It it's really the the biggest wave that I've seen in my lifetime, and I think that's because of the way it has so much impact on almost every possible business and almost every possible life. The potential for AI to just dramatically change the way we do everything is enormous, and we're still at the very beginning of it. We barely even scratched the surface. Hi, everyone. Welcome back to another episode of Glasp Talk.
Today, we are so excited to welcome Chris Yeh. Chris is an entrepreneur, investor, and author with deep experience in Silicon Valley startups. He is the coauthor of the bestselling book, Widthscaling, with Lieder Hochman, sharing how companies grow fast in competitive markets. Chris is also a founding partner at Britscaling Ventures and teaches entrepreneurship at Stanford University. Today, we will talk about his journey and his views on scaling companies and the impact of AI.
Thank you for joining us today, Chris. My pleasure. It's great to be here. Thank you. So first of all, let's talk about AI because it's eating the world and its impact on businesses and startups is massive and huge. So I think you've been involved in startups since the mid 1990s and have witnessed multiple tech revolutions. And how does this current AI wave compare to earlier errors, like the rise of the internet or mobile?
What stands out to you about AI? Could you share about this? Absolutely. It's really the biggest wave that I've seen in my lifetime. And I think that's because of the way it has so much impact on almost every possible business and almost every possible life. Now the Internet absolutely was transformative as well. I think it's almost difficult for people who are young to remember that there's a period in time when we weren't all connected together, when it wasn't possible to see what was going on.
People, literally would spend their entire lives and and never actually communicate with people unless they physically saw them face to face. So the Internet was certainly big, but the reason why AI is potentially even bigger is because of the huge impact it has on the daily activities that we have. So the Internet made it possible for us to do things online. AI makes it possible for us to do things that we couldn't do before, whether it is writing books more quickly, whether it is, you know, transforming the way we manufacture things by having AI control robots or what have you.
The potential for AI to just dramatically change the way we do everything is enormous, and we're still at the very beginning of it. We barely even scratched the surface. Yeah. And I can't even imagine, living without AI now. I everyday use ChatGPT or Cloth for coding and Gemini sometimes. But do you use any AI tools for your daily work or life? So I do apps. I use I use ChatGPT and Gemini. I also play around with the tools like Perplexity and a few other things.
I'm not using as many as most developers because I don't develop software, And so I'm not using Claude code or codex or anything like that right now. Although I do try to follow the developments. But nonetheless, it is a 100%, you know, something that gets used on a daily basis. For example, just the other day, just two days ago, as a matter of fact, I was working with, my editor, and we were working on coming up with book covers for a couple of books that we're gonna be publishing.
They're gonna be self published books, but, you know, obviously, we want good covers for them. And the ability to generate those covers in in AI to work back and forth quickly on them and refine them, it's quite remarkable. In this case, we were using ChatGPT, but we were also using Canva. They have a fascinating tool. Actually, was I'm not even sure it was Canva. It might have been something else. There was a tool that my editor showed me called FlowState where it just generates images continuously.
So you have a prompt, and it just keeps generating images, you just go through until you spot one that you like. And that's the sort of thing that never could have been done before. It's remarkable. Yeah. And before this interview, you talked about you are publishing a new book next Yes. Year or in 2027. Could you share a little bit about the new book and your work? Yes. If you can share. Yeah. Absolutely. So, of course, everyone is very familiar with blitzscaling, which is a bestselling book and has been very influential since it came out in 2018.
And what Reid and I are working on this year and probably publishing next year is a sequel to Blitzscaling. That'll be called Blitzscaling in AI. It'll be focused on how AI has changed blitzscaling, how blitzscaling has enabled AI, and how people can blitzscale the AI companies of the future. Wow. That's the question we exactly wanted to ask next. Everyone wants to ask about it, which is why we're gonna write the book.
So could you share a little bit about the how AI changed the idea of blitzscaling? I I remember the the book has been. It's been seven years. Right? Over seven years since you've published over seven years. So there are so many things that are true. First of all, AI is a technological revolution, and that always produces blitzscaling because it creates new markets and new market opportunities. The second thing about AI is that it is also changing the way we blitzscale by enabling individual people to just do so much more.
I see smaller teams than ever before. This is some not just me. A lot of people have commented on this. Smaller teams can accomplish more. They can get further than ever before before they need to raise money or or hire new people. So this helps address one of the things that tends to limit blitzscaling, which is the organizational scalability of a company. There's only so fast that you can hire human beings.
