Event: How to Build a Successful AI Native Startup on 08/09/2024
We had a talk session with John Whaley, the founder of Inception Studio, an AI accelerator in San Francisco. The event was held by Glasp and JETRO in Tokyo, Japan.
About the Event
We had a talk session and networking event with John Whaley, the founder of Inception Studio, an AI accelerator program in San Francisco.
Since the start of the program in November 2022, over 27 companies have been formed through Inception Studio, with over a dozen successfully raising funding from top venture firms and angel investors. This is an extremely valuable opportunity to meet the founders of San Francisco's local AI accelerator in person. We hope you will join us!
* The language used in the event is both English and Japanese.
Application to Inception Studio Japan:
About John Whaley
Dr. John Whaley is a prominent figure in the field of technology and AI, known for his extensive experience and leadership in the startup ecosystem. With a strong background in computer science and entrepreneurship, John has been instrumental in founding and scaling multiple successful startups. He is currently focused on fostering innovation and growth in the generative AI space through various initiatives and collaborations.
Career Highlights:
- Experienced founder. John has founded and led three startups, leveraging his expertise in AI and technology to drive growth and innovation.
- Stanford Faculty. John is an Adjunct Lecturer in the Computer Science department at Stanford University.
- Accomplished speaker and industry thought leader. John regularly gives keynote and invited lecturers at major industry and academic conferences in AI, security, and entrepreneurship.
Follow John on social:
Inception Studio
Inception Studio is a non-profit accelerator built for the most experienced world-class entrepreneurs and builders at the earliest stages of starting their next AI company. We organize exclusive 72-hour cohort retreats and other events to help our Inception Founder community find cofounders and launch their companies. Inception specializes in stages from "I just became fully committed to starting my next AI company" to "Our company is ready to raise a pre-seed round".
JETRO information
JETRO is a government-related organization promoting FDI in Japan and business partnerships between Japan and the world.
JETRO LinkedIn pages:
- https://www.linkedin.com/company/jetro-japanexternaltradeorganization/
- https://www.linkedin.com/company/jetro-collaborate-and-invest-japan/
- https://www.linkedin.com/company/jetro-global-connection/
Organizer
Kei Watanabe
Co-founder of Glasp
X: @KeiWatanabe17
Transcripts
JETRO: Accelerator but currently this is open collaborating so if you have a friend who works in AG SE please introduce this program. The application date is August 19th, so yeah, I can introduce myself. And other than that, this year as a new initiative, collaborating with Texate and 500, we launched a new program, like an investment-based program. So, collaborating with the Japanese potential L investment, like a real estate farm, they created a fund in Japan and they do the investment-based acceleration program. Unfortunately, the deadline for the text is over but this initiative continues in the upcoming three years, so maybe you can apply for the next year. We also collaborated with a tech conference like CES, Disrup, or something like that. And over there, in order to support marketing in international markets, we just created a Japan and you can show your product in the JA. We also have our own and over there, we just promote our upcoming program to apply or some other services we currently have on the pipeline of the overseas VC cap CVS. So we know the kind of startups they are looking for and some of them are actively interested in the Japanese startup ecosystem. So we sometimes set up one-to-one meetings between Japanese entrepreneurs and overseas capitalists. So if I reach out to some entrepreneurs here by LinkedIn or some, yeah, please feel free to, you know, get accepted to our offer because we can sometimes try to get connected with startups and overseas investors. Sometimes, yeah, we offer a startup information map, like this kind of market map. Yeah, that's it. So this is my card, if you like to know more about our services, please scan my QR code and feel free to contact me. Thank you. So I’ll just leave it there, okay?
