How to Scale a Profitable Agency with 0 Employees (Using AI Agents)

- June 6, 2025 (9 months ago) • 01:02:50

Transcript

Start Time	Speaker	Text
00:00	Shaan Puri	My brain kind of exploded a little bit. I was like, "Oh—so any business, you can flip the sales model on its head." Absolutely. You could basically, instead of reaching out saying, "Hey, would you like me to maybe do this for you if you pay me and hire me?" say, "Check this out — I made this for you." </FormattedResponse>
00:13	Matt Mazzeo	Yes.
00:14	Shaan Puri	"You want to work with me. That's right. Obviously, that's going to be a better sales pitch. But now AI made that scalable. That was not scalable." </FormattedResponse>
00:19	Matt Mazzeo	"That's exactly right."
00:25	Shaan Puri	If I think about Matt Mazzy, I think about the word early. So, whether it was investing in companies early—you were part of a portfolio that had Uber, had Twitter, had tons of hits—you were early at CAA to the internet trend. Like you're saying, the first one to sign, kind of like YouTuber talent to a traditional Hollywood agency, or getting Hollywood talent to put stuff online. Yeah. Now you're early-ish and early on the AI train. You've been texting me photos of myself and my house but redesigned using a, like, AI-style agent… [sentence cut off]
01:04	Matt Mazzeo	We talk about this. Well, it starts with my wife making fun of me. She's like, "Your entire closet is black and white. You don't buy clothes ever—you default to the same exact things. It's laughable." So we took a few months and lived in Japan and Korea. Everyone on every corner is the most stylish person you've ever seen; they're all put together. I'm like, "What is it? They all look like they're out of a magazine and I look like a schlub." I realized that I've just never taken the time to do it. So I'm like, fine—I'm going to try to train this AI model, O3, which is spectacular now and has all the vision modeling. I'm going to try to train it like a Korean color theorist, and that's where I'm...
01:40	Shaan Puri	I'm going to start.
01:40	Matt Mazzeo	Have you ever heard?
01:40	Shaan Puri	I've seen on TikTok the idea of color theory. They hold up a bunch of swatches and say things like, "This one brings out your skin" and "This one clashes—avoid these." It's like an allergy test for fashion. That's right. Perfect example.
01:53	Matt Mazzeo	And people spend hundreds, even thousands of dollars doing it. If you get a stylist in the U.S., it's hundreds or thousands of dollars. It's one of those classic things rich people do that eventually AI will open to everybody. I'm gonna see if the vision models are good enough. So I run the color theory on myself and my wife, and it tells me [to me and my wife], "Okay, you're a soft autumn," — which, you know, I was gonna say I appreciate that for what it's worth — "you're a dark winter," which I think is super, super interesting. For me, it was like: here are your primary colors. Then I was like, okay, great — what do I need to avoid? It gives me a list of colors I shouldn't wear and tells me how to put them together. As I dig further into it I'm like, "Oh, can you give me a capsule collection that I could do? Give me brands that I like. Here's my measurements, here's brands that I wear, here's how I think about it." My wife and I are doing this and we start building it for each other, and then we start productizing it for friends.
02:45	Shaan Puri	You sent it to me and you were like, "You basically sent me an AI version of me." It had me dressed way better than I actually dress. Then it was like, "Yeah, you can just push a button and it'll buy this outfit."
02:56	Matt Mazzeo	That's where we're going to. Yeah. So it's not there yet, and this is part of why people laugh at the "AI rapper." But I actually think there's a huge moment in time to just be the rapper — you be the rapper, right — and sell the service, sell the outcome, and do the work for somebody before they even ask for it. I didn't ask you for selfies; I just found selfies of you. I didn't ask you for your fashion sense or your colors — do you want a color theory? I just ran color theory on you, and I gave you a final output: "Here's a bunch of looks that you're going to look great in," and I got to the endpoint of what you actually want.
03:27	Shaan Puri	So, let's put the style thing over here for...
03:30	Matt Mazzeo	Yeah.
03:30	Shaan Puri	This is a golden nugget to me. Number one: this is the one I picked up from you in the last few months, which is this key insight — AI is not just a product; AI can actually be the go-to-market. What you're doing is very, very smart. I think you have a few ideas that fall into this bucket, but you're like, "Alright, what if instead of just using AI to do the job, I use AI to sell the customer, to get the customer on board?" What you're doing is basically saying: instead of the customer having to find you, decide if they want to try it, sign up, pay you, and then you do the work, you're saying, "Let me just grab pictures of you… let me grab pictures of your house." Yeah — you got my house off Zillow or whatever. Then you're like, "Hey, here's what your house could look like. Wanna use my service?" And that — like — does the work up front.
04:19	Matt Mazzeo	Yeah.
04:20	Shaan Puri	Seems to be like... that's not casual advice. My brain kind of exploded a little bit. I was like, "Oh — so any business? Any business now, you can flip the sales model on its head." Absolutely. You could basically, instead of reaching out and saying, "Hey, would you like me to maybe do this for you if you pay me and hire me?" — versus saying, "Check this out, I made this for you."
04:38	Matt Mazzeo	Yes.
04:38	Shaan Puri	"Do you want to work with me? That's right. Obviously, that's going to be a better sales pitch. But now AI made that scalable. That was not scalable. That's exactly right."
04:44	Matt Mazzeo	I think you've nailed it. That's actually the thing that triggered me: when I think about new technology unlocks I try to ask, did it come with a distribution unlock too? I got to start investing at a time when mobile and social were born, and both of them came with totally new distribution mechanics — the feed, the algorithm, social sharing, and the App Store. They were all these greenfield distribution hacks. Gotcha.
05:10	Shaan Puri	So, not just the apps are new, but the App Store is also new.
05:13	Matt Mazzeo	That's.
05:13	Shaan Puri	Right — not just doing a connected tool, but... oh wait, social sharing.
05:17	Matt Mazzeo	Which is going to make this thing grow. Zynga lets you share natively as part of the product.
05:20	Shaan Puri	Not just a social game—it's a social share.
05:21	Matt Mazzeo	That's right. All of it was baked natively into the products. AI doesn't really come with that, right? AI has—this is why the incumbents are moving so fast—it makes your existing product distribution much better. You can add auto-suggest, auto-correct. You can put Gemini in every product. You can tell your Gmail to just respond "as if you're me" and do all of these things. And if you're ChatGPT—if you're novel enough, early enough, and powerful enough—you're now the dominant and default for the category. But it didn't come with new distribution for everybody yet. There were the G custom GPTs that were trying to get there; there's MCP that's trying to get there, but nothing obvious. [Note: "G custom GPTs" and "MCP" marked as spoken terms from the original transcript.] What did change was it let you do the work in advance—the cost of work dropped.
06:05	Shaan Puri	The cost.
06:05	Matt Mazzeo	Work dropped so dramatically that, instead of selling the promise, you can just sell the finished work.
06:11	Shaan Puri	Right. So you sell the proof instead.
06:13	Matt Mazzeo	I think a lot of people are still stuck in this paradigm. Consider all those headshot apps that came around — that's a perfect example. Nowadays it's 2 to 3¢ to create a high‑res headshot. Everybody was creating these headshot products and saying, "I'll make headshots for you." You still have to go in, use the tools to make the headshots, and upload the thing, but their photos are already out there. There's probably an addressable market of—call it—10 million super high‑intent people. They all need headshots. You know exactly who they are: you can audit LinkedIn and see their headshots are terrible. "We can make it much better." That's your target market. You could literally, for a penny or two, create a low‑res version of a new headshot and sell it direct. Don't ask them, "Do you want to set up a whole new headshot?" Just do the thing: give them a new headshot and then use that as the entry to another part of the business. "I'll do the rest of your LinkedIn. Your LinkedIn could be optimized." The second‑order service can be the service that you actually end up selling.