It takes a while to find the right people, bring them on board, train them, get them up to speed. But in a world where you can instead accomplish things with an army of AI agents, all of a sudden, your growth can be even faster than before. And we've seen this in the AI native companies growing at these tremendous rates. I see. And and yeah. And in the bridge scaling book, so you yeah. You were writing about the concept of, like, you know, company scale, like, from so family, tribe, trade, society, nation.
Yes. In the AI era, so, yeah, how does it change? Also sorry. Yeah. Also, so there is a concept talking about, like, you know, sort of in unicorn. Do you think it's gonna be, like, ideal, also doable in the future? Excellent question. So a couple of things. First of all, as you mentioned, in blitzscaling, we talk about levels of organizational scale, and it goes by orders of magnitude from family to tribe to village to city to nation.
And what is true is as organizations grow through those sizes, they are gonna experience the same things that we saw before. Because, fundamentally, if an organization has a thousand human beings working and if there's certain ways in which they're going to work together, it's just based on the way humans seem to relate, and those things don't really change that much. However, what is different is what you're referring to, which is if each human being is supported by a team of 30 AI agents, even if those AI agents aren't as fully productive as a human being, maybe they're only one third as productive.
Each person is effectively doing the work of 10 people. And so companies can grow, and they can grow with much smaller staffs than before. So imagine if, like, before, it would have taken a 100 people. Now you can do it with 10. And that has a tremendous difference because it allows you to stay informal. It allows you to stay, smaller and nimbler longer. And I think it is to the benefit of companies to stay smaller and nimbler longer.
Now I do think that saying that three people are going to be able to run a billion dollar company or even one person, as as Sam Altman and other folks say, I think that's a bit much. I think that if they do that, it'll be by outsourcing a lot of those functions to other companies where there are human beings working. So it's kinda cheating when it comes to getting that one person, $1,000,000,000 company.
But I think it's quite possible that in the past, it would take a 100 or a thousand people to get to a billion dollar company. Maybe instead, it takes one tenth as many. Maybe it takes 10,000 people, and that's 1,000. A thousand person company, maybe it's a 100. A 100 person company, maybe it's 10. I think a team of 10 can do tremendous things. I see. Yeah, totally. Yeah. And so, yeah, we are talking about maybe can make a unicorn, but so therapy unicorn without people in the future.
That is good And to That that may very well happen at some point in time because what will occur is, you know, again, you'll be able to with a world where so many inputs and outputs are mediated by AI agents and APIs, maybe MCP servers, what have you, you'll be able to create certain kinds of businesses. I started my career in, the hedge fund industry at D. E. Shaw, which is a quantitative hedge fund.
If you're building out a quantitative trading platform, you may very well be able to do that without any human beings. Totally. Yeah. Thanks so much. And I I was curious, like, the concept of bridge scaling. So because bridge scaling ultimately relies on the assumption that in software business, marginal cost trends towards zero and the scale's advantage improves unit economics. So given that, so in today's AI era, so user growth often drives cost to increase linearly.
So do you think bridge scaling should still be applied? So I do think that what we are going to see is a classic pattern that we see throughout all sorts of software and and computerization, which is that right now, we are seeing, you know, a lot of costs to AI because everyone is focused on improving capability. And the race is to improve capability as quickly as possible regardless of cost. The fact is that the cost of serving, like, a previous generation model just plummets thanks to the continued drive towards ever more powerful chips and ever more Moore's law still seems to apply at this point, although Jensen's law is also applying.
So if people needed to find a way to make AI more cost effective, they easily could. They could use smaller models. They could find all sorts of ways to drive that cost down. The fact is nobody cares right now. That's why the costs are rising. And I think that in the the realm of blitzscaling, the point is over time, it absolutely will be the case that the cost to serve will decline, and that will give us those high margins that we look for.
I see. Yeah. It makes sense to invest in AI. Right? Absolutely. Because if you don't invest in AI, what exactly are you going to invest in? Those are that's right. So as you mentioned, I think in the future, many startups or individual solar printers are using AI agent to do business and to help their business. But in that sense, what can be the new moat? How can startup differentiate with other startup? Because if everyone use same AI agent, and Mhmm.
They could do the same thing. Right? So True. Although everyone can use employees, and they do the same thing as well. Right? We shouldn't distinguish. Every company could theoretically hire human employees. Well, they're all the same, aren't they? But they're not. So I think that what will be true is we believe very strongly that the technology moat is not going to last. And even if somebody has the best model today, maybe another model comes out two weeks from now that's better.
There's gonna be a lot of switching back and forth. And so what we think people need to do is they need to, develop other motes. So there are classic motes like the the data moat, or I prefer to call it a context moat. Right? The data moat, people really talk about data. I'm like, no. It's not really data. Data is just the means to an end. The end is context, understanding the context that allows you to drive greater value.
