I can catch up on like years of television and movies that I've missed. Late nights, make time to catch up on Pluribus over your late nights. My wife steals the television and watches HGTV, and no offense, that's not what I want to catch up on. So I just let her have it. But on the point about AI and what's real and what's not, I want to take us all the way to the 59-minute mark of this video because we get into a very interesting conversation about AI and what's going to happen, where it's harder to tell what's real.

I want to bring up this clip and the one that comes after it, not because this is the most technical, practical, practical startup advice ever, but because I think today we spend so time thinking about AI. It's like the thing. But in 2020, it was a smaller industry. It was not the entire world. And it's good to see how smart, engaged minds consider the technology before it became the only thing we can talk about.

And so that's where I think it goes. And I think that we don't have a search problem in the future. We have a matching problem. So we are going to use AI to match you to whatever you want. And I think that the other thing is I think we're going to be really good. hopefully identity authentication. You know, the problem with AI is, what's the first word in AI artificial? The biggest risk before machines come after us is they become us.

And we can't discern who is the machine and who is not. And I think that we're already seeing that. I mean, when that Pope photo circulated, I thought that was a real photo at first. I did too. I was like, that's weird. The Pope is going for it. Yeah, that was like pretty awesome. He's got Balenciaga and go for him.

But that's today's technology. So we're about to live in a world where someone can sound like me, they can behave like me. And as we spend more and more time online, it's going to be harder to discern what is real and what's not real. So I think in the age of artificial intelligence, the other thing people want is authenticity. And authenticity is whatever is real and whatever's authenticated. So I think our brand is kind of authenticity. We're not going to ever be the most digitally immersive company.

That's going to be social media or entertainment. Our value is really the physical world. We get you online, offline with people. different from you all over the world. And that comes with starting with knowing who you are, authenticing your identity. Lon, I'm impressed by how prescient that was, although it was only a couple of years in the past, but still, I don't think I would have had that clear perspective of what's coming. But his point about AI can sound like you, look like you, we are seeing AI models in the voice category today advance to the point to which you can feed them a very

small snippet of audio and ape someone's entire speech patterns. We're seeing people create, I think we used to call them deep fakes. Now we just call them realistic videos of people. You can actually have avatars kind of like stitched over yourself digitally. And we're seeing people's parents get called by fake callers that sound like their children. So everything he called for is coming true. And I do think that he's correct that we do want more authenticity. I just think it's funny that he was talking about the physical world in a brick and mortar sense.

But today when we talk about AI, the physical world, we're mostly talking about robots or embodied AI, if you will. I mean, I think this prediction in particular, we've been hearing since the early chat GPT era, which is like, this is going to lead to a revival in in-person experiences, authentic experiences, human experiences, like live music is going to make a huge comeback. And we're all going to start hanging out together and going to block parties more. And like, I really like that sort of take in like 2021, 2022. Here in 2026, I feel like that's not. That is not. So far, that has not really seemed like it happens.

We put up a lot of people have become much more hostile to AI outputs where they insist on, I only want to read substacks that are 100% written by humans. I don't want slop on my timeline. Like, there's definitely a thirst for authentic content, but I have not seen like, I'm going to go to five extra concerts this summer, or I'm going to host more dinner parties. Maybe I'm just not popular. But I don't feel like we've seen that as a society yet. You're absolutely right.

And here's the problem. I think AI and the general erosion of trustable media would lead to more people being outside more and together. But the problem is we're all working more than ever to keep up with AI. So no one has time. Right. Like in five years, maybe, we'll all work a little bit less because agents will take on so much of our workload that we have more spare time to take a walk during lunch. Go see a friend. Long, I could fly down and see you and take you out for barbecue.