07:11	Shaan Puri	Right, but you...
07:12	Matt Mazzeo	Close them. You got the opportunity to meet with them and close them on those other things, right? Because you gave them the value without ever having to...
07:19	Shaan Puri	So, theoretically, what you could do is you could basically say, "I'm gonna do the color theory part proactively and, essentially, forever — free for everybody," and then I might even suggest a couple of looks. Yeah. And then if you actually — if...
07:33	Matt Mazzeo	I can do it for you every day. Do you want it every day?
07:35	Shaan Puri	Yeah, and you want me?
07:36	Matt Mazzeo	To keep track of things that are in style, right? Do you want to...?
07:38	Shaan Puri	Personalize it over time.
07:39	Matt Mazzeo	Right. Like, those seem like... and then that's the upsell.
07:42	Shaan Puri	But the kind of initial sell is pretty powerful. Then you string it together with tools like Clay or whatever, which enable mass-personalized outreach. So what do you have now as an entrepreneur? You have the ability to say, "Alright, for any given market..." Let's just take real estate agents as a market. Okay — I know who the agents are. I know what their listings are. </FormattedResponse>
08:00	Matt Mazzeo	Yeah.
08:00	Shaan Puri	Let's say I had an app that could, with AI, take your photos off your listing and stitch them together into a cool, Instagram-ready reel. Perfect. Alright—easy. I can then go into a tool like Clay and search for "real estate agents, California" and get all those. I can put an AI agent that will go scrape all their listings, because that's all publicly available: get their listings, get their photos, do the AI wrapper thing which says, "stitch these photos into a social-optimized video." Then I literally just email that person: "Hey Steve—saw you had that listing on 410 Montgomery Way built you…" [utterance unclear/unfinished]
08:37	Matt Mazzeo	A reel I built you.
08:38	Shaan Puri	A reel for it — I think it'll be really great. You could share this on your Instagram. "Hope you like it. Let me know if you have any suggested changes." Then the guy could be like, "Wow, this is incredible." You'd say, "Cool. Would you like me to be doing this for all your listings? I think this actually will help you sell." Sign up for this to end like that. Yes — the sales process is now, like, what a totally does versus what a sales guy does. [The word "totally" in the original is unclear; it may have been "tool" or something else.]
08:57	Matt Mazzeo	That's right.
08:58	Shaan Puri	Let's do another example. When I heard what you were doing — which we haven't talked about on this yet, but that's okay — I texted Ben and I said, "Ben, this is a billion-dollar secret." Yeah, because if you actually learn a differentiated, scalable sales mechanic, it's like our friends who started doing Google Ads back in '05. They're like, "Dude, I used to buy clicks for a penny." And now friends who were doing e-commerce on Facebook back when it was, you know, $1 CPM. </FormattedResponse>
09:20	Matt Mazzeo	Yeah.
09:21	Shaan Puri	Right — "less than a dollar CPM" — it's like those are billion-dollar secrets. There's this arbitrage that can exist right now, where there's a new sales model and a new distribution model that's less tapped today. So, what's an example of where you could use this?
09:34	Matt Mazzeo	I think the beauty of most enterprise software is that you're either helping them make more money, save money, or save time. Most businesses are pretty focused on the first of those: just help somebody make more money. So you can go and help almost anybody generate new business and new clients.
09:51	Shaan Puri	Right.
09:51	Matt Mazzeo	And so, the exercise could just be: "Hey, I delivered you new clients." Imagine I am the person running sales for Company X. You can do this with O3. I want to help them generate new business. What are two or three new, innovative ways that I could use O3 to generate business for this company? Just do that playbook for every business. It could be like: start a listserv that targets all of the potential customers of that client. It could be: create an Instagram account that highlights that business. Don't even ask permission — be like, "Best HVAC service — destined."
10:26	Shaan Puri	Right.
10:26	Matt Mazzeo	And, like, create an Instagram account for that person and do it at scale. If anybody comes in as DMs, be like, "Hey — got three new customers potentially for you." I think those kinds of arbitrages now exist for basically every company. There's going to be this lag — there's always a lag between what's possible today and how long it takes the average person to end up figuring it out.
10:48	Sam Parr	Alright. A few episodes ago I talked about something and I got thousands of messages asking me to go deeper and to explain — and that's what I'm about to do. I told you guys how I use ChatGPT as a life coach or a thought partner. What I did was upload all types of personal information: my personal finances, my net worth, my goals, different books that I like, and issues going on in my personal life and businesses. I uploaded so much information that the output is a GPT I can ask questions about issues I'm having in my life — for example, "How should I respond to this email?" or "What's the right decision, knowing my goals for the future?" I worked with HubSpot to put together a step-by-step process, showing the software I used and how I set this up. I had ChatGPT ingest all this information, so it's super easy for you to use. Like I said, I use this 10 to 20 times a day. It's literally changed my life. If you want that, it's free. There's a link below — just click it, enter your email, and we'll send you everything you need to know to set this up in about 20 minutes. I'll show you how I use it again 10 to 20 times a day. Alright, check it out — the link is below in the description. Back to the episode.
12:02	Shaan Puri	"So, you had some ideas that you sent me. One of the ideas you talked about—you gave me a bullet point, which I think is called the 'Mario Kart theory.' What does the Mario Kart theory mean?"
12:12	Matt Mazzeo	Yeah, the Mario Kart theory is actually that you could have made a lot of money investing in tech over the last handful of years — or at least in the SaaS space.
12:18	Shaan Puri	"If you just followed this one..."
12:19	Matt Mazzeo	"You just followed—everything's better in multiplayer." Build things that are multiplayer by default: Figma, Notion, Airtable, Slack. Take the communication tools or the enterprise tooling and make them real-time and cloud-based. Those things have network effects to them... instead of Mario.
12:40	Shaan Puri	Kart's better with friends.
12:41	Matt Mazzeo	Mario Kart's better with friends. But when I look at AI today, it's all siloed—it's almost like I'm having an individual conversation. Where's... where's group-chat GPT? You and I are in a conversation all the time; why isn't GPT in there with...
12:56	Shaan Puri	In there.
12:57	Matt Mazzeo	Yeah. Why isn't it living with us in those social moments? Why do I never see your ex exhaust out of your feed? Like, there's a bunch of stuff you'd be fine with sharing. There's a bunch of stuff that we'd all get better at. Like, Midjourney was right. Midjourney built—it's almost lost to history now—that three years ago Midjourney was on Discord. They were social by default. They had channels where you could see everybody's outputs. Yeah, it was kind of a magical experience, but almost nothing else went that direction.
13:26	Shaan Puri	Yeah.
13:26	Matt Mazzeo	"Like, it's all isolated and private... hidden. It's a travesty that the outputs for the free tier of GPT are private by default."
13:41	Shaan Puri	Majority is a great example of "bad is better" in product design. You think, "Wait, so okay, I download your app," and then it's like, "No—you open up Discord. You join our Discord, and in the Discord there's going to be a thousand channels." There are different rooms; just hop into any one and there's a flood of completely random people generating completely random images. It was actually kind of amazing, because you go in and suddenly you're inspired. You see, "Oh, that guy's using this to design T‑shirts." That person's creating anime. Then you see that same person change their prompt and make their images better, and you're like, "Oh, I never thought to add that to my prompt—let me copy‑paste that." So we were all teaching each other. This is not the chef's kiss product design that looks perfect in a mockup. No. It was a hacky way to get going, but it actually had this tremendous bet: the one thing it was great at—putting you around other people doing the same thing—outweighed all of the downsides of that initial user experience.
