Transcription
We're in the age of AI, where building is a commodity. Technical skills are now being somewhat devalued, and strategic skills are more valuable. The business leaders who have that strategic insight and who have that leadership ability are going to have the opportunity to really outexecute the technical co-founders.
If you had to break it down for founders who are about to start their company: if you don't use these tools, what are the risks you are kind of exposing yourself to? AI is not going to replace founders anytime soon, but founders who use AI are absolutely going to replace founders who don't. The next wave of great application layer AI companies are going to be those companies that have those insights into customer requirements. Startup selling to startups can be a very dangerous game, as we've seen time and time again in history. We're a startup selling to small businesses; having an understanding of the common consumer is a winning formula.
Welcome back to the podcast. Today, I'm really excited for our conversation with Jeff Buskang. He co-founded You Promise, which was acquired for $300 million. He was also an executive at Open Market, which went public. He's also co-founded the VC firm Flybridge Capital Partners, and he's also a senior lecturer at HBS. That's quite a mouthful. Well, usually when we do an intro, like, you know, it's usually one thing for one person, but I guess it just goes to show how much you've just accomplished over the course of your career. So Jeff, welcome to the podcast.
Hey, thanks so much, guys. Maybe it's just a sign of how old I am as well, but thank you very much. I appreciate you guys having me.
Yeah, great. I guess the natural starting point is, you know, again because of your breadth of experience, you've seen, you know, you've seen the .com, then you've seen cloud, now you're seeing AI. What are some kind of similarities, maybe, and some differences that you've seen through this evolution? Would love to, would love to get some thoughts on that.
Yeah, so similarities: we're in the midst of a hype cycle because there's just so much energy and enthusiasm about AI, just like there was about the internet when my company Open Market went public in 1996. We had a $1.2 billion market cap with 1.8 million of trailing revenue. So when people talk about bubble times and revenue multiples that are a little absurd today, I lived that cycle, and it is really hard to grow into your valuation. And so, hype cycle and then working your butt off to grow into that valuation and realize that vision, that promise—that's a similarity. A huge difference, guys, is that the moment we're in right now is nothing like we've ever seen because of the proliferation of 8 billion of these devices and 2 billion laptops and a billion and a half cars in the world, which means that the AI capabilities are available instantly as soon as these new models and new applications are deployed. They're distributed to every single laptop, car, and smartphone in the world. And in the early internet days, our distribution was far more limited; the market we were focused on was much narrower, and it just wasn't global. So those are the huge differences. Today, this is something I've been thinking about quite a lot because we're seeing a new wave of companies that are growing really fast, like we've seen Cursor, a few others that have basically gone from 0 to 10 or 0 to 100 million revenue within like 1 year. It just never happened before, but it's happening now.
One of the hypotheses that I have is just the one that you said: it's just like because we have so many mobile devices, because we have so many laptops, because this distribution is just built out over the internet that apps can scale a lot faster. Do you think that's an accurate claim for why a new, the new generation of products are growing so fast, or do you feel like there's a little bit of nuance that I might be missing there?
I think that's accurate, but there is nuance. So yes, distribution is happening faster because these distribution channels are more efficient. Heck, even X and Facebook and Instagram are distribution channels for SaaS companies and for consumer companies. But what's really different is that we're in an era of ERR, not just ARR. ARR, everyone knows what that means: annual recurring revenue, and those are the annual contracts that you can count on year-over-year if you're an enterprise software company. ERR is experimental revenue, and the dynamic we see now is that enterprises are experimenting with a lot of AI tools. There's this top-down mandate to run a bunch of experiments within your organization, and so organizations are grabbing all these AI tools and they're experimenting with them. And so you see these flash bursts of revenue, but then they're not renewing; they're not deploying them enterprisewide. So there's this moment where we're in sort of the, you know, pilot ghetto of AI applications, and to move out of the ghetto and get into the mainstream, get into the enterprisewide deployments—that's really where the heart and soul of these growth companies are going to be.
We have a lot of builders, a lot of founders who are listening to this, who are going to listen to this episode. What advice would you have for them?