I love that. But right now, you and I are both so far behind at work that there's not the capability to do it. And so I think that's the real sticking point of AI for me right now. You and I are both AI users. We're fans, you know, we're not far from haters. But I do think that all it has done for a lot of people like ourselves is ratchet up the intensity and duration of work. And that is not quite what I had in mind when we got this amazing advancement in AI. Yeah, it is.

one of the rare points of total agreement between myself and Jason's all-in bestie David Sacks. We don't agree. We don't agree very much. But one of the things we do agree on is that rather than costing people their jobs, at least so far, what it seems like AI has done has made everybody a lot more productive at their job and therefore a lot busier at their job. So you're getting three times as much done. You're maybe working twice as much, but you're not being replaced. So that, that, on the plus side, you still have a job. It's just harder now. Can I, can I challenge that ever so slightly? Please. All right. So I was here at the, the good ship launched a long, a long period of time. Several years in podcasting time,

you're doing three shows a week. We get a lot of work together. So I've been through a lot of iterations of the show. And one thing that was a consistent thing during my time was, I wrote the show most, most of the time. And it was a lot of work. Prepare the docket, We don't have new ideas, we changed the format. It was this evolving living process. And now long, that's no longer the case. And we've done the modern thing and allowed really agents to write a chunk of it for us. Because it's just collecting information and putting it down on the page.

And I don't think you would go hire someone to prepare the show notes in the what we did when I was when I was brought on. And I think that is a change in the example of less job. I mean, yes, but there's a guy. his name is Zach. He works for us now. And he is the one who's making the dot. Like I used to write the docket by hand. It was my job.

It took up a big chunk of my day. You would do it for different shows than me. And, you know, I spent three hours putting this document together by hand with all the pieces of research or whatever. So Zach does that now. I'm freed up. I'm working on the substacks or whatever.

Check it out. TWI Startups. Subsdack.com. Best headlines in the game. No one outshines lawn in the pun game. Read everything he writes. Makes me smile every week. You're very kind.

But so, you know, Zach is now doing that job. Maybe he's spending an hour or less a day on it. But it's the job is now not only putting the stories. You still have to pick the stories. Claude's not good at that. Still need a human to curate what we're going to talk about. You give the stories to Claude. There's a Claude skill that turns those stories

into the docket, Zach had to write and brainstorm that and create that skill. And then every time we have feedback on the docket, he needs to go back and fine tune and improve that skill. And then he's the human in the loop. So whatever the skill comes out with, he's got to go through line by line and be like, didn't nail this, nailed this, didn't nail this. It is probably like 20% less work over the course of the week. but it is still a job.

And I don't think in two weeks we're going to iterate our way out of that job. I think that's going to be Zach's job for at least the next few months. And then it'll be on to the next skill and the next skill and the next workflow that he can improve. And maybe there is, here's what I would say. In three years, in five years, maybe we'll reach a point where these models have gotten so good that we won't need a Zach anymore to fine-tune them. we won't need a micro one with all their experts to help the frontier labs make the models better, or a TACE Labs and all their experts to help make the models better designers,

or maybe we'll be done with that. I don't know, maybe. And maybe then we'll be talking about, okay, now that all the architects and accountants and lawyers train the models to do their jobs for them, now nobody has a job. I don't know if that's going to happen about it. I don't buy the... But I don't think that's happening yet. No, no, I'm just seeing little bits of work get automated.

Like, just to put another example in this, I write a lot. And I have created some skills inside by a codex instance that help me do what I call pair editing. So I abuse commas and I need someone to go through my drafts and just remove the commas. Like, this is a bit a problem. My entire life is a writer. I know it. But I never want to sound like AI because I do all my own writing. So I've made skills that edit my work, leaving the tone, the jokes, the way that I

right, but just solving the grammatical issues that are, that pop up here and there. When I was at TechCrunch and I started off, I had to take my writing and bring it to a human and have them do this. And here's the thing that blows my mind. You know what's really good, Codex? We'll just chat GPTS, I suppose. It's incredible.