John: Thanks everyone for coming. I literally just arrived yesterday in Japan. I'm still really jet-lagged, so sorry for being a little bit out of it. But anyway, a little bit about me. I lived in Japan for one year a long time ago. This was before I started my PhD. I graduated from MIT with my undergrad then came to Japan for one year, worked at IBM in Japan, and then started my PhD at Stanford. I started my first company while I was a PhD student at Stanford. That one eventually got acquired about 10 years later, and then I started my second company and that one got acquired about four years ago. I just started my third company about a year ago. All three of the companies have been in cybersecurity. I'm also on the faculty at Stanford. I teach in the computer science department about generative AI. The class that I taught is called CS224G, G is for generative like GPT, and last quarter it was the class that filled up the fastest of any class at Stanford. Within 3 minutes, it was completely full because this is a very hot topic right now. Everybody is very interested in large language models. I'm lucky enough to be able to teach one of the most popular classes at Stanford. I also run an accelerator program called Inception Studio which I'll talk to you a little bit more about. We started about two years ago. We started right around the time our first event was about two weeks before ChatGPT came out, so our timing was perfect. We've had a lot of success through that. Being there in the heart of Silicon Valley, and also being there now for almost 25 years, teaching at Stanford, and running the Inception Studio accelerator, I get a very unique view of generative AI from the research standpoint, the practical standpoint, the academics, the startup, and the commercial standpoints. I think this is an amazing time to go start a company. There are going to be some trillion-dollar companies or companies that will be formed this year or next year around this time frame. But there's also a lot of hype right now. We're kind of in that bubble, and I would say most of the companies that are formed right now, maybe 97-98% of them won't go anywhere. They'll end up just going out of business or getting acquired or not being successful. There's a lot of hype but there's also a lot of very interesting. It’s a very exciting time to be working in this space. One observation is now is a really great time to be building. I think this is actually the best time to build that I've seen in my lifetime so far. We're at this very interesting time in history where we have this brand new capability called large language models, generative AI, and foundational models. No one fully understands what they are capable of, even the people who are creating the models. But I think almost everyone agrees that these are going to be very important. This is a time when there are really big benefits to being a builder, like an engineer who can actually build things. Now the definition of an engineer has changed because it used to mean you had to write code, like Python code. Now because of prompt engineering, if you can write just natural language you can often get the LLM to do very interesting things and actually build software using natural language. It’s a really interesting time to be in this space. People who have that builder gene, who are interested in experimentation and building things, can explore the edges of the system, understand what the actual capabilities are, and being fully fluent in this type of technology becomes a superpower. This is going to be true for the next few years. People who really understand generative AI and how to use those tools, what they're good for and what they're not good for, are going to have a great advantage compared to people who don't. I see this already. There are many startup companies that may have three or four people but can act like a much larger company because they're fully embracing these generative AI tools. But there are not really any books or courses that cover this, and because the space moves so quickly, every week there are some new capabilities. The best way to learn is to just try to build something real. When you try to build something real, you're going to explore the edges of the system and understand a lot more about how this works. There are lots of excuses you can tell yourself about why you shouldn't build now or why you shouldn't start a company now. It's like, oh well, there are a lot of tech layoffs happening, big companies are cutting their workforces, maybe I don't want to leave my stable job. Or sometimes they would say generative AI is very overhyped, there are too many startups, and most of them are not going to survive, so it's not a good time to start. Or, there’s no really proven successful business model yet. The only business model seems to be like you light a bunch of money on fire, or you use up a lot of GPUs and spend a lot of money, but are you actually going to get the revenue after that? That makes it very risky. Or, the incumbents like Google and Meta have way too many advantages because they have access to data, access to compute like GPUs, distribution, they already have users, all of this. So it makes it seem like small players can't compete with the big ones. The truth is those statements are all partially true. It’s true, that most people should not leave their jobs and try to do a startup or join an early-stage startup. But if you're very entrepreneurial, if you have ambition, if you have some conviction on some idea that you really want to build, and you're at that kind of life stage where doing a startup makes sense, now is a great time. It's probably the best time that I've seen in my lifetime so far. The strongest benefits come to people who can build, people who are either engineers or have an engineering mentality, to be able to build things. The big winners in the future, the winning products and the winning ideas, are not going to come from some big company where there's market research