Yeah, so first, you need to find a toehold into your value proposition, into your enterprise—something that's really must-have, not just nice-to-have. And then, secondly, you need to have the conversation with your customer: if this works, and what are the tangible metrics that we should be aiming for? But if this works, and if we hit those tangible metrics, what's the deployment path enterprisewide? What's the expansion path look like? Too often, founders get so excited, even giddy, over signing these pilots, and these are substantive pilots—50K, 100K pilots are the order of the day—but then when it comes to enterprisewide deployment, everybody's looking at their shoes and saying, "Oh gee, I don't know what to do here." So really being tangible about the milestones for enterprisewide deployment and the steps that that deployment might look like and the timelines associated with those steps.
So obviously you've like built like these phenomenal companies; you were Open Markets, took it to public as well, and then you decided to start a VC fund. Tell us a bit more about like why, how did that happen? Why did you start that?
So I co-founded the firm with Chip Hazard, who you guys know well. Chip had been previously at Greylock; Greylock had backed my previous two companies, Open Market and You Promise. Chip and I knew each other at Harvard Business School, now over 30 years ago. So we've been colleagues and partners for three decades, which is kind of crazy to say out loud, but that's the way time flies. And Chip had a thesis, and I had a thesis that post .com crash in the early 2000s, we were in a moment where all the smart money had shifted to the west coast, and there was an opportunity for a seed-stage firm to focus on the New York and Boston ecosystems. And so that's what Flybridge has been doing as an enterprise software focused and now AI focused early-stage seed investor.
And how did like—so like obviously Chip comes from like a very prolific VC background, you came from like a builder background. How did like this combination allow you to pick the best companies? You guys preceded MongoDB, which kind of anchored the New York tech scene—a very prolific company. Invested in Falcon, which is like a big crypto infrastructure company; Redox Health; Firebase; so many like amazing cloud companies, right? So like how did you guys find all these opportunities?
You know, Hassan, it's it's a little bit um, kind of old-fashioned in the way uh that I would describe and answer that question, which is we have a macro theme or thesis that we believe in, which is a hypothesis about the future, and that manif begins to manifest itself into a detailed investment thesis. And that's a top-down lens or a framing that we have. So over a decade ago, it was this notion of big data and cloud being an opportunity for new applications, new data structures, new middleware. And then we meet the most amazing entrepreneurs that we're very blessed to be with in the Harvard and MIT ecosystem here in Boston or in New York, coming out of the crash of Silicon Valley and the new surge in the New York tech scene. And we, we meet those entrepreneurs, and if they articulate a vision and a hypothesis about the future that matches our hypothesis about the future, then we have the potential for a match. And so I think as an early-stage VC, you always have to have an inquisitive mind, insatiable curiosity, a belief about the future, and then an openness and a willingness to meet amazing people and and partner with them for that journey.
So obviously, I'm sure you have heard like hundreds or maybe even thousands of pitches at this point. Are there like some pitches which stand out for you, and like what was the reason why they stand out? Or maybe, and you didn't end up investing, but they were still amazing pitches, like you know, maybe some lessons from there which you can share for some of our audience? That would be awesome.
I always viewed VC as a black box when I was on the entrepreneur side. I never really understood how decisions got made, and then when I came over to the venture side, I began to see behind the curtain, and I began to realize the algorithm that VCs undertake, and it's a pretty simple one: it's around team, TAM, and business model. And so a team—there are a number of things that we look for in the quality of the team, and it's really about having somebody who can grow, not just into who they are today, having an earned secret about the domain that they're operating in, a grittiness and a set of attributes that we think will allow them to execute in the moment, but also the growth mindset as Stanford professor Carol Dwek talks about, which allows them to, to us to have a belief that they will grow um to be something special. You know, when I met um Ragu, the founder of Falcon X, in my HBS classroom over 11 years ago, Ragu was a young uh executive, a middle middle-level product manager coming out of the Google ecosystem, had been in a startup that hadn't gone anywhere, and I had to have a hypothesis about Ragu's potential to grow into the executive that he is today, running an a company valued at $8 billion with, you know, a dream and a vision to go public in the near future. So it's really about the growth journey that you think that executive, that founder, can be on. So that's team, and then TAM, you know, we look for waves of technology, secular waves that we think will allow for these companies to be propelled so that they don't have to have the the the entire burden of of creating the market on their small little balance sheet that they're a part of some secular wave uh and then also that it's going to be a wildly disruptive opportunity with some really interesting fundamentals about business model quality, and that and that leads to the business model conversation. Um, I I am a believer in the MBA; I am a believer in strategy and competitive advantage and competitive modes, and so we do apply all these classic theories—you know, Clay Christensen's disruptive theory, Michael Porter's five forces, um the strategic thinking out of the BCG ecosystem where I started my career—like we apply that thinking to markets to determine where will value accrue, and we're in a moment right now where the AI market, the AI software market, is so dynamic and so cluttered and so chaotic that having incisive strategic thinking and a hypothesis about where value will accrue ends up being a quite important attribute.