14:39	Matt Mazzeo	And they could focus on just making the model better. They didn't have to focus on resources or how to think about it. We know it's not perfect, right? We'll eventually get to the thing. But that was like model five or six by the time they built the destination site.
14:53	Shaan Puri	Right.
14:54	Matt Mazzeo	And it, frankly, feels less social to me in a lot of ways. They have the feed there, but I'm not seeing it the same way I did when it was native in Discord... and so...
15:01	Shaan Puri	Right.
15:01	Matt Mazzeo	I bet there's a bunch of tooling like that. The style idea would probably be better as a social environment in Discord. You and I are different color types, right? Similar body types. Maybe we should have a "similar" channel for ours, like...
15:13	Shaan Puri	Right.
15:14	Matt Mazzeo	For our area.
15:14	Shaan Puri	"Yeah, because if I saw you actually—this would be kinda magical, right? If I went in and, let's just pretend I was also soft autumn, and I'm like, 'I'm in with my soft autumn — the soft autumn boys; we have our channel' — and I see you actually proactively doing stuff."
15:29	Matt Mazzeo	Yeah.
15:29	Shaan Puri	"And you're prompting it — you're like, 'Oh, I got a wedding coming up,' and you prompt it. It gives you a really good recommendation. You actually take action on it: you upload a photo of you in it. I'm like, 'Damn, that actually worked out. Yeah, should look good.' Your success kind of triggers, inspires me, teaches me how to use the tool. Then I go back in and use it. That's a fundamentally better experience — it's going to be a more powerful experience than...
15:51	Matt Mazzeo	That's right.
15:51	Shaan Puri	You doing this on your own — me having to discover this, learn this, remember to use it — and then only the benefits are siloed in my experience, right? </FormattedResponse>
15:58	Matt Mazzeo	I redesigned your house from Zillow photos. That’s probably better as a social experience. I can just deliver the work: “Do you like this room? I can do the rest of your rooms.” I used to have to sell you on the idea that I’m a great designer and have you call my references. Don’t do any of that. You have a different budget, you have iterations—I can do it all for you. That should live in a social environment. Almost all of these experiences—outside of health and legal ones—are probably good enough that a free social tier would be accretive to the value. Right now we’re like bonobos. You ever see the videos of bonobos with ants on a stick? They’re a great example of social spreading. They put the ants on the stick, pull them out, and another one takes it. I was like, “that looks like an ant lollipop—I like that.” They move on and share the next thing. Instead, we’re cut off from everybody else, so I’m the only one eating the “ant lollipop,” and nobody else gets it. </FormattedResponse>
16:54	Shaan Puri	I feel like that's a really big idea. I think you could—without creating a new model or even a new use case—just design a social experience around the same use case.
17:06	Matt Mazzeo	Social by default is always better.
17:08	Shaan Puri	Right. So, is there anyone doing this right now? Have you seen any cool social projects? I basically want to take something like Claude — which is so powerful — but Claude is kind of losing the race of the individual.
17:21	Matt Mazzeo	Yeah.
17:21	Shaan Puri	User thing, but I'm just thinking: what would be the Figma-type of experience? You know—what would be the use case? Is it something like for work? For example, I have a lot of friends. We're all looking at how we can use AI in our—yeah—in our company. Everybody's kind of self-[unclear phrase: "doing the law the stick in the tree to get the ants"] where it's like, "Oh, you set up a thing where you connected it to your calendar and it…"
17:46	Matt Mazzeo	Yeah, the workflows.</FormattedResponse>
17:47	Shaan Puri	It researches who you're meeting with and then sends you the brief through a text message. Then you have that and you think, "Oh — I want that. How'd you set that up?" I have to ask them for a tutorial, whereas I really should just be able to fork that. Yeah, right. That's what I think Gumloop and some of these other guys are going to get towards...
18:01	Matt Mazzeo	Yeah.
18:01	Shaan Puri	One person makes a great workflow. It's like, "That problem should now be solved for the rest of us." Yeah — we should just be able to fork that.
18:07	Matt Mazzeo	People are doing — I mean, like, Zapier: there are tons of recipes or workflows that you can use. The thing that's missing is the social proof. The thing that's missing is: how many people liked it? What are the comments on it? What are the likes? What are the remixes? Who's using it? If you're using it, that's a big signal for me. If I'm doing it and I'm getting value out of this "fashion thing," that's probably a good signal for you. Those are things that tip people, and it's what's missing. These are like lessons of Web 2.0 that are kind of gone, right? We've almost collectively forgotten a bunch of these things — that were like "social by default." It makes a ton of sense: all the social apps, all the iOS apps went social. It feels like a collective forgetting, and it does feel like a big market gap of the...
18:56	Shaan Puri	Things I've picked up from you is that you have this... your process is sort of like, "I go try to do a thing." Firsthand experience in doing so. I run into the problems of doing that thing — maybe it's like styling yourself or whatever it is — and then learn by building. It's so funny because I thought of you: you started your career with CAA and you're this kind of charming guy who lives in LA. I just thought of you as, "this is like Hollywood pretty boy," and, like, actually you're more…
19:27	Matt Mazzeo	"Appreciate that."
19:28	Shaan Puri	Hacker. Tinker. Yeah, you focus on the pretty boy.
19:30	Matt Mazzeo	Yeah, no.
19:30	Shaan Puri	It's... but you're actually a hacker-tinkerer type, and I'm curious. I don't know the full story, but I would bet that that's led to some very good things in your career.
19:41	Matt Mazzeo	Even though you're...</FormattedResponse>
19:41	Shaan Puri	Not building the end company—yeah, but it's the learning by tinkering.
19:45	Matt Mazzeo	Oh, dude.
19:45	Shaan Puri	That led you to the investment, that led you to the opportunity—all of that. Can you tell a story that... </FormattedResponse>
19:49	Matt Mazzeo	Yeah, yeah, right. So during COVID [COVID-19 pandemic], I have a buddy, Brian Wagner. He's one of the best product guys I've ever met. Back in the day he was building a company I thought wasn't the right company, but this is a guy I wanted to spend time around because I just love jamming with him. I love the way he thought about product. We'd always build different products together. It was like our weekend hang—the thing we'd do. During COVID we got obsessed with this idea of a product called "Road Trip." It's a music-listening product that you could call into, and it was like you could talk while listening to music. It was this experience where, when we got into it, we'd be in a room... I don't know if you've experienced this: you probably sit on a phone call for like five minutes of silence, two minutes of silence, and it's like, "I gotta get off this phone call." It's the most awkward thing. But if you sit in a room with somebody and there's music playing, it takes all the edge out and you can last for hours.
20:39	Shaan Puri	In a.