and focus groups. We're in a very disruptive time right now. The winners are going to be these AI native companies, the ones that are fully embracing AI, really understand the problems, and are actually able to come up with very innovative solutions. They are AI natives. They are going to be reimagining the way industries work and developing brand-new capabilities. Those types of things are not going to come from big companies, they are going to come from startups and people who are going to be very disruptive. Big companies have a huge, huge, huge advantage because they already have existing revenues. They're not going to cannibalize their revenues to risk their revenues. They are naturally tied in terms of being able to be as disruptive as startups can be. It's a great time to build. Now to be clear, we're also in this massive hype cycle. I'm not sure if you've seen this before, this is called the Gartner Hype Cycle. You start with the innovation trigger, go up to the peak of inflated expectations, then the trough of disillusionment, and then eventually you kind of flatten out. I'm curious where people think we are on this curve right now. Some people maybe think we're here, but we still have a long way to go. I don't really think so. I think we're probably somewhere around here or here or maybe here. Maybe we’re at the peak, maybe a little before the peak, actually a little bit past the peak. I've seen there's obviously a lot of hype and stuff happening, and I get the sense we already went past the peak. The hype around generative AI is just massive, so huge. We have people in generative AI like Sam Altman who say, well, we better work on UBI, Universal Basic Income because in the near future, people won't need to work anymore because AI is going to do all of our jobs. I'm not sure if any of you have actually tried and used any of these systems or even understand the trajectory of these. There is no possible way that is coming anytime soon. We’re so far away from that. Now that being said, I think that the promise is impossible. The promise, the things that are being promised, will not happen within the next five years, at least five to 10 years. I can say that with absolute certainty. We’re not going to have a situation where AI is going to take all of our jobs within the next decade. It's not going to happen. But that’s the promise, they said this is going to be the biggest revolution that's ever happened. Then people make the argument, well, this is getting different. They say, in the past, every technology followed this curve but AI is different, we’re just going to keep going up and to the right forever. It’s going to be an exponent. I don't think so. There is a lot of hype right now. People are spending huge amounts of money on AI systems and they're already starting to see the results of, this doesn't work as well as was promised. We’re spending all this money, where is this going? Where's the practical use of this? You're already starting to see some of that. The higher the peak, the worse the crash is going to be, and so this is going to happen at some point. That’s the other thing I can say with certainty. It’s going to follow this pattern. The only thing you don't know is what the scale on the X-axis is. Is this over the course of one year, or 10 years? When is this going to happen? I don't know, maybe it could happen this year, next year, or the year after. It will happen, and you have to be prepared for it. What that means is sometime in the not very distant future it's going to be very uncool to be an AI startup founder or to be working in AI. People will say, oh, another one of those. It's kind of like when web3 and blockchain were so huge and then it became not so cool to say, oh I'm a web3 founder. Not so cool anymore, right? Because they've already gone down this slope. This will happen for AI as well because there's a glut, way too many startups, and honestly, ones that are not very good. They don't have good strong fundamentals, and they're not going to be successful companies, but they're still getting a lot of money because there's so much hype right now that a lot of not-good founders and not-good ideas are getting funded. Your goal should not be to get funded, your goal should be to be successful in the long term because you don't want to waste your time and your life on something for two to three years that is not going to go anywhere. Even if you can raise money, it does not mean it's a good idea and it doesn't mean that you should spend your time on it because your time is worth a lot more than investors' money, especially now because there's lots and lots of investors who are excited about AI. This starts with the LPs, the limited partners, the ones who give all the funds to the VCs. They are saying, I don't want to hear about anything other than AI. I feel really bad for anybody who's trying to start a non-AI company right now. It's really hard to do because all the oxygen is being sucked out of the room by AI. On the flip side, all you have to do is take your not-very-good business model, add a registered AI domain name, and suddenly you can instantly raise money at a high valuation. We're definitely in a very interesting time. It’s an incredible time, an incredible opportunity. The next trillion-dollar company, I think, is going to be founded this year or next year. The real question is, how do you find the winners? Who's going to win in this? I mentioned there are probably 98% of the companies that are going to fail and 2% that will succeed. Let me say this: For the early stage, this is not for the later stage, once you have Series B and Series C and those types of things, those metrics are different. But for the early stage, it's all about the team and the early team, the people you have. It’s because the landscape is changing very quickly, it's highly competitive. The AI companies are so early, they may have some data but it’s not meaningful yet because the data points are so few. Even if you say, I got some