And so I do think that we will see a slow erosion of previously high-touch labor categories that are in the knowledge of a game that will change the employment system, structure. It just isn't going to happen in 18 months, like a Thunder Club like everyone thought. But I wouldn't go all the way the Sox of view that, you know, only more jobs, no job disruption. I would say more jobs net long term, medium term, turn. And I think that's probably a great tradeoff for what we're going to get down the road.

I'm not worried about this. I'm just chewing on as we go through it. Yeah. No, I mean, I think there's some wisdom to what you're saying. And we're definitely going to see some movement in general, especially like I said, in the two, three, four year time range when all of the expertise and fine-tuning work that we're doing now on models is baked in and done. And like maybe we don't need to keep having accountants look at the output and be like, oh, mess that part up again. The model will just know and benefit from their expertise and get that,

get better to that point that we don't need them anymore. I think that is definitely, I think that is definitely possible. Now, one more clip before we go, sticking to the theme of Chesky being a little bit early to things that are now very obvious. We're going to play a clip of him talking about what's going on in the realm of AI software development. Now, of course, here in 2006, we are deep into our Claude Code era. Everyone uses an agent to write code. Sure, but back in 2023, very much not the case, but here is what Mr. Chesky had to say then. On the side of a friend, few told me,

hey, you got to try this place in San Francisco, Cafe, Okinawa that I've been going to to get Tonkatsu sandwiches. I told like three people about this. And because I love these when I go to Japan. they're just everybody goes there and they're like oh yeah you're right that was great and now they spread to the next person and that kind of word of mouth i think like this ai is going to find little things like that where it didn't know that this was something that you would be delighted by totally and this is such an opportunity i think it's amazing and by the way i think i heard another one of your podcasts you sent something that really resonates that i think it's worth calling out the cost to develop software's about to go down and maybe a good analogy jason would be like the

camera. 130 years ago, very few people could operate a camera. And so therefore, you had to be a specialist to take a photo and there weren't a lot of photos. Suddenly, computer programming is just basically telling a computer to do something in its language. AI means you can now tell computer to do something in your language. The moment you don't have to learn a different language, you can use your natural language to English. Suddenly, kind of everyone in a way can be a programmer like everyone's a photographer. You don't have to actually have a skill. Now, there will be programming skills, but suddenly anyone can do that. Now when it's possible, there's going to be so much more software.

We're going to live in abundance of software. I think in a podcast you reference like a ski instructor or a ski company, like a niche business that couldn't have developed hired great software engineers, but now they can the cool, the best engineer you've ever had developing their app. And so suddenly the skills start to different. We have to understand skiing, human psychology, taste, design, and how to operate and have a conversation with that tool but you don't have to necessarily build the infrastructure just like we never built servers because we had a ws but if we launched five years earlier we would have been really good

at like kind of building out servers rack and stacking servers and your first half a million would have went towards a co-location facility yeah and now it's like you're just going to talk to this thing and it's going to make your software for you so i think this is going to create i think this is going to create millions of startups i think that like entrepreneurship is going to be a boon i think there's going to be millions more I think anyone can essentially do the equivalent of what software engineering only allowed you to do only five years ago. It's going to be awesome for many people. It's going to be wildly disruptive for others.

Albert Einstein used to have as saying the best way to keep her balance in a bicycle is to keep moving. And I think the best way to keep her balance in the world of AI is to keep moving forward and just adopt the tools. Don't fight it. To put some context on this lot, I looked it up in mid-20203. The leading AI models were GPT4, Claude 2. and Lama 2. So this was way before they were good at coding, but he was able to see clearly three years ago

where this was going to go. And I think they were dead on. Here's the question, though, for you. Because you are deep in the launch ecosystem, accelerators, you know, founder you, et cetera. Sure. So Chesky was saying we're going to see millions of new startups. And on one hand,

I do think we're seeing a lot of experimentation in the market, but it doesn't feel like we're seeing that many more new companies come up. So I'm just curious what you're seeing. and if that prediction from this Airbnb founder has borne out? I, you know, I'm always a little dubious of this prediction, especially because a lot of tech people tend to use it to advance the argument that, like, as like a pushback against being anti-AI. Well, like, oh, but it's going to lead to massive job disruption.