20:39	Matt Mazzeo	Room with a grown man and have a conversation that dips in and out as long as there's music playing in the background. It's almost like the music creates the substrate that lets you pay attention. We started building this thing and, you know, it ended up not working. We couldn't figure out retention. Turns out live is really freaking hard. Synchronous is really, really hard. But in the process of building it we were like, "God, Firebase just sucks." "Firebase sucks." Every — you know, it's like I'm in the Google universe — like everything is tied to the u universe [unclear]. I remember this moment with Brian where he's like: > "If ever you see a company — I was, you know, investing early-stage at Codetube — it's like, 'Dude, if you ever see a company that is like an open-source Postgres-like Firebase alternative, like that's our ticket.'" A week later... </FormattedResponse>
21:21	Shaan Puri	By the way, from the outside it just seemed like: "Alright—solve the problem; Firebase does this." It's probably fine, right? If you weren't building, you wouldn't know. </FormattedResponse>
21:29	Matt Mazzeo	I would never have thought. He's like, "We built so many social products." It's like every time you have to go through the same thing, right? You're stuck in this Firebase universe. Eventually you try to cycle off of it, and it's a pain in the ass to migrate if you ever do. They weren't upgrading the product much. Now they're actually putting a little bit more attention into it because it's like their Replit competitor, but at the time there was no development there. A couple weeks later, it's YC [Y Combinator]. I'm looking at the list of the YC companies; I'm just ripping through it, and I'm like, "Supabase — open-source Firebase alternative."
22:00	Shaan Puri	Right.
22:00	Matt Mazzeo	And I'm like, "Okay — this... this feels like a great moment in time." I'm talking to — I'm talking to one of my partners. He's like, "I love this idea." And Ben Tossell, who's an incredible tinkerer, builder, and investor, was an angel in the company. He made the connection, and we ended up leading the seed round at Co2 and did a Series A there as well. It's a great company. I love those founders, and I think it'll be a great investment. But I never would have gotten close; I never would have thought about that idea if I hadn't experienced the pain first. And so do you when you're going into them.
22:33	Shaan Puri	Is it, "I'm doing this to learn," or "I'm doing this because I think this is the next big hit," and it just turns out that's wrong — and then the learning becomes the consolation prize? Mentally, how do you approach that?
22:43	Matt Mazzeo	It's curiosity. Sometimes it's because I'm solving a problem. Sometimes it's because I have an idea that's been sitting on me and nagging me and I just have to build it and get it off my brain. Sometimes it's like, I just think this is going to be a category that's going to be huge and I have to understand it. For example, I built a live—or video—shopping product. I got obsessed, so I spent a lot of time with the team at ByteDance. I used to spend a few hours a month with them and we would share notes on what I saw coming in the U.S. and what they saw coming out of China. It was a really fun jam with the team at ByteDance. One of those sessions focused on video commerce. This was four or five years ago, and I thought, this is going to be such a massive category. It was huge in China and nobody had cracked it in the U.S. You'd seen attempts, and it felt like we were all close. I decided I was going to build in that space so I could understand why it's not working here and what's broken. So we built a shop. My wife is a Korean expert and knows beauty really well, so the first shop we built was around Korean beauty products. I went through the whole process: I found wholesale suppliers, I spent hours in different Korean beauty shops to understand the products and the full stack of a Korean beauty routine. They told me how bad my skin was, and I went through the whole process with them. I built the store on TikTok, did the fulfillment myself, and went deep into the operation. There are two reasons I do this. One: it's fun to learn these things. Two: the only way to understand it is to get close to the metal. If you approach it out of curiosity, it's a win no matter what. It may turn into an interesting business, and at the very least you've seen what the pain points are.
24:22	Shaan Puri	Most VCs don't do this.
24:24	Matt Mazzeo	I don't know — everybody's got a different process. My process has always been that I try to better understand the pain points. I think it resonates more with founders if I can cogently talk about how I've used the product. Those things matter. You have more authenticity. I don't know if other people do this, but it's my best way of learning: I actually know what the product does, I actually use the product, and it actually makes me better. When I was talking to Amjad from Replit, he did an awesome podcast. By the way, I love that episode — I love that product. We were lucky enough to lead the Series B, and I was a board member at Replit. Partly that was because I love tinkering. It was a tinkerer’s dream: I was technical-ish, but I could never have built some of the products I wanted to build on my own. The idea that Replit would help me get closer to unlocking that felt like something everyone would want. It also turned out to be perfect timing for the entire AI wave. The way I got to spend time with Andre was by just using the product. At the first meeting I ever did, I built a repl — it was free to hack — and I learned Python to figure out how to send them an app in a repl that could, you know, pick a meeting time, and it was...
25:36	Shaan Puri	VC courtship — I like it. So what's their story with Replit and with Amjad? He seems like a real force of nature. When he came on the podcast, I was really impressed. He seemed like somebody who has a shot at building one of these generational companies. I don't know if Replit will — I'm an investor in it; I hope it succeeds. I think it could. But I just think he was really unique, the product was really unique, and the story of how he...
26:01	Matt Mazzeo	Got there. Yeah, the origin story is amazing.
26:03	Shaan Puri	What stood out to you from him, or from that?</FormattedResponse>
26:05	Matt Mazzeo	I mean, just that he was always—it's not like it was a new thing for him. He didn't have access to the tools of technology, but he knew how powerful they would be and would find whatever way to get them. Partly, he did that by working in "pc bombs" [unclear term] or whatever in his hometown, just to get access to a computer. It was like the same thing as when people spent time at research labs in Silicon Valley to get access to their first computers. It felt like he was going through that hero's journey.
26:32	Shaan Puri	Right.
26:33	Matt Mazzeo	Just to get access to compute.
26:34	Shaan Puri	As a kid in Jordan.
26:35	Matt Mazzeo	As a kid in Jordan, he was obsessed with spreading it to everybody. He was one of the first guys at Codecademy. He was a mixture of a true believer and someone technical enough to build it. He knew the pain points for himself and for everyone else who was coming up. He had a real mission to make a billion developers. If he had anything to do with it, he was going to get to that endpoint. He was almost like an endurance athlete in the sense that he was willing to wade through a decade of boredom. It's sexy now because there's a Replit agent and you can build your idea and you can write code. It feels like a wild roller coaster ride right now. But it was years of waiting: going to schools, building tools for teachers to teach kids how to code, building the infrastructure needed to full‑stack it, and a kind of religion around the idea that you had to full‑stack it — you couldn't outsource these things. It was all that minutia. There are a lot of people out there who are like, "I'm gonna build the ultimate 'vibe coding' platform." This whole moment feels like it's the idea guy's time: everybody says it's the time for the idea person. Now you get to skip what would have been the first six months to validate the thing, but it really just gets you to the entrance of the maze.
27:54	Shaan Puri	Right.
27:55	Matt Mazzeo	And it's like, now you've gotta go through...
27:56	Shaan Puri	What's up, Idea Guy? Welcome. Yeah.
27:58	Matt Mazzeo	Welcome to.
27:59	Shaan Puri	You ready?
27:59	Matt Mazzeo	Yeah — you up for this? Because the maze sucks. You're gonna tear apart every idea. What matters is the number of iterations on the experiment you can run: ripping that idea you had in half and having the world tell you, "your idea is garbage." Getting the brutal, honest feedback from your customers that they hate your idea. Maybe there's a kernel of something in that idea that leads you to the next one. The idea guy's time has come — we can validate ideas really fast now. But, I don't know if you've experienced this, just because you can build the thing... we've moved from Apple Notes to REPLs, with the ideas fully — the MVPs built.
28:36	Shaan Puri	Right.
28:36	Matt Mazzeo	That doesn't mean that you want to spend the next — you know — decade eating glass for that world. I have a lot of friends who are like "idea guys" — builder/product people — who say, "I'm just looking for the person who wants to take this idea over, right?" And people respond, "No, that's actually the job." Yeah. It's like the job is going and spending time courting a dentist to sell them a product.
28:56	Shaan Puri	It's the same thing with people who are like, "taste is all that matters now."
29:00	Matt Mazzeo	"Oh, dude — the taste copers are..."