traction, it doesn't really translate into meaning that you're going to just follow this straight line. It's unlikely to happen. For a startup to be successful, the idea approach is probably going to change. Look at Slack. How did Slack become Slack? They tried to build a gaming company and then they accidentally built Slack, and then now Slack became Slack. Or Instagram. Instagram used to be a check-in app where you could go to a location and check-in, like the Foursquare, that kind of app. Then they pivoted and changed, now it became Instagram, and then obviously became what it is today. Usually, the story of a successful startup is something like that. You start off going after this problem and then you end up working on something else. Given that, the idea is going to change. The other thing I'll say about this is that I teach this class at Stanford where we have groups of students that form teams and end up working on projects. It's amazing how many of those projects, these are students, these are just kind of undergrads, they come up with exactly the same idea as a startup that later on goes and raises 50 million or $100 million. Exactly the same idea. That’s the truth. If you say something like, hey, I want you to apply generative AI to the legal profession, go whatever. There are probably about five ideas that everyone is going to have the same idea. There's nothing really in it, there's nothing unique to copy. If that's the case, what that means is ideas are so cheap. It's easy to come up with ideas. What's going to actually make you successful is execution and being very responsive to what customers say, different opportunities, being able to pivot and change, and being agile. That's what's going to make you successful. Hustling and getting customers, that's all the stuff that's going to make you successful. Your idea is not the thing that's going to make you successful. I can guarantee, especially in Silicon Valley where it's like everyone is doing a startup, you can go to any random coffee shop and you'll see at least half the people there are founders or would-be founders who are working on their generative AI startup. The ideas are so cheap. It's all about the execution. If you say that, okay look, it's about the execution, it's about the team, the team that you have, what makes a good team in this kind of AI Renaissance? These are the traits of successful early-stage AI founders. Number one, you want to find somebody who's impressive in some way. They have some type of special skill, it's like you meet them and say, well, this person is an impressive person. This becomes important because you need to find co-founders, raise money, get customers, hire people, you have to convince somebody to quit their job and come work with you. So there has to be something impressive about them. They should have some deep expertise. That could either be domain expertise or about industry, could be around technology, but something. You need to have depth, you can't just be shallow across everything, you need to have deep expertise to be successful. You have to be autodidactic, which basically means you can learn by yourself. You don't need somebody to teach you, you can just grab the documentation or you can try it out, you can learn by yourself and be naturally curious. You have to have a high-magnitude, high-horsepower type of person. If you move too slowly, you're not going to be successful at the early stage. You have to be capable of first-principle thinking. You have to not just follow exactly the pattern that you've seen before, you have to be able to think from first principles and reinvent things that you do. You have to be a good storyteller. You have to convince people, your co-founders to join you, your teammates, you have to convince investors to give you money, and you have to convince customers to sign on with you. If you don't have the skills as a storyteller and be convincing, you're going to have a hard time. They should have raw leadership skills to scale an organization. It's one thing to be successful as an individual or a small group, but to be truly successful you basically have to have some leadership skills so that you can have the organization grow. You have to be at the right life stage to be ready to do a startup now. For example, if you just had a new child, it may maybe hard to be successful in these cases because you're not at the right life stage. You have to be ready, have a little bit of savings and money in the bank, go without getting a salary for the next six months or 12 months, or maybe your family is sick or your parents are not well and you need to take care of them, it's going to be really hard to have a successful startup at that time. You have to be at the right life stage. You have to have the right level of ambition. So many times I see this, especially in Japan, where people think very narrowly, they think about a very small problem. In order for this type of venture-scale startup to work, you have to think big. How is this going to be the next Google or the next Meta? Have that level of impact. Not just thinking, oh I can potentially make a sustainable business from this. That might be fine but you're not going to do this type of tech high-growth startup by doing that. You need grit and perseverance, you won't give up easily because there are a lot of hard times when you're building a startup. If you have co-founders, you want to make sure that the co-founders have complementary skills but mutual respect. If you have co-founders that are too similar, it's kind of like the way I think about it, the founders at the beginning and the first set of employees are the gene pool of your company. You want to think about what DNA do I want to have in your gene pool? If the DNA is all the same, then you get a bunch of inbreeding and it leads to all sorts of problems. If your company is all engineers and you have nobody who knows anything about sales, you're going to be in trouble. On the flip side, if you have all