Like, well, those people are just going to start companies. And it's like, well, everybody can't run their own. Like some people need to work for other people's companies. So, like, you need users who have their own money from other jobs where they work to buy. Like, I don't know if a capitalist society where literally every single person owns their own company is sort of functional. I don't think that's the way our society functions. So it always seems like it's missing a little gap there in some ways. But I do feel like we are seeing an explosion in entrepreneurship.

We are seeing a lot of new companies. and we're seeing a lot of people who have started one company are now starting three companies or are now starting five companies. And so we're seeing that grow exponentially, where if you are already entrepreneurially minded and a builder, you are getting great utility out of this and using it to do a ton of projects. I think what's missing is the non-entrepreneur people. The prediction I think that Chesky and others are making is that people who have never started a company who maybe had ideas, but always like, I don't want to start a square space.

That's too complicated. Now they'll be like, oh, but with AI, I could start that bagel shop I've always wanted or I could open my own dog grooming business like I always wanted. And I'm not sure we're seeing that yet. Maybe it's just that the word hasn't spread to those communities and those people that like AI makes it that much easier to start your own company. But like, I think we're seeing a lot of of people who are already in the, I'm going to one day start a company mindset using these tools to do it faster and easier. I just don't know if that's become universalized yet. Yeah. On your point about not everyone can become a founder. I think it was Yo Gotti, who said that everybody wants to be a general, but nobody wants to be a soldier. And I think

that does apply here that, you know, you can't have an entirely top-heavy organization in the economy. People have to have jobs versus CEO titles. The thing that I've been thinking a lot about, and just pairing off of what Chesky was saying is that now that you can program in your natural language, your normal language, people are building a lot more out of sight. One thing that I keep being shocked by is when my non-technical friends,

people that I've worked for off and on for a decade plus, say things like, oh yeah, so I was using Claude yesterday, and I made this app, and I'm like, what? You? You couldn't even open Google Docs five years ago. What are you doing? And they're making their own stuff,

things that they need, little things. And so I'm always trying to figure out, like if we're not seeing this explosion of startups that we expected, are we seeing instead people with jobs essentially making micros software companies inside their own role for themselves? And that's maybe where we're seeing more of this impact and rubber hitting the road and so forth. I'll tell you what else I think we're also seeing is people who are already doing things like they're a content creator making substack posts or they are a dog groomer with like a van that they drive around.

Like, I think a lot of people who are already doing that are becoming much more effective with the help of AI tools. Like, one thing that I am always shocked by, as a person who's been doing this job for 15, close to 20 years now, one of the massive headaches and also advantages like 15 years ago that I had was I was pretty good with Google Analytics. Like, I knew how Google Analytics worked and I could look at the pages and make comparisons to be like, oh, we need to do this or this page is doing really. really well, so we should make three more pages like that.

Like, I learned how to do that. And that was a very, like, not everybody had that skill who was making blog pages or whatever. And so that could put me at an advantage. And today, you could just download your Google Analytics or your YouTube analytics or your X analytics. I know because I do this all the time. And you could just give it to Claude or Grockbot or whatever and say, what kind of post should I be making or what can you tell me big?

You don't have to know one thing about how any of it works. You could even go, I did this once too. You could go to Claude and say, I have a YouTube channel. I want you to tell me what kinds of videos I should be making. What should I do in analytics? And it will tell you exactly what charts to pull for it and what metrics it wants to look at to make you a full report. It takes all the inherited knowledge out. So nobody has to show you how to do that stuff.