On that personal story about like the Falcon X fund, right? So like 11 years ago, you see this guy coming out of like, you know, obviously Google's product is very well-known, especially 11 years ago, very competitive as well. So clearly he's at HBS, again a very competitive school. What, what were like the the the things he was doing which kind of like gave you confidence like, "Hey, he's going to build a massive unicorn company," which obviously is very rare uh as well, because I think you meet a lot of smart people and you meet a lot of like experienced founders, but it's hard to kind of predict like you're going to end up building an $8 billion valued company, which is which is very rare by itself.
Yeah, it's a great question, and look, luck is involved for sure. But as you noted, there are 900 graduates of HBS every year; there are thousands of employees at Google. Very few of those individuals actually create a unicorn company like Regu has done at Falcon X. And I think the growth mindset is one thing, but there's something else that I'll uh point to: there's this interesting combination that I see the best founders having of a bit of stubbornness and almost arrogance, but just this high conviction that they are right and the rest of the world is wrong—a willingness to be a contrarian thinker and to stick to their guns in combination with an open-mindedness and an ability to adjust and listen to feedback and grow from that feedback and grow from experience. And it's that, that really interesting combination of, you know, as F. Scott Fitzgerald talked about having, you know, the sign of a rare intelligence is someone who can hold two truths in their mind at the same time. It's this rare two truths that I see founders that are really quite brilliant being able to hold, and Reggu is an example of that. He believed that crypto would be an institutional asset class, that the Goldman's and the JP Morgans and the Black Rocks of the world would embrace and adopt crypto, and we've had a few crypto winners and such that that hypothesis in 2017, 2018, 2019 looked pretty dumb, looked pretty wrong. Um, and at the same time, he's also been able to adjust his model and evolve the Falcon X story in such a way as to navigate to where the opportunity lies. So um, having these two opposing views, I think, is a really interesting attribute of great entrepreneurs that I've seen.
Now, how do you test for that prima? Like that's the really the hard thing in in conversations with entrepreneurs where you're testing their high conviction of ideas, but also testing their willingness to listen and to adjust. That's really the the art and science of the business.
This is a little bit of a tangent, but I think this is an interesting tangent to to go on. I think in Silicon Valley, it's almost like cool to say that you know that you shouldn't do an MBA, that you shouldn't go to college, that you shouldn't do a postgrad. The irony of these uh situations always is that people who say this have always gone to Stanford or or Harvard. Um, so I just wanted to, you know, get your take on that, right? Clearly uh knowledge is is pretty abundant and ungated at this point, so you know, is is is the knowledge argument there? The credential is obviously very important; the network is obviously very important. So I just wanted to I guess get get the insider take, right? So what, where do you stand on on the value of a of a postgraduate, graduate, or Harvard MBA today?
Look, we're in the age of AI where building is a commodity; anybody can build. And so I would argue, and I know this is a bit maybe controversial, that technical skills are now being somewhat devalued, and strategic skills, leadership skills are more valuable. Because if the tools to build are easier and easier to use and more and more powerful and cheaper, then execution capacity is expanded tremendously, and now the question is what do you execute on? What product do you build? What strategy do you employ? What target market do you pursue? And so I think we're in this really cool moment where we're back to the revenge of the MBAs. We're in a moment where the business leaders who have that strategic insight and who have that leadership ability are going to have the opportunity, if they embrace the builder mindset, are going to have the opportunity to really outexecute the technical co-founders. Now, we back plenty of technical co-founders, but it used to be that we would insist on PhDs and deep understanding of machine learning for our AI-based startups. Now we're much more open to a domain expert who has a builder mindset and who has a really high throughput, high capacity to execute with strategic insight.