And if you have a context moat and you build up that context moat and you are the only one who's collecting that cutting edge data that will allow you to really understand the context, that's one of the ways you're gonna have a moat. There's also traditional ways like having a network effect where the more people who are using a platform, the more valuable that platform is. It doesn't matter whether you are serving that platform with AI agents or human employees.
At the end of the day, if that platform becomes more valuable, then people are just gonna use it. If we think about companies like, for example, an Airbnb. If you have an agent going on to Airbnb to find the Airbnb for you, but you still purchase the Airbnb, what difference does it make? In fact, you might be more willing to use Airbnb because you'll be able to find exactly what you're looking for quickly.
I see. But how do you, you know, find a context mode then? So yeah. Because AI is changing, model is updating, product is differentiated, then every two weeks new product launches. Right? So how do you keep Well, the context mode is the business context. So thinking about your business, most of the knowledge about businesses is not in any structured form right now, and it's not easily accessible. Obviously, we've gotten used to AI firms pretraining on the contents of the Internet, but all the business context has never existed there.
Right? It's existed within the walls of the enterprise and even more so within the walls of small businesses. So if you are able to go out and build a product and an organization where you can go into these places where there's all this dark data and have to transform it into useful context, then that's where the value is gonna come from. I often tell people, listen. You know, the same AI, the thing that distinguishes the different AI from each other is not gonna be which model it is.
It's gonna be how well do you understand the context and how well you can apply, because of that context, AI to real business problems to create value. And so if you understand the context of your customers, of the companies that you serve, and you will add to that context by gathering data from the interactions, from what you're doing to serve them, then you can absolutely still build them out that way.
But in that sense, I mean, regarding the data mode and and context mode, and, eventually, like, a big companies like Google, Microsoft win because they have so much data. Do they? They use Do they? So let me ask you this question. They have a lot of data. You know, Google has always had a lot of data. Microsoft has a lot now. This is one of the benefits of moving to the cloud. Right? Your customer's data is on your servers as opposed to before when Microsoft, the data was on the computers themselves, a little harder to access.
But nonetheless, do they really have access to all the data? Well, it's difficult because they do not have that you know, nobody is so standardized that there's just one platform that they use and where all the data's on that platform. Google may come the closest because of Gmail and Google Drive, But even then, you know, you have, on a daily basis, I'm gonna use email. I'm gonna use Slack. I'm gonna use text messaging.
I'm gonna use WhatsApp. I'm gonna use all these different things, which are all run by different companies. And the context only comes together in the company itself or in the individual itself, not at the level of the the those current vendors. It's also the case that people whenever there is a new technology revolution, they always say, well, the incumbent players are gonna win. And the fact is this is not true.
Right? This has happened many times in the past. IBM, once upon a time, was the technology company. In fact, the saying was that the industry was IBM and the seven dwarfs because IBM was so dominant, and there were seven other companies like Burroughs and NCR and other places like that that nobody really cares about. And did that persist? Did IBM remain the dominant player? Is IBM the leader in AI today?
No. Not at all. In fact, IBM was then replaced in people's minds with companies like Microsoft. And by the time I was active in the late nineteen mid to late nineteen nineties, people were like, oh my god. Microsoft will just dominate. Microsoft will just do everything. Right? Why would you bother building a web browser? Why would you bother building a search engine? Everyone's gonna use Microsoft. That wasn't true either.
And then we get to another era and people are like, oh, well, everyone's just gonna use Google forever or Facebook forever. The fact is that as new technologies come along, it's difficult for incumbents to fully dominate and embrace them. They've done a pretty good job. I mean, again, I you give a lot of credit to Microsoft for building this relationship with OpenAI, give a lot of credit to Google for being able to really do a lot of pioneering work in AI, and now Gemini has become an extremely strong player.
But believing that that means that they're just gonna dominate everything forevermore is just incorrect. That hasn't happened in the past, and it won't happen in the future. I see. And in that sense, as an investor, I think your job is, like, to find the next unicorn or next Google. Right? So how when you talk to founders, when you see startups, what aspect do you see? Oh, and then figure out, oh, this company or startup will be the next Google.
Is there any yeah. Anything Well, so we're always guessing. Right? Remember, every venture capitalist, they make investments. They never make an investment saying, well, this money is wasted, but I told them I'd give them money. So here you go. Right? They always invest because they believe the company is gonna be successful. But the fact is 90% of them are not. So it's very difficult to know. Now I will say what we look for in companies, we wanna see that there's a winner take most market.