29:02	Shaan Puri	Because it's even more—it's even... they're all the same. With the "idea guys," they're like, "Yeah, no, I don't even make the idea; I just taste it. I just judge the ideas." [Unclear: the word "taste" may be intended as "test".]
29:08	Matt Mazzeo	I hear another person talk about taste in this market. You know, have you ever read the articles on Lee Sedol from a decade ago? Do you know this story? No? Oh my God — it's an epic story. Lee Sedol is like the Korean master of Go. Do you understand this moment that happened a decade ago with AlphaGo? AlphaGo — it's like everybody was like...
29:25	Shaan Puri	Best at Go.
29:25	Matt Mazzeo	Chess — chess is a known thing. But no one will ever get to AlphaGo. There are "more possibilities in AlphaGo than there are atoms in the universe" — that's the quote I remember hearing. At least, it all plays AlphaGo. AlphaGo makes a move and it's this absurd move, and nobody says, "This is a beat — AlphaGo made a mistake," you know? All those possibilities mean it could make a mistake, and then 78 moves later AlphaGo beats the human. The human sits there, looks in the mirror, and retires. It's a sad story. He says things like, "I never truly scratched the surface of Go," and he feels embarrassed and helpless. The feeling of helplessness — everyone is going to have that moment. All those people talking about VCs — saying "it's safe, it's all about taste and psychology" — they're going to experience a taste moment where they're going to be "least it all" [phrase unclear]. They're going to have a moment where their taste and the AI's taste are presented to the end consumer, because ultimately taste is in the eye of the consumer — it's in the eye of the customer. The customer is the arbiter of taste. You think you're a tastemaker until you put the burger in front of the person and they say, "I like this burger more." "But that's a McDonald's burger," you say. "Yeah, I like the McDonald's burger." Billions of people might like a McDonald's burger more than your Saint Charles burger. You might prefer the other thing — it might not win the award — but it wins the choice of the consumer. So I think all these people who say "taste is our salvation" are going to get "least it all" — it's going to hurt real bad. They're going to run into the brick wall of consumer choice on taste.
31:09	Shaan Puri	Well, dude, this just happened. Remember — rewind five, seven years ago — it was all about self-driving cars. People were working on it and asking, "What's gonna happen to the truck drivers?" It was the truck drivers and the grocery-store checkout people. Folks were like, "We're just gonna have to find them new jobs." And then, all of a sudden, Dolly and these other tools came out, and it was like, "Oh shit — it's actually us first."
31:30	Matt Mazzeo	It's every... oh no — it's us first.
31:32	Shaan Puri	It's the graphic designers, the copywriters... like, you know, the first things that ChatGPT — and the modern wave of tools — were good at was writing. They even did that fake podcast of Joe Rogan talking to Steve Jobs. I was like, "Oh, podcast — alright, that's gone." It's just a matter of time. So I don't really think there's any sort of safe space. But I definitely don't think it's "I have unique taste," because you're right: whether the AI maybe has better taste, or it takes one million shots on goal while you take one, it's just going to play the law — like the law of large numbers — and get it right just by doing so many variations.
32:10	Matt Mazzeo	I—I mean, tried to write.
32:11	Shaan Puri	A movie script was — I was like, "Oh, I wanna write the opening scene of this idea for this movie I have."
32:15	Matt Mazzeo	Yeah.
32:16	Shaan Puri	And so I'm like, "Sit down." I'm like, "Oh shoot — the idea was really great." I told, like, five people the idea; they're like, "That's incredible. This is great. You should do this." And then I sit down. I actually have to do the work. I'm like, "My god, this is so hard. How do I — what do I start? Are they in a café? Are they in a bar? Where are the characters? What do they say?" So I go to ChatGPT and I upload a script of The Social Network. I upload, like, all of Aaron Sorkin's work, and I'm like, "Alright, you're kinda like Aaron Sorkin but you're also me. Here's the idea for the movie — give me a draft." It gives me a draft. Kinda sucks. I'm like, "Give me another one." It gives me another one instantaneously, with no ego about it. Gives me another one, gives me another one, gives me another one. Eventually it got pretty good at doing that scene and pretty...
32:58	Matt Mazzeo	Good measure, by your taste.
32:59	Shaan Puri	By my taste.
33:00	Matt Mazzeo	"As enjoyed, right?" </FormattedResponse>
33:02	Shaan Puri	"But what I'm saying is, even if—let's say—my taste may be right or may be wrong, either way, it took ten attempts before I could even lift a finger."
33:09	Matt Mazzeo	"The volume was there."
33:10	Shaan Puri	The volume is gonna win. Yeah, and in a taste game it's gonna take 100,000 shots on goal. Same thing with ads, right? If you look at fashion brands that do well with Facebook ads, go look at their ad library. It's not like they have one David Ogilvy sitting in a room and he just comes up with this one beautiful ad. They're running, like, a thousand ads in a catalog—quantitatively testing which one gets the highest ROAS—and they're doubling down on that.
33:33	Matt Mazzeo	It's so funny because we've gotten over the idea of "taste" in all these really fuzzy areas of human consumption. It's like taste on video. It turns out people preferred the algorithm of ByteDance and TikTok to following your programming on television. Your years of study of what people are enjoying got replaced—big time—by the algorithm. I can't tell you how many mistakes I watched.
34:00	Shaan Puri	That's so true. The algo is literally the definition of taste—taste already proven. The algo dominates. That's right. You know, whether it's editors inside the company or it's me, for myself, it's better than what I could do for myself, even.
34:13	Matt Mazzeo	It's funny — we make these mistakes all the time. I'm watching it, and I mean, you could argue that Apple lost the entire podcast business based on their religion around taste. Apple started the entire podcast moment. They owned every podcast that got launched on Apple, and they were like, "We really need to own the future of entertainment." Then they asked, "What is the future of entertainment?" It's great shows — HBO — tasteful creators, tasteful makers. Things that are just like... no slop. But on the other side of the house, the volume was being made by people who didn't look like the top Hollywood creators. With all of them there was no product iteration, no help for monetization, no featuring of them, no studio production, no tools, no budgets, no metrics or data back — none of the things you would do if you had a million free creators building awesome audiences. Those audiences, it turns out, can now sway elections, and many of them start on your platform. Then you fast-forward: where's most of that consumption? It's YouTube and Spotify. You realize they fumbled. They fumbled because they didn't think it was tasteful and they missed it. YouTube, back in the day, did the same thing. They actually had half a billion dollars ($500,000,000) for content... there was like the
35:40	Shaan Puri	Originals. Right—originals. It was like a half.
35:42	Matt Mazzeo	$1 billion — they came to all the agencies and said, "We want your best creators." We all went and pitched them 5 million channels. They gave a big budget, and they also allocated about $5 million for the native creators. Phil DeFranco and Michelle Phan got small infusions of cash here and there. The only channels left standing were the creators — the native creators. It didn't match the taste profile of the advertisers upstream of all this shit, but those native creators were the ones creating stuff that people wanted. I mean, Sam Altman has the intuition for basically every new business that's getting built. He ran YC [Y Combinator] for a long time, so we're not that far away from him saying, "Hey, why not just build a funding model into the next GPT [Generative Pretrained Transformer]? If somebody's talking about starting a business and I actually like that business — do you think it'll make money? Let's provide them credits, let's provide them a little bit of funding to get going." I don't know — does that change how investment works? Is that maybe upstream of all of the tastemakers?
36:40	Shaan Puri	So, what do you think it looks like? Let's say AI just keeps getting better. Yeah, there's no life raft for the "taste" guys or the "idea" guys or whatever — like, AI's getting better at all these things and you're going to have... so what happens? What happens to us?