Yeah, yeah. I guess that's a a good segue to ask about your new book, um, *The Experimentation Machine: Finding Product Market Fit in the Age of AI*. So yeah, can you tell us a little bit about a book, and then you know we would love to dig in a little bit deeper about about it as well.
So the book is, in essence, this combination of the timeless methods for finding product market fit—strategic thinking, must-have value proposition, customer discovery, deep insights into customer value prop and go-to-market opportunities—with these timely AI tools and methods um to accelerate the journey.
Awesome. Can you give us a little bit more, maybe a little bit of uh of I guess a tactical or more practical view? So let's just say you know Hassan and I are you know both co-founders. If you had to give us some very tactical advice today on how we can use AI tools to accelerate finding product market fit, um how how do you think we should go about it?
What I encourage my students and my founders to do is leverage the AI tools to the max. First, create a user persona. Have the AI tools help you create a user persona for your target user, and then feed it all the data you can—market research data, perhaps even interviews from your customer discovery calls, perhaps expert calls and transcripts from TGUIs or other expert network interviews, and then begin to even podcast interviews. And then begin to have these queries. You can use Notebook LM, you can use custom GPTs, um to use this rag-based system to constrain the the data set to the content that you feed it, and then have it create this user persona, so that 7 by 24 you have access to your customer, to your user. Um, probe on must-have value propositions. Use the AI to insist that what you're doing is a must-have, and then use the tools naturally to prototype. As everyone knows, you can use the modern software development tools and the no-code tools to create prototypes such that you don't need to write a single line of code. Build those prototypes quickly and efficiently, and then get those prototypes in front of your customers. Create multiple demos, that versions of the demos that are customized for your customer using the video generation tools like HeyGen or ElevenLabs. Um, and so you know you sort of think of this pipeline of value prop discovery and experiments, go-to-market discovery and experiments, and then finally business model profit formula discovery and experiments. At every step of those three components, you can use AI to automate your process and to run more experiments efficiently and effectively than ever before.
So Jeff, like you kind of mentioned this 10x founder concept, right? Like, and and one thing I I think about is that like you know it's very crucial now for folks to use AI tools, and I think you talk about that in the book as as well that like, hey, this is kind of like critical for you, right? Like this is kind like which allows you to become that 10x founder, you can get competitive advantages. If you had to break it down for founders who are about to start their company, maybe like simple three lines like, hey, like if you don't use these tools, what are the risks you are kind of exposing yourself to? Like what are your competition kind of doing? I think that'll be very helpful for our audience to hear.
Look, I start the book with this notion that AI is not going to replace founders anytime soon, but founders who use AI are absolutely going to replace founders who don't. There's this notion Sam Altman talks about of the billiondollar solo entrepreneur. I don't know if that's going to happen anytime soon, but 50-person, 25-person, 10-person teams that are leveraging AI to the hilt and allowing them to execute as 10x founders, they are absolutely going to outexecute the larger teams that don't.
In any kind of like key story, maybe like from your portfolio where you kind of seen this kind of play out where like founders have actually been very successful in leveraging these kind of tools you've talked about in the book, or maybe like even any other company which you kind of like interviewed as part of this uh—I think that'll be very interesting to hear as well.