salespeople and no technical people, you're going to be in trouble. This is why you want people who have complementary skill sets. Maybe they have a shared vision and ambition, but their skill sets are different. Otherwise, you run into problems where the question is, who's in charge of this? Oh actually, you're both really similar, so the question of who's in charge of something becomes a really hard question. It’s much better if your co-founders are people who are a little bit different from you. This is statistically shown and proven. There's a book called The Founder's Dilemmas and if you're interested in this, I suggest reading it. All the best teams, all the most successful teams are ones that are more diverse as well. There's also an interesting data point about the companies with mixed gender on their founding team, they do better than companies that are founded strictly by men or strictly by women. I think there are a lot of benefits to having diversity in your early team. Around first principles thinking, just more on that, this is especially true in this world now where GPT is available to everyone. All the public information is going to be indexed and incorporated into these latest foundational models. How do you build that differentiation? First principles thinking is where everything has changed and everything is new to everyone. If you can fully understand how the systems work and think from first principles, you’ll see new opportunities that other people won't see because they’re caught in their traditional thinking. If you think from first principles, you’ll come up with truly innovative solutions that most people won’t come up with, and that becomes a true differentiation. Not that many people are actually capable of doing a lot of first principles thinking. Around a company, yes we're in this new era of AI, with all these new capabilities, but the fundamentals of a startup don't change. Who is your customer? It’s amazing how many generative AI companies I see, they don't know the answer to that. They don't know who the customer is. What is the burning need problem you're trying to solve for that customer? Are they actually willing to pay for it? Not just, hey I had this free thing that I downloaded and tried out. No, they have to be willing to pay, and there are other ways to pay. They could pay attention, to advertising, but how much are they willing to pay and how much are they willing to pay? What's the size of the market? I'm solving this problem for this customer, how many of those customers are there in the world? How do you actually identify and reach those customers? What's the message that resonates with those customers? Do you know how to reach them? Do you know where to find them? When you do find them, what's the message that's going to work? Is your solution 10 times better than the alternatives? It can't be 10% better, it has to be 10 times better for your company to be successful. None of those things change. This was true before generative AI, it was true during AI, and it will be true after. If you don't have good answers to these questions, you're not going to have a successful company. I think a lot of people in the new AI world get caught up in the technology and forget about these fundamentals. These things have all become very important. I’ll talk a little bit about the benefits of being part of a community or accelerator programs or startup communities. There are a lot of benefits from that. Recently, there was a Harvard Business School report, an academic study, about startup communities. They found that founders who are members of a startup community or join accelerator programs end up with measurably better outcomes, even when they compared them directly to companies that didn’t join, versus ones that did. They typically raised more from investors, were more likely to be alive, and more likely to be acquired than similar startups that didn’t participate. They did the analysis and called out three very important things that turned out to be the deciding factors about whether those companies were successful. Number one is getting very rapid intense feedback. The founders who got that were typically much more successful than ones who received feedback over a long period of time. It’s a little bit counterintuitive, but the thinking is that rapid intense feedback helps to overcome natural tendencies to resist feedback. Everybody has their idea, their ego is tied up with it, and if they're only getting feedback slowly over time, they're unlikely to listen to it. But if you get a lot of people saying the same thing in a very short period of time, it helps to overcome those tendencies. It makes it easier to compare advice from different sources, and getting that person kind of pushes thinking more abstractly and helps you see higher level patterns. This is part of the reason why they think that getting rapid intense feedback is better than getting feedback over time. There’s also this notion of friendly rivalry and support network. When you have regular interactions like in the structure of an accelerator program, you end up learning from others who face similar challenges. It gives subtle pressure to execute because you see your friends or other people you know doing well, progressing, and now you feel like you better keep up. Being in that peer group makes it easier to identify missed opportunities or bad assumptions you're making because you can calibrate with everyone else and end up getting better information. This is another reason. The third reason they said these types of programs are helpful is because of deadlines. The fact that you have a fixed schedule and tight deadline allows the founders to have this trade-off between broad learning and execution. You have to do both. If you're just blindly executing without really thinking about things, you're losing the forest for the trees. You need to be able to do both, but having those tight deadlines allows you to emphasize execution, which is extremely important.