36:54	Matt Mazzeo	"What do you do? How do you think about that?" I think you just keep creating things that people want. Use the tools to piece them together and get to the answer faster: do customers want this thing? Give the customers what they want. That's the true north — build something people want. It's the YC mantra: "build a product people want." It's about how fast you can get the product to the people who want it, and how many times you can iterate on it until you eventually get the thing that people want. Just laser-focus on that and cut everything else out of your life. That's how you build most of these scaling businesses.
37:27	Shaan Puri	Right.
37:27	Matt Mazzeo	To me, that's never going to go away. People are always going to want things, and you can use these tools to provide those things to them. There's probably a good arb [arbitrage] for the next long while. We've got a long time before this stuff filters through the economy, because it's not how fast the technology is moving—it's how fast people adjust their reality and their habits. That's the friction that actually exists here. Technology is mind‑boggling, yeah. But my mom learning how to use a second brain is the same as you and I learning how to use a second brain.
37:56	Shaan Puri	"It's really hard. Do you have any other AI-related theories or frameworks that are interesting that you've been noodling on?"
38:02	Matt Mazzeo	Yeah. I've only—I've made three AI investments [interpreted "post-COVID"] as an angel. All of them follow the same kind of framework. One of them is for a buddy of ours in the travel space, and the other two are founders that I backed before. All of them are basically agent platforms. One is called Cognition, which makes a product called Devon. The founder, Scott, is someone I backed once before, and Scott is—he's a genius.
38:32	Shaan Puri	Correct. Yes, he's actually.
38:34	Matt Mazzeo	Mathematical genius. Among many things, one of the things I was riffing on was that it's possible Devon is just Scott in the background answering all the coding questions because he's so good. Another product is called Augment. It's Harish Abbott. Harish was a founder of a company that I was on the board of called Deliver, which was like third‑party logistics—it was like Amazon Prime for everybody. He worked at Amazon in fulfillment, then built Deliver and sold it to Shopify for a couple billion dollars. Harish is one of those humans where, if you could imagine the ultimate logistics hire, you would hire Harish. No detail escapes him. He doesn't mind staring into the details of a warehouse and figuring out how to optimize every minute thing: product comes in—where does it go, what platform, how do you get optimized, what's the routing down to the minute. He's that brain. It's his gift. It's his dark gift. So both of these people—actually, all three of these people—have these dark gifts. You could never, in your wildest dreams, hire Scott. You could never hire Harish to do your logistics. But what you can do now is they've built platforms that are agentic. Imagine if the reinforcement training was done by the ultimate employee, and that was priced at tokens.
39:48	Shaan Puri	"What does that actually mean? So, they're billing an agent — that's going to be, let's say, your logistics employee. But he himself is literally auditing the answers."
39:59	Matt Mazzeo	He is building and saying, like, "Hey, we know exactly what goes into building world-class fulfillment and logistics. We've done it at Amazon; we did it at Deliverr. I know all these details." Now you've got a company that's an SMB—it's an e‑commerce company—and you don't have the knowledge that I have. But I can program it into an agent so that all the workflows you're eventually going to have to do, this agent can help you with. It will provide the things that would otherwise take you a lot of time, even if you were the best in the world at it. It'll do those things automatically for you. So I think we're moving to a world where the best at a task will be able to productize themselves as these agents and deliver it.
40:39	Shaan Puri	So, you know, when social media came out, one of the cool things that happened—and one of the reasons why there's this word "influencer"—is that prior to social, the only people you knew who did a thing were the people around you. You were limited by your geography. The superpower of the internet is that it vaporized geography. So, when we had a baby, we started following these "it-mom" influencer types who teach you how to get your baby to sleep and offer little tips on swaddles, diapers, and all these small things. I'm getting the best mom coaching—and dad coaching.
41:12	Matt Mazzeo	Yeah, from...
41:14	Shaan Puri	Like, not just my geography but the world. Yeah. Those people became—well, the best teachers were not teaching in schools anymore. The best teachers were teaching on Instagram, YouTube, and other places, and they were getting millions and millions of followers because the best teacher actually should have a few billion students. So the best got rewarded in that way. That's what the internet—like the last 15 years—has done. And now, with agents, it's kind of the same thing: the best employee is no longer limited to who you can afford or who's nearby, who you can hire locally. Actually, globally—who's the best at travel planning and booking? Who's the best at logistics, workflows, planning, and analysis? </FormattedResponse>
41:53	Matt Mazzeo	Who's the best coding engineer?
41:53	Shaan Puri	"Who's the best engineer you know?" Right — it's actually not "who's best engineer"; it's "who's the best engineer."
41:59	Matt Mazzeo	Yeah.
41:59	Shaan Puri	He can program himself in the same way that content creators were able to make pieces of content that scale. I can go to sleep and millions of people can listen to this podcast. Now they're doing that with a piece of software—a piece of code that's going to actually take their brain and be like, "Yo, I got that guy's brain here in this piece of..." [trails off]
42:16	Matt Mazzeo	And do the digital work that they would have done for you.
42:18	Shaan Puri	Right — and they can. Just like the best teachers have millions of students, the best employees now actually work for millions of companies. That's what's going to end up happening. "That's amazing. I want to do some stories with me. So, I have here — oh, this is scary. These are scary. We're going to draw a card, and then I want... Are there people that you've encountered in your life? I want you to tell me some stories."
42:38	Matt Mazzeo	"Or story timing."
42:39	Shaan Puri	"Take a card." "Yeah. Alright—show the camera. Who'd you get?"
42:44	Matt Mazzeo	"I got Peter Thiel."
42:45	Shaan Puri	"Peter Thiel. Do you have any Peter Thiel—either lessons learned, funny encounters, interesting stories, or seeing him in action? What have you got on Peter Thiel?"
42:53	Matt Mazzeo	The only time I've ever encountered Peter Thiel was when we were both pursuing the opportunity to invest in Replit. When you go up against somebody like Peter Thiel, you generally plan to lose in most cases. I didn't end up losing the opportunity to invest in Replit. Part of it is the realization that, in many games, people count themselves out—but on any given Sunday you have the opportunity to pull out a "W." I love that product. I love it authentically; I use it all the time. I love the vision of empowering a billion people to build the companies they want to build. I feel viscerally the tinkerer's dilemma: I'm not an expert, but I wish that I were, and I wish I could bring this idea to life. That software is like a bicycle. I'm on a tricycle—training wheels at best—and this thing makes you feel like you're on a motorcycle. I love that feeling. If you can tell that feeling to somebody and get them as excited about it as you are, and they use the product and obsess over the details, then great. I'd send Amjad every product thought I had. I'd have shower thoughts about it. It wasn't work; it wasn't "the art of the VC deal."
44:22	Shaan Puri	I was like, "I was just genuine."
44:23	Matt Mazzeo	I genuinely loved it. Then I think there was a moment where he was like, "I can have a guy who's probably one of the best, you know, investors of all time, or I can have Matt." Matt loves me, and that was his choice. So I think the lesson is: if you really do the work and you really love it, you might not win them all, but any given Sunday you could pull out the dub. Bondo — who you got? David Bonderman.
45:00	Shaan Puri	Okay—tell us: who is he, and what's David Bonderman?
45:02	Matt Mazzeo	"What's your interaction?"</FormattedResponse>
45:03	Shaan Puri	With him.