Yeah, so in the book, I I give a lot of case studies—case studies that I've written about for my HBS class or case studies that I've experienced through the investments at Flybridge. I'll give one example: there's a company in New York, actually between New York and Boston, called Topline Pro. Was founded by two former HBS students who dropped out. I had nothing to do with it, but they decided after their first year to drop out, go to Y Combinator, and start the company. And they started the company pre-ChatGPT, and they had this thesis that service pros—think landscapers and painters and electricians—would need a more robust and modern tool and platform for marketing and customer service to build their websites and to interact with customers, to book appointments, and then to serve those customers. And so they set out to build this tool just as ChatGPT um was about to come out. They built it originally with uh leveraging the GPT-2 SDK and then GPT-3 SDK and then 3.5 and ChatGPT, and suddenly you get this Cambrian explosion. So that today, Topline Pro allows you to automatically build the website for the pro, scraping their Facebook pages and Instagram pages for photos; it creates the website, and then it creates a video of a sales rep, personalized and customized, showing the website to the pro, addressed to them and saying, "Hey Hassan, I built this website for your landscaping business; I used your photos; I used your copy and your descriptions of your services. Does this look interesting to you? I'm happy to let you subscribe to this and take over this service." And it's an incredible response rate that they've achieved; they've gone from a 1% response rate to a 10% response rate plus, and then they have a chatbot that services the pros and allows them to enable the pros to change the website, throw up new photos that they can just text to the chatbot, change the copy, add a new service or customer testimonial or customer case study, and handle upsell and cross-sell opportunities. And so that company has grown to being a very substantial company serving thousands of service pros with just a few dozen employees.
That's amazing. Like I I I love like how they've kind of taken different types of like tools in the AI kind of like plethora, like massive tools we have, and like figured out like sales outcomes, uh, combination of like conversational and the like. Most probably these service pros feel like, hey, there's getting like a custom outcome, uh, and like someone serving them at a customized level, which is which I think AI is like really helping you do.
I think that's right. You know, you there's this notion of uh service level being proportional to revenue in h in software companies and service companies, where the more revenue they can obtain, the greater a service they can afford to provide. AI is bending that curve, and now companies can provide a really high service level and really high customized, almost personalized capabilities for very small businesses. It's a pretty special moment in that regard.
As a founder myself, I think one challenge I have is always thinking about like, hey, how do you balance the need for speed with like, you know, hey, let's build a better product because I'm serving an enterprise or like thoughtful decision-making; let's get more feedback from the users. Maybe like if you can share a bit more about that, because I think especially in today's age where shipping velocity is so fast, you're seeing all these companies like getting to 10 million ARR in like months, which used to take years. There is some sort of fog as a founder as well, like, hey, maybe I need to like ship fast, uh, but then you like, how do you, how do I balance that?
Yeah, it's a great question, and this is why I think there's this balance of AI-fueled execution with fundamental old-fashioned strategic thinking. You've got to really ground yourself in: what's the customer problem? Is it a must-have versus nice-to-have? Is it valuable? Is there a high willingness to pay? And is there some sustainable competitive advantage if I can solve that problem and retain that customer in a very sticky fashion? And that
Requires slow thinking; execution requires quick thinking. Uh, Daniel Kahneman, the behavioral economist Nobel Prize winner, had this great book, *Thinking Fast and Slow*, and so I think founders need to have that sort of similar kind of approach, which is execute fast but strategically think slow.
Yeah, fair enough. I wanted to get your take a little bit because, you know, you are a venture capitalist; you invest in a lot of companies. I'm sure you have very strong opinions. Uh, one of the most kind of, you know, surprising things over the last 12 months—you know, anything that was built on the app layer was just called a GPT wrapper—but it looks like, you know, it's been a very sharp 180 on that opinion, and it's like, no, no, no. It's like now it's like, you know, apps are valuable. Uh, maybe it's models that are getting commoditized. So, so you know, in this, I would say, tension between the app layer and the model layer, like where do you see things heading? Like, do you feel like models are going to get commoditized, or feel like that's maybe an overreaction at this point? And the app layer, you know, definitely seems like it's on the up and up right now. Does there need to be any caution on that side, or do you think like maybe we were maybe underestimating the app layer heavily over the last 12 to 24 months?