I’ll talk briefly about Inception Studio. We actually limit the number of companies we support because we want to keep the quality bar very high. We have a very limited set because we want to make sure that everyone who joins is high quality. We’ve had quite a few companies form already. Many of the founders are repeat founders with successful exits and a lot of experience. So we’ve had 50 companies form so far, but only 22 have raised because many of them are still in the early stages and are bootstrapping. When you've had a successful exit in the past or you've been working for 10 or 15 years, you usually have a cushion and know the benefit of not involving investors early on. So many of them have not raised money. The ones who did raise money ended up being very successful in their fundraising. That’s because we focus on quality. If you get a great reputation for quality, it has a halo effect. Once upon a time, Y Combinator was like that. It used to be very hard to get into Y Combinator, and there was a halo effect, so anyone who joined could instantly get a lot of interest. Now they kind of scale, now they have 200-400 companies every batch. They do one in summer, one in fall, and they do a lot of them. The quality has definitely dropped. Among investors, they often talk about this and say, maybe the recent batches are not as good as the ones in the past. By focusing on quality, the companies that form typically finish their fundraising in two to three weeks, their fundraising is done quickly because we focus on quality.
Audience: Two things you haven’t really addressed. Since you’re going to start your branch in Japan, there are cultural and regulatory hurdles that are potentially quite different from the US. We know that founders, especially foreign founders in Japan, maybe they can't apply the same formulas that they used when founding their startups in the US or in Europe. Approaching customers or dealing with VCs here is a little bit different. Maybe you've researched that. Then there are the regulators. In the EU, you have the AI Act that came into force and this is going to limit some enterprise endeavors as well. In Japan, it's also a little bit different. Can you speak to these aspects?
John: This is kind of a work in progress, we’re still figuring some of this out. Our goal is not to target Japanese founders for Japanese investors and a Japanese-style company. This is more about finding people who want to build the next great set of companies worldwide. People who want to have stronger connections to Silicon Valley and other places in the world and have that level of ambition. When I came about a month ago and went to IBS in Kyoto, I met a lot of startup founders, successful startup founders, people who want to start companies, investors, and people from the government and JETRO. A common thing I saw in Japan was that many Japanese investors think in a really different way than typical Silicon Valley investors. They say, even for early-stage companies, I need your full financial plan and projection of revenue, and this sort of stuff. They’re very focused on the numbers. I think part of that is because many Japanese VCs, are not as founders, they are from finance. But this is not uniform. There are quite a few investors who think a little bit differently, and the common theme is that they feel like we have plenty of capital, we have plenty of money, but there are not enough good companies to invest in Japan. This is a common thought. The thing that I've seen is sometimes the founders in Japan are thinking a little bit narrowly. They think about how can I apply this problem only to the local market, or I’m going to take this idea that was successful in another country and bring it to the Japan market. That’s very narrow thinking. The model, if you're an investor, the way Silicon Valley VCs think is like this: I’m going to give you a million dollars right now, and in five, seven, or 10 years you're going to give me $100 million back. It’s not about doubling your money, they need 100 times their money back. So tell me the story about how you're going to build a company that is going to give me 100x of my return. That’s the kind of story Silicon Valley VCs are thinking about, at least 100x. In order to do that, the only path is building a global brand, not a regional company. We’re building something that’s going to be the next Google, the next Uber, the next Airbnb, the next Meta. You have to think in a very ambitious way. I see that is not that common among founders in Japan. They often think much more narrowly about local markets. There’s a selection thing that happens. The people who are truly ambitious about that, guess what, they end up not starting their company in Japan, they go to Silicon Valley or somewhere else. The people who are left are the ones that are maybe a little bit less ambitious. I’ve also talked to a lot of incredibly talented and ambitious people who feel like there wasn’t as strong of an ecosystem. For various reasons, maybe they want to start their company in Japan, sometimes it’s because of family, sometimes it’s because of other reasons, but they want to start a big company, they are extremely ambitious. When we're thinking about Inception, we're not thinking about doing this for Japanese founders for Japanese investors and a Japanese-style company. This is more about people who want to build the next great company who happen to be in Japan now or who want to come to the Inception Japan event. But when you're doing the pitching, yes you're pitching to in-person Japanese VCs, but we also have Silicon Valley VCs there. All the pitch coaching and other things we’re doing are aligned towards building a globally successful company, one on the scale of at least a unicorn and beyond.