45:03	Matt Mazzeo	David Bonderman passed recently. David Bonderman was the founder of TPG. He was a guy I got to know later in life. He worked with a very close friend and mentor of mine named Rick Hess. During my sabbatical, one of the things I did was decide I was just going to go and try to spend time with people I had always admired and never really gotten to meet. David was on that shortlist. David had ultimately been part of the group that acquired a majority position in CAA, and I got my first exposure to him there. He had an incredible, storied career: he was a huge investor in the airline industry, had owned sports teams, and was a titan of industry who did it his way.
45:52	Shaan Puri	What was his come-up, by the way? Because, you know, that's the cool stuff you do at the fountain.
45:55	Matt Mazzeo	Well, he worked for the Bass family. He helped build the investment strategy for the Bass family in Texas and ultimately pivoted that into, I think, the first investment: Continental Airlines, where he made his first real big exit. He ultimately ended up building TPG into one of the titans of the private equity industry. I got to spend a bunch of time with David—hours at a time. He was the kind of guy who, later in life, still had it. He'd have days where he was as hot and fired up as anything. Some days it would be: "Let's just spend time talking about his job and his career and lessons along the way," and then other times... wait
46:33	Shaan Puri	Sorry. How did you do this? So you're on your sabbatical and you're like, "Yeah, I want to spend time with these people." What—call them up? You're like, "Hey, can I just come hang for a couple of days?" What do you...
46:39	Matt Mazzeo	I called Rick Hess. Rick was building a new strategy for investing; I was just helping friends. I was spending time doing—treating all of my friends and family like portfolio companies. I was in that portfolio VC mindset: I'm going to treat them all like portfolio companies, go do things for them that matter, and help. So I would help friends start companies. I would go over and clean out their garage. I had time for the first time in my twenty years, and I just wanted to spend time with people I loved. I wrote a list of all the people and would say, "Hey, I have time for the next couple weeks. Can I help you on something?" Rick was like, "Well, I'm building this thing and I'm partnering with David. Why don't you go over and talk to him about it?" On one of those trips he said, "David's really interested in AI—will you put together some thoughts?" I said 100% yes. I got to spend a three-hour session with Bondo, and I laid out where I saw AI going. I showed him Midjourney, I showed him ChatGPT, and I walked him through all the new products. He was quietly riffing on it and asking a few questions here and there. At the end he asked, "So what are you going to do?" I said, "I don't know. I'm going to invest in a couple friends who are smart and building stuff." He asked, "What would you do?" He goes, "I'd buy railroads." I thought, okay—maybe he's gone crazy. I didn't understand what he meant, so I asked him to explain. He said there are two ways to play it: > "This world is moving so fast that everything is changing. If you build software or a product today, it feels like it could be obsolete in six months—it's shifting sand. You can play toward the bleeding edge, but that's where everyone will be, because it's shiny and fun and distracting. > Or you can buy a railroad. Think about the things that, in times of the most change, are still going to be here. Focus on the industries that aren't going anywhere." That moment felt like talking to business Yoda. It was a really interesting framework. It applies: the business I ended up spending a bunch of last year working on is in a really old, dirty industry that's going to be around forever, but it applies these tools to it. I met him in the middle—I used the tools that are available today to help an industry that I knew would be stable in the face of all this change. He was one of those guys who could pull out "invest in trains" and have it be one of the most incredible pieces of business advice.
49:24	Shaan Puri	Alright, we got...
49:24	Matt Mazzeo	Three more. I think the thing that I pulled from Mark was the idea that where venture was heading—the institutionalization of venture. It kinda feels to me like we're at almost the apex of that. When I joined this industry, there was all of this “tribal knowledge.” There were things that were hidden, secrets. What's an example?
49:44	Shaan Puri	What was hidden—secrets?
49:45	Matt Mazzeo	Like, how to write a term sheet was unknown. How to run a closing process was unknown. The real mechanics of the industry were unknown—let alone the metrics: what is good, great, bad—growth metrics, retention, net dollar retention. A lot of venture today feels really institutionalized. Even further out, we’re at a point where the knowledge is all shared and public. People have posted as much institutional knowledge about the industry as possible. We have documents like the SAFE that revolutionized things. YC created a whole process for how to institutionalize fundraising—the “handshake protocol.” We created protocols for the industry, and we’re at a more mature part of this industry. When I got into it, there were about ten seed funds in the market. You didn’t have to be great—you just had to be alive to go find these deals. There wasn’t a ton of competition; there weren’t mega funds doing seed rounds. You met a founder, you liked the founder, they went to YC, there were fifty other people in the room. You kind of said, “Hey, I like this one—let’s do it,” and you got to a handshake on a note. Mark, I think, saw where the world was heading on this. He saw that all these asset classes were maturing and that we were going to build a much bigger platform for them. He was inspired by a lot of the CAA mentality early on. CAA shifted the way agencies had been done. It went from “I represent one client with one deal,” a boutique model, to something more strategic. Ovid’s had this insight: if I rip an agent away from another agency, all their clients come. At CAA, one client sat in the middle of ten agents with each of their own personalized expertise. In VC, the same thing kind of existed—you had a one-on-one VC, a boutique: “I’ve got this board member, that’s my person.” Then Andreessen comes along and it’s like, “No—you’ve got all of us.” We all come with a value proposition: recruiting, marketing, and all these services. It was the realization that VC was a service industry and the institutionalizing of that service that changed how a lot of venture was done. Everybody rushed to create platforms, teams, and structure because they were beating people on that—even if it wasn’t always effective. They pushed the edge and owned the brand of what a platform VC could be. On the other edge, you had Benchmark and Founders Fund, where the message was, essentially, “You need me,” or, with Founders Fund, “You don’t need anybody.”
52:24	Shaan Puri	You know.
52:25	Matt Mazzeo	So I think what I took from that was: you can look at any industry — VC included — and bring a new institutionalization to it. I think they're still doing it. I think they just acquired Turpentine and Thornburg. It's the realization that one of the edges you're competing against is distribution and media. If you can bring customers, right — what do startups really care about? It's customers and attention: the resources, attention. So it's like, "we're going to build attention into the offering," and that feels, again, ahead of the customer-first, customer-centric response. Most VCs were selling things that weren't necessarily what the customer needed.
53:07	Shaan Puri	Right.</FormattedResponse>
53:07	Matt Mazzeo	It was capital.
53:08	Shaan Puri	What's that guy's name—Jilly something? He invests in Wiz and, like, every security-service company. Basically, he was like he unbundled the Andreessen model. He's like, "Cool — we're not gonna help you with, you know, marketing or recruiting." "Yeah, like, we have a network of CISOs at the top 500 companies and you'll get to $2,000,000 in ARR if you invest with us," he said. And then, to those guys, he was like, "Hey, you're gonna own a piece of this portfolio."
53:32	Matt Mazzeo	Yeah.
53:33	Shaan Puri	While you're sitting in your job, you're going to also be a venture investor—a venture partner in our fund. Yeah. And you kind of created a little "bribery network" that actually worked and, like, turns out huge, huge results, like in a way outperforming results by figuring out that platform. </FormattedResponse>
53:47	Matt Mazzeo	Yeah, and I think those are what he's continuing to do: find what those gaps are and what the customer really wants.
53:54	Shaan Puri	Right
53:54	Matt Mazzeo	And it's like, if you could tell me, "Oh, I want revenue. That's what I really want," and you can deliver that, that's pretty incredible. And part of...
54:02	Shaan Puri	What the startup founder wanted was, like, you know, the brand — a16z.
54:06	Matt Mazzeo	I want cool. I want cool.
54:08	Shaan Puri	"I want to be at the... I want Andrea Aguerdala to be, like, at our— I want to meet him. I want these different things that, you know... Again, just work backwards from what actually resonates."