So at Flybridge, we're really focused on the app layer and some of the enabling tools around the app layer. We've always believed that there would be a massive commoditization at the infrastructure layer. We saw that in the Web 1.0 era. Today, we don't talk about companies that are multi-billion dollar companies built on SSL protocol and HTTPS protocols. And in the cloud era, we saw massive deflation of prices and cloud wars and the cloud dividend such that the app layer has been incredibly rich. And I think similarly, we're seeing this in the AI world; it's just beginning to shift, and that's a natural technology life cycle adoption curve dynamic. What I think we're grossly overestimating is absorption speed. Companies are made of humans; humans have organizational inertia and brain inertia. It's difficult for humans to change habits, and it's difficult for organizations to change habits. And so there is a bit of hand-holding and patience that companies need to exhibit in order to get their customers to change over to their way of doing things. And so I think the best founders are building world-class distribution and world-class service in addition to world-class and innovative products. And I think the founders that only focus on building products are going to really struggle to scale past the few experiments and the few pilots that they get. It's easy to get to that $1 million or $5 million ARR with pilots; it's hard to get to $50 million or $100 million. My partner, Chip Hazard, refers to this as the SaaS valley of death. It's that moment between $20 and $50 million where you have to execute on repeatable distribution and go beyond one product and one distribution channel tactic to multiple products and multiple distribution, um, uh, motions and different, uh, you know, tactics and partners and globalization. So that's going to be really the test for a lot of these companies.
Yeah, you mentioned something there about building distribution. So one of, you know, sort of like the very strong opinions that Hussam and I hold is when we started out in tech, it was everyone just wanted to invest in technical founders; nobody wanted to invest outside of technical founders. So Hussam is technical; I'm not technical. So this was, you know, sort of going in very bad news for me, cuz if I ever wanted to become a founder, I needed to find someone like Hussam who was, uh, who was technical. But it seems like that's changing, right? I think 10 years ago, distribution was abundant, like, you know, you can just show up to Facebook ads and spend a bunch of money in Facebook ads and scale it way to like $10, $20, even $100 million in, in, in revenue in some cases. That's fully saturated now, right? Uh, TikTok seems like it's saturated; a lot of just paid ads is saturated; SEO is is pretty hard; it's pretty competitive; pretty saturated. So it seems like, you know, owning, having an audience, or owning distribution, or even having a podcast at this point seems like a very valuable skill, whereas like 5 years ago, it was sort of, you know, um, sort of like looked down upon as like a cool side project.
Maybe building on that question, like, you know, over the next 5–10 years, as we're sort of, you know, entering this AI wave now, uh, is is technical less valuable, or is technical just as valuable, and distribution skills are getting uh more valuable? We'll just love to get your take on that as well.
I think brand and distribution is getting more valuable. I think technical skills are becoming more commoditized when it comes to the execution portion of the technical skill, but the strategic thinking behind the technical skill—that intersection of customer need and products to deliver on that need—I think that's going to be highly valuable. There's an amazing product management leader, Marty Cagan—both of you guys may know—where he talks about this new AI era, the role of product management. Marty recently wrote a blog post where he said, "Look, product execution is becoming commoditized, but the conceptual parts of the product creation process—that's still highly, highly valued." I talk about this a bunch in the book, which is that you've got these timeless methods and timely tools. The timeless methods—timeless methods of strategic thinking and judgment—that's essentially human, and I don't think that goes away.
No, no, I think Geoff, like that, that's so spot on. Like, I find myself, like, as we kind of are building things on our own, like most of my time is kind of spent on like system thinking. Even like engineers we interview or like people want to bring on, I think folks who have done system architecture level stuff, they're so much more valuable, right? Like if you have like end-to-end perspective, how to architect a product, how to experiment with those things—that that has become so much more valuable because that's still hard. Like you can leverage AI to like execute code or like, you know, help you with a specific feature, but you still have to work with them to kind of think about like, hey, what is what is the system you're trying to build there? How do I want the outcome to be delivered? Like those things are still need to be thought about—the strategic perspective, talking to your customer, kind of getting it into like a product level, uh, architecture. Like, so spot on.
Yeah, and I, I really um struggle with the following question, Hussam, which I'll throw out to your audience and to you guys and see what you think. I was at OpenAI last week talking about my book, and we were having this conversation with the OpenAI leadership team, which is that strategic thinking and judgment and systems thinking comes from many years of experience and many reps. And so mid-level and senior-level people who have done that hard work when they were younger now have that holistic systems-level view, but if the AI tools of today are going to do that work for the junior engineers and the junior product managers of today, how will tomorrow's mid-level and senior-level people develop that judgment and that insight?