And I'm sure those kinds of processes are making people much more powerful at the task they were maybe already doing it. Yes, as long as someone's already done it and it has been ingested and is, you know, chewable by AI, then sure. I'm always so excited about progress. And then I think about that. I'm like, well, you know, I could just prompt my whole newsletter. And I bet it would be 70% of the day. I will tell you, as a decide, and then what we have to wrap.

So I'm doing these substacks for launch now. We have a this week in VC one. We have this week in startups and this week in AI. 100% human written. I write every single word. I don't let AI do any writing for me, but I do have GrockBots helping me out at every step of preparing those newsletters.

And I am able to, I'm so much more. I would never have been able to churn out three well-researched, fairly lengthy substack posts in a day five years ago, but I can do it today. Help people out, Lon. What is GrockBott doing for? Is it clicking the buttons to publish them? Is it doing research for you, pulling facts?

It's that one. So I'll give you a great example. In the This Weekend AI newsletter, this weekinAI. com, if you want to check it out for yourself, we do a demo or a model comparison or a tutorial every day. Like, what is there like a hot piece of content about AI that's hands-on and practical that we can share every day? The old workflow would have been me loading up X and literally scrolling

my for you, looking for like, what are people talking about today? Let me spend 15, 20 minutes just getting a good sense or that and then go to YouTube and spend another 10 minutes scrolling over there. What are the, oh, Sonnet 5.5 demos are the thing today. Now I have the XAPI tied into a grok bot and I just tell the guy. I already have this process figured out because I do it every day. So I just come in in the morning. It doesn't even need to be prompted.

It just tells me, here are the big trending demos. Here are the big trending model walkthroughs. And it doesn't even just pull the links. It actually summarizes for me. So it's like Lon Sonnet 5.5 demos are super hot today. I found three times as many of those and everything. It's doing a research report for me. So, and then I can go through that.

I could be like, what part of the transcript of this YouTube video does this come from? It'll give me the time codes. I could take that. I could give that to my clip bot. And I could say make a video of this video between this and this time code. And it just makes it. So all the tiki tax. stuff is done and I just get to focus on the actual right.

No, I think it's a brilliant example at all. I'll leave everyone with one more AI hack while we're sitting here. Have you ever run into a very annoying paywall because you don't subscribe to that particular website? Well, go to your favorite friendly chat GPT or GROC agent and say, hey, can you find me a gift link of this article link? It will go on to X and find you one. Yeah.

Perfectly legal, ethical, and a great way to avoid the paywall. I'll give you one other producing tip with Brockbot. It's really good at surfacing people's contact info. So if you don't want to pay for like a service, you know, there are those services where you pay them 20 bucks a month. And if you're like a salesman, it'll help you find anybody's email address or just ask Grokbot and it'll get really good at it. And now I could just, if there's somebody we want to book, I can say, find me some contact information. but it won't even just, it'll figure out their assistance name and look up their contact information.

It'll figure out their press context name and find them on X or LinkedIn. Like, it'll go three levels deep and spend an out, like what would have taken me an hour of research and I can zero in on that person. And I'll tell you, we're very excited she's going to be coming up on one of the shows. Dr. Fay-Fa-Lee of World Labs. I've been trying to get her booked on a Jason Calacanus podcast for many, many months, with the help of Grockbot, I was able to find an email address I hadn't found before, got right through.

Surely you mean AMD's, Fafing. I do. World Labs is a division of AMD, exactly. Yeah, yeah, yeah, yeah. She's the World Lab CEO, an AMD company. Chief Scientist of A&D is her new title, actually, I believe. If you don't know what we're talking about, her startup was sold to AMD in an $8.2 billion all-stock deal, I'm sure it was on the show.

It was. But that's what we're joking about. All right. I think that does it. I think we've talked more than enough about Brian Chesky. We enjoy getting together and hanging out, chewing the fat, a little bit. So thank you to Alex Wilhelm for being here. I am Lon Harris.

Check out our substacks. Please, TWI Startups.substack.com. Go take a look. Give it a read. Give us a subscribe. And we will see you next time on this week at startups.