Audience: Hi John, thanks for the presentation. Personally, I’m ecstatic that you're here and doing this. I think I align personally with a lot of things that you described. My question is just, what brought you to do this program in Japan specifically? Naturally, for a lot of reasons, it's not, again, personal alignment notwithstanding, right? It's a place where there are a lot of roadblocks. I think a lot of people push back, saying power laws don't work here, the envisioning of the venture, etc., isn't aligned with public market evaluations and exit patterns and whatnot. It’s the same debates that I have with people all the time. Was it the money?
John: The honest reality of it is yes, there are a lot of challenges in Japan doing this type of thing, and honestly, there are other markets where it arguably makes more sense. There are language challenges here, so other places with more native English speakers would be a little bit easier. Culturally, there are challenges here as well. That being said, I used to live in Japan 25 years ago and I worked at IBM Japan, which is a big company. When I was living here, I saw many people who were extremely talented, maybe doing research or they’re engineers, and they’re stuck in these big, saturated roles. Especially back then, 20 years ago, it used to be that the very best people would go into government or big companies, and then people who were not as good would go to small companies, and it was only people who could not get a job that would go start their own company. That’s what it used to be. It made me so sad to see this. I know there are a lot of challenges, family challenges, where parents will tell you, I sacrificed so much to put you through school, why are you giving up your stable job for this crazy dream? Who's going to take care of us when we're old? These are real challenges. That being said, if we wanted to optimize for choosing the group where it’s going to be the statistically most likely to succeed, we would choose all men, all people from Stanford, and all technical founders. That’s very shortsighted. I’m a big evangelist and I have a very soft spot for people who, through no fault of their own, have talent, ambition, all the pieces, but because of societal pressures or other things, like discrimination, they end up not being able to succeed. I can't let that happen. This is why I've already had success. I’ve had two prior exits, and now I can do what I want. I want to see a world where it’s not 97% of the funded companies going to all-male teams, or where the only way to build a company is to go to Silicon Valley and be in Silicon Valley. I don't think that's good for the world. I want to see a world where you can live in Japan and start a world-class AI company. I want that world to exist. That’s part of the reason why I’m thinking about doing it in Japan. Also, my daughter is living in Nagoya right now. I’m actually going to pick her up in a couple of days and take her back to the US. I have connections to Japan and I want to encourage an ecosystem here to develop. It's going to take a long time, but we have to start somewhere. That’s part of the reason why I’m thinking about doing it in Japan.
Audience: Hi John, my name is Max. I'm from South Africa. Thank you so much for your presentation. You spoke a lot about founders who already started. I think a lot of people here have not necessarily started, but perhaps have some ideas in their minds. I wanted to ask, we’ve seen a lot of companies in the US like OpenAI, DeepMind, and other startups doing some really interesting work. What do you think are the next set of ideas that are going to be transformative, especially considering that we might go into a depression, where it's no longer cool to work on AI? What are the next set of challenges that you think AI can change the world?