So I totally agree with you. I think a lot of times I pattern match, like, oh, I did this in this startup; this is how we solve this problem; I can see we can solve this problem this way. But now we have these AI tools; let me go like experiment with them and see if there's a better way or not. But there's a lot of pattern matching, like, hey, Mike Smith at Snowflake building previous startups, but also like we just recently interviewed Aaron Levie, the Box CEO, and one, one thing he mentioned which I, which I thought like made a lot of sense is like when I'm hiring like younger individuals, my expectation is that they can ramp up faster with these AI tools, and they can leverage and get to answers faster. So my kind of thought process there is that if you're a junior person, you're leveraging AI and getting into these more meaningful conversations with more insight, so you can kind of have more osmosis faster. So like what osmosis took me maybe 2 years to learn starting out can take someone 6 months. But you still need to have an appetite to learn, ask the tough questions to like other folks who are experienced, but with AI you should be able to ask better questions and maybe go get to like the right outcome. Maybe that's that's one way of thinking about it. I don't know, Zen, if you have any thoughts on that.
I think there's there's always like these layers of abstractions that are happening. Like the amount of things you can do with the press of a button or with like a command just keep going up and up and up. So I think, you know, I think the the really most hardcore kind, you know, software engineering must be that like done for hardware, right? And then, you know, if you're writing in C, like that's obviously very difficult. Most software developers or engineers haven't really touched a lot of those foundational languages; they're, you know, probably using some kind of, you know, framework like React or or JS or something along those lines. So I just think maybe it just becomes a case of that it's like, you know, I built multiple businesses on the internet without having written any code myself. So, you know, for example, like we're recording this podcast right now; this is a product; I have zero idea uh about how the code behind any of this stuff works, right? So I think I, I think obviously I'm not saying that foundational technical knowledge is not going to be valuable going forward, but I think as the layers of abstraction go up and as infrastructure just becomes more reliable, a lot of companies are just going to be built with people who really are uh skilled with that layer of abstraction. So yeah, I think people, you know, um, people build $100 million uh Shopify businesses like that's that's doable, right? So without knowing any any technical skills. So yeah, I think there's there's value to technical stuff, but I think with the layer of abstraction that's coming in, we're just going to see more and more businesses uh being built.
Hussam and I, actually the first uh podcast interview that Hussam and I had done uh was with this guy called Blake Anderson, fresh out of college. And what he basically did is he created uh a dating—it was sort of like a dating app; you basically take a screenshot of your conversation and you upload it into this app, and then it basically tells you what your next pickup line or so is going to be. But the app was very successful; he got it to, you know, I think $5 or $6 million, fresh out of college. So we're like, "Let's interview him; let's see what's going on." The most surprising thing about the interview is that he was not technical at all. What he'd do is he'd go to Figma; he'd draw up an image of what he wanted the screen within the app to look like, and then he'd go to ChatGPT and say, "Hey, I want to build an app, and this is the screen that I want to build; can you help me generate the code?" And every time he ran into a bug, he'd go back to ChatGPT and say, "Okay, help me figure out this bug." So without ever having coded himself, he basically built this app that was doing $6 million ARR at the time that I had talked to him. But I try to explain this to someone 5 years ago; they it would it would make no sense, like, you know, try telling someone, you know, I, I tried talking to an internet chatbot and it helped me build like a $6 million app; it would make no sense. But that's just the layer of abstraction and the kind of businesses that they can get built today.
So I love that because it's a great story which reinforces my comment of the revenge of the MBA and the revenge of the liberal arts major. It's having this ability to have insight, conceptual thinking, systems thinking, levels of abstraction thinking, and then also the interpersonal skills and the understanding of humans to have that nuanced view of what will be successful, what might not be successful—100%.
So, so Geoff, like I, I was very excited about asking you two questions on this podcast, right? Because you, you teach, you, you like teach a lot of cases at HBS, and I think your class, you invite a lot of founders to come in; some, some of these founders have done really interesting things. So I'm very curious to know like maybe one story from a founder which, you know, is very memorable over the years which you can still remember, if you can share that with our audience, uh, that'll be the first one. So let's start there, and then then I, I'll ask a follow-up.