54:16	Matt Mazzeo	Most VCs were selling "cool," right? They were selling—maybe not cool—they were selling the promise of de-risked, right? Credibility. Yeah, the promise of credibility that we were something that had been validated. They were selling that if you got a Sequoia investment, that's like, "This is a real business." They've looked at billions—how many trillions of dollars of market cap are from those people? We're among them. You sell the association with market cap.
54:45	Shaan Puri	Alright, we got two more. </FormattedResponse>
54:47	Matt Mazzeo	Derek Jeter — I have one absurd Derek Jeter story that's super funny. One of the things you realize at a talent agency and being around A-list stars all the time is that a lot of people go a little bananas. There's almost no equivalent, though, for sports stars, because something embeds in your brain when you are a little kid that these people are larger than life. I was a young VC at Lowercase, and Derek Jeter and his partner were starting a sports media company. They said they wanted to spend time, and I had seen all of this in media and tech for years. So we went out to dinner at a nice restaurant in Beverly Hills. It was the three of us. Up to the table walks a man with the thickest, deepest New York accent you've ever heard. He says, "Hey, Gus," and when he sees Jeter he's about to talk about the specials, and his face melts. The guy almost—time is a flat circle—and he is lost. It's one of those moments where you watch him reverse-age into an eight-year-old. This is the moment of his life. He's going to these, like, captures and like, "Mister Jeter, Mister Jeter, I'm a big fan," and he goes through the whole thing and at some...
56:18	Matt Mazzeo	We all ordered the same thing — we ordered the special, the guy recommends those truffle pastas. He comes back to the table a little bit later and he's like, "You know, it's fresh truffle on top," and he gives me a little bit of fresh truffle. He gives the manager a little bit of fresh truffle, and then he's over at Mr. Jeter. He looks at Mr. Jeter as he's putting the fresh truffle on, and his brain freezes but his hand doesn't. It gets to a... where it is — it's so funny that we just start laughing as it's happening, because it is an ongoing mountain of truffle on this plate. The whole plate is black. It must have been hundreds — hundreds of dollars of fresh truffle on Mr. Jeter's plate. He's just looking at Mr. Jeter and he's like, "I think you got it," and it was just one of those moments where you're like, it's really funny to watch people in that moment of their life. I don't have... I don't have, like, a good...
57:08	Shaan Puri	Lesson: At least have a different level of cool.
57:11	Matt Mazzeo	It's like—it transports you back to being a little kid. I’ve had one of those experiences myself. Yeah, this might be the one. I grew up as the most diehard Lakers fan. I grew up in Los Angeles. I went to college during the Shaq–Kobe years. In high school, during those Shaq–Kobe years, Christmas was like: "Laker game on TV"—that's what we did. He's one of those people I got to watch, and then I got to be obnoxious at college—watching him beat the crap out of the Celtics. It was really fun. Again, at Lowercase, I had an opportunity. There was a brief moment in time where I was managing director of a fund called "Lowercase Thirteen," which was a Kobe Bryant partnership. It was, for a split second...
58:09	Shaan Puri	How does that happen? Kobe reaches out to you?
58:11	Matt Mazzeo	I had a friend who was advising Kobe—her name is Betsy Skolnick. Betsy said, "Hey Matt, Kobe is... I think he'd been injured at the moment. He's thinking about what comes after basketball, and he's always been obsessed with tech and venture. He really wants to learn. Will you come down and have dinner and spend time with Kobe and teach him a little bit about venture capital?" This was my Jeter-like truffle moment. I was like, "Let's go—I'm ready. Do you want me to come now? I'll come right now in the car, and we can do it." She said great, we'll set up a time in the next couple weeks. Can you come down? Chris and I go down and have dinner with Kobe. He was asking questions. You could tell some people have a second gear of curiosity. Sometimes people ask surface questions and stop there. But there's another layer—people who ask every relevant follow-up question you can imagine. Having been in the industry a long time, you can tell when someone has real depth of curiosity: they want the details, not just the surface answer. Kobe did that the whole time. It was like a three-hour dinner. By the end he was like, "This was awesome, thank you so much. I'd love to find ways to do this again. Will you send me stuff I can read?" I said okay, I'll send him some stuff. I put together the essential reading list—every blog post I ever loved. I went through all the girly posts and the Wits Wilson posts [phrase unclear], all the stuff, thinking he would never read it. At 2 a.m. I got a text: "I read that post." It was like, whoa—he was in it. He would text and we'd have conversations in the middle of the night. I was like, when is this guy sleeping? He must be going crazy because he was injured—just texting. Eventually he said, "I really want to go spend some time and meet..." We suggested he could come with us to San Francisco and meet a bunch of startup companies. He said, "I'd love to." Again I thought he'd never do it, but he said, "Let's go—when are we going?" So we took him up to San Francisco and saw a couple startups. We spread it out: early-stage app startups, and then we went to Twitter. Twitter had just bought Vine. We were walking around, onboarding him to Vine, and having conversations with Dick Costolo and Adam Bain. He wanted to see the whole place, so we went down to the engineering floor. He walked past one of the hang rooms for engineers and there was a kid in there doing paper-shot basketball [paper-toss game]. Have you ever watched those people who are savants at paper toss? They can do it with both hands; they go into another gear—it's like the Matrix—and they're both-handing and...
01:00:57	Shaan Puri	"Yeah, they look like the Tesla humanoid robot—that's..."
01:00:59	Matt Mazzeo	Yeah, just like—you’re like, "this guy is that to the tee." Kobe looks in and he's like, "I gotta do it." He walks up behind the guy and he goes, "I got next." All the engineers see, like Dick and Adam, and then they see Kobe and they're like, "Holy shit, Jimmy's about to play Kobe!" They rush in. It's like Vine has just come out; they're all taking video of this thing. Kobe goes and he's like, "Bang." </FormattedResponse>
01:01:22	Shaan Puri	"Bang, bang, bang, bang, bang."
01:01:23	Matt Mazzeo	And he hits, like, 80. Then Jimmy goes — "bang, bang, bang, bang, bang" — and he hits, like, 82. The room erupts. It's the greatest moment in this guy's life. He has never been... it's the whole thing — no one will ever believe this moment, right? He's like, "I beat Kobe at Papa Shotty, that guy." You expect Kobe to be like, "Oh, congrats, dude," high five. I just remember the look on Kobe's face — he was legitimately pissed.
01:01:51	Shaan Puri	"It was like that. It wasn't like a fake piss."
01:01:53	Matt Mazzeo	It was like, "I'm buying one of these things."
01:01:54	Shaan Puri	Right.
01:01:55	Matt Mazzeo	"Never fucking happened."
01:01:55	Shaan Puri	"Yeah, I will destroy you the next time I see you."
01:01:57	Matt Mazzeo	"I'm never letting that shit happen again, and he... walks out."
01:02:01	Shaan Puri	Delete his Twitter account, yeah.
01:02:03	Matt Mazzeo	Anyway, it was pretty incredible getting to spend the time with him. He was like, "Hey, I really want to build a shop." For a brief moment, we had a side fund that we were putting together with Kobe, and we made two investments. One of them ended up being Stripe out of that fund. But he had a business partner who tried to renegotiate, and we don't do those kinds of things. So, anyway, Matt.
01:02:29	Shaan Puri	Full of stories. This is amazing — super fun. Thanks to everybody for doing it.

Discover

Listen

Catalog

Guides

Connect

How to Scale a Profitable Agency with 0 Employees (Using AI Agents)

Transcript