John: There’s been a natural evolution. There was a big glut of companies that are basically building tools for AI. A lot of the people who are starting companies are engineers or AI engineers, they’re the ones who understand AI, and they say, what problems do I have? Oh, it's hard to use RAG, okay let's build a system and make it easy to use RAG. Okay, that's great, we need systems like that, but there are about 40 of them and only two are necessary. So some of those are getting overcrowded. Now I’ve seen this evolution where many companies started from building tools and picks and shovels for this gold rush to starting to do more vertical AI. They’re using AI to solve problems in construction, agriculture, finance, and other verticals. I’ve seen this shift start to happen. But the problem you should work on should be something that you actually genuinely care about. That’s the real question because these things take a long time. The median time to an exit for any company that doesn’t shut down is 10.7 years. You should plan on working on something you’re excited and passionate about for the next 10 years of your life. When you're starting a company, you’re trying to do it with an eye toward being successful. If it's successful, you’ll be working on it for the next 10 years. What problem are you passionate enough about to work on for the next 10 years? That will naturally take you in the direction that is going to be successful. Yes, we’re going to have that downturn, and a lot of companies are going to die, and a lot of founders are going to give up. They’ll say, I thought I was going to be able to get rich quick, and now it doesn’t seem like it, so they’ll leave. When things get tough, it becomes unsexy to keep going. Any company you start right now, I'm pretty sure you're going to have to go through that downturn. The best way to survive the downturn is if you care enough about the problem you're solving that even when it is unsexy, you’ll still continue to work on it. So the answer is going to be different for everyone in this room because everyone will have a different answer to what that problem is.
Audience: Hi John, I have a question regarding the approach to scaling a startup. You mentioned that at Inception Studio you're looking for ideas that might go global. As an entrepreneur, when I see problems around me, I think maybe AI can solve them. At the time, I was not thinking from a global perspective, I was just thinking about how I could make things better in my surroundings. It may or may not scale, but that depends on whether the PMF is there or not, and if that model can be applied to other markets. Is that thinking narrow?
John: There are tons and tons of problems to solve and tons of things that AI can build. Have you ever been to this website, “There’s an AI for that”? Literally, there are tons of problems you can solve with AI, but is that worth your time? The question is, of all the problems you could work on, which one should you work on? Because there are a lot of different things you could do, having good taste in the problems you’re working on is important. You don’t want to just build something because you can. You could theoretically come up with a list of 50 projects to work on, but is that worth your time? You want to focus on the problems that are going after a huge problem and a huge market. There are many cases where people solve a problem and completely own a market, but that market is tiny. That’s not worth your time. By tiny, I mean less than multiple billions of dollars. Yes, you can build a sustainable business, but if you're on the path of raising VC money, there has to be that story of 100x return. The only way you do that is by having something that can hit that level of scale. Understanding the size of the market and the opportunity will naturally orient you toward the problems worth solving. That being said, you should have a personal connection to the problem you're trying to solve. You have to find the problems you have a personal connection to, and if you solve them, it can actually be global and change the world.
Audience: Hi John, I’m Ryan. Thanks for sharing, your views are very insightful. I was also wondering about building something, but when trying to find the problem, it always feels like a bias from my experience or my stereotypes. How can I avoid this kind of trap?
John: I highly recommend that if you have some interest in an industry or area, spend time with the actual customers in that space. Sit down with them, live their life with them. That’s where all the insights are going to come from, and that will help you break out of your own bubble. There’s a real danger in just thinking on your own. Startups are not a thinking game, they're a doing game. It doesn't matter what you think. If you sit by yourself and theoretically think of the perfect startup, perfect idea, or perfect company, you're going to fail. That’s not the road to success. The road to success is actually talking to customers, understanding their problems, and then coming up with some innovative solutions. You don't do that in a vacuum or in a bubble, you do that by engaging and talking with customers. If you find yourself in this trap where you're doing that, it probably means you're spending too much time with yourself. You want to find people who are going to be your target ideal customers and try to understand their problems and their pain. But realize they’re not going to come up with the solutions, they don't know the solution space. You want to understand their problems and then use them to generate solutions. “I saw you struggling with this thing, you’re wasting a lot of time or spending a lot of money on this. I think I know a solution that would be 10 times better than what you're currently doing.” That’s where the insights can come from.
Glasp: Thank you for the presentation and Q&A. Until 9:30 p.m., we can network and interact with John and the audience. Tomorrow or the day after, we’ll send an email with the application for Inception Studio. John has prepared many business cards, so please make sure to thank them.
Application to Inception Studio Japan