So one of the most popular cases in my class right now is Sober Sidekick, and the founder, Chris Thompson, built this app which is for addicts to help them get sober and stay sober. Um, he built this while himself in a halfway home with a broken laptop. He coded this thing up by himself, built it, provided all the customer service, all the maintenance, um, and has built it now into an app that's serving hundreds of thousands of daily active users who are in the process of st getting and staying sober. And what's so amazing about that company—Chris is an extraordinary 10x founder; he runs the whole business on Notion; Notion AI has a handful of employees, and he's achieved so much—but what is the grittiness of that story? Chris was so passionate about serving his customer that he pivoted his business model from getting paid by the clinics that took the addicts in, who get uh paid and compensated in almost a lead-gen mercenary way, uh, to shifting to health plans and helping outcomes and reducing the cost of servicing these individuals as they get sober and get healthy. And so he, he just showed incredible grittiness, but also strategic thinking in framing the business model, keeping his true north and serving the addicts while finding out um who had a high willingness to pay for this service.
That's a great story. Thanks for sharing that. So, so obviously you mentored a lot of students as well as part of this journey, and any, any kind of students like, you know, you kind of like, you learned some interesting lessons from them, kind of like still remember, and you kind of changed your class because of that, or your investing strategy changed, or maybe you kind of like thought, "Hey, I should write this book and maybe include this in the story." I love to hear maybe a story like that as well.
Yeah, one of my most successful joiners—so not just founders but joiners who are employees, 2 to 200—one of my su most successful joiners out of my HBS class is Zach Kirkhorn. Zach joined Tesla pre-IPO, worked in the finance department for a couple years, came to HBS, got his MBA, and returned to Tesla in finance, and over a few years grew to becoming the CFO of Tesla and had an amazing 10-year journey helping take the company from a billion-dollar public company to a trillion-dollar valued company at the time when he retired just recently. And I was asking Zach in a little conversation I organized with about a dozen of my students, I said, "How did you know to pick Tesla?" Because joining a rocket ship and having that judgment is just as important as the judgment of coming up with a good startup idea. And Zach said, "You know, I just loved cars; I was obsessed with cars," and it just reminded me and reminded my students that sometimes it's not just thinking rationally but also thinking emotionally, with your heart, about what are you passionate about, what do you want to spend the next 10 years of your life working on, and go find a great company that's on that mission. That that's an amazing story. And like, congrats to Zach on all the success from like from like that early days of Tesla; I'm sure like that was like a crazy ride. And like a great conversation for your students as well.
Uh, last question about the class and uh, what is the craziest thing which has happened in the class?
Okay, so one of the craziest things that's happened in my class is that one of my former students was indicted uh and and put in jail for fraud by the SEC and by the US Justice Department. And so it's just a reminder of there's a right way to go about building your business, and the hustle culture and the fake-it-till-you-make-it culture has a point, a line that you need to be very careful about. And I think in this age of AI, having our business leaders think strategically about ethical issues and about systematic biases is going to be more critical than ever—than ever.
Just one last question before we wrap up. Uh, we're seeing a lot of a lot of AI companies in Silicon Valley. I see Gary Tan and YC; obviously, I'm sure they have a vested interest in saying you should be in Silicon Valley, but you're obviously in Boston; you're in the New York ecosystem as well. What's your take on on SF versus basically any other place on Earth right now when it comes to uh building an AI startup?
Look, I think Gary Tan is wrong. I think it's an amazing time to be in New York and in Boston and building an AI company. To build a great AI company in today's environment, you need to be close to customers, and the next wave of great application-layer AI companies are going to be those companies that have those nuanced insights into customer requirements, and that's going to come from being close to customers. And startups selling to startups can be a very dangerous game, as we've seen time and time again in history, whereas startups selling to big companies, small businesses, having an understanding of the common consumer—that's a winning formula.
Jeff, thank you so much for this; this has been a very insightful uh conversation. *The Experimentation Machine*, the book, please go check it out, uh, and with that, let's wrap it up, and we will see everyone again next week. Take care.