We wanna see there's a reason why there's gonna be one big player in five to ten years because we wanna own that big player. And we look for those network effects that are gonna drive one big player being the winner. We look for the virality that's gonna allow them to outgrow the competition, maybe some great distribution or things like that. We look for the ability to charge premium prices and earn high margins in a big market, and we look for the ability to actually scale up the company to serve the customers.
Because if you fail to do that, your business will fail as well. So we absolutely look for a bunch of things that we hope will help us find the best companies. But at the end of the day, even after we've invested, there's no guarantees. The vast majority of startups will always fail, and the default outcome for any startup is that it's gonna die. And the question is, can the founder find a way to pull it off?
And that's why the thing that will never go out of style is, can you invest in great people? Is the founding team brilliant, hardworking? Do they have the capability to build something amazing? Do they have special insights into the world that other people don't have? Do you or your team at BridgeSkating Ventures use AI, in processing or or investment? Like, a diss deciding up the company to invest in. Yes.
We use AI to help us with the research. We use AI to help us compose investment memos. We use AI to find competitors. We use AI for all of these different things. We still haven't, like, built out our own, like, commercial platform to do this. We're using other people's tools, but we are exploring that. I mean, one of the things I would like to do this year is to spend more time building internal tooling and infrastructure.
In fact, this is one of the things that is different about AI because AI makes it so easy to build internal tools. You should probably be building more internal tools than you did in the past. We're used to waiting for a commercial software product to come out so we can use it, and that's because it was so hard to build something internally and so hard to maintain it. But in an era of AI, you may very well wanna not just you know, I I'm not a big believer in vibe coding your product, but you may very well wanna vibe code your tooling and use vibe coding to help you drive much better tooling and much better productivity.
Yeah. That reminds me of your recent blog about too many companies confuse buying AI with adopting AI. And Yes. Could could you could you elaborate on this with the audience who don't Absolutely. Get them grow up yet. Is something that, you know, we're we've seen a lot over the past couple of years. There's been a tremendous pressure put on companies from the board and and from investors to adopt AI. And so what they've done is they've gone out and they found somebody like Microsoft or Google or some provider, and they bought the AI.
And that's great. They got to say, we did something about it. But that's not the same as adopting the AI. In fact, what happens is the people who are dealing with the board by the AI, they often don't even bother checking with the actual workers who are gonna use the AI, and the AI sits on a shelf. It's never used. That's a big issue because in the end, nobody's gonna pay for a product that isn't used. They're going to eventually churn out, and that's gonna be a big problem for those companies.
There is a saying these days that people are in pilot purgatory. So they've been able to sell pilot projects and build up their revenue that way, but then the pilots don't turn into lasting expansions. And that is the big thing that you have to do as an entrepreneur as well. You cannot mistake that pilot purgatory, that bunch of trials for traction that is going to be lasting. Maybe it'll be lasting, but you have to actually see that pilot turn into a long standing, long term contract.
Do you have any tips or advice to startup founders who wants to turn their product into lasting product in the enterprise? Yes. Is it the context context building context? So there's the context element of it, but the other element is being obsessed with the user behavior patterns and how the users are using the product. It is very easy to say we made this sale. We've got this revenue. We've got this ARR.
I'm like, yeah, ARR. That's great. That's fine. But that doesn't tell me that you're gonna retain that. Right? The net revenue retention is more important, and you need to be able to make decisions more quickly than waiting a year and seeing if they churn or not. So you need to be able to monitor people's usage of the product and see whether they are continuing to use the product, in which case they're likely to stay, or whether they've stopped using the product, in which case they're likely to go.
So the number one advice I have for founders is stay close to what your customers are doing. Make sure that you carefully instrumented your product and are tracking the actual level of engagement that occurs because that's gonna be the best proxy for what will eventually matter, which is net revenue retention. And I have a, like, a interesting story. And when I talk with the founder and so one of my friends are selling b two b product to company, but he is afraid of how to say in the future, the enterprise realize, oh, we can build this tool by ourselves, so we don't need to buy this product.
As you mentioned, you know, AI, the company should explore and leverage AI and build something for themselves. Right? So in that sense, the more company don't buy, like, the AI tools, or how does the market shift? Well, here's the here's the key. Remember that as people are building things, they're vibe coding them up and what have you, maintenance is a challenge. So vibe coding and all those things are great for tooling, things that are temporary, things that are not persistent over time, things that don't have to be production ready.
If something requires production ready software, you still need real software to do it. Now what I think will occur is that the individual business functions or tasks are going to fall into different categories. There's going to be ones where there is value to standardization, where everyone in the industry does it a certain way. And if that's the case, people are gonna gravitate towards commercial software, and people are gonna buy that commercial software and have common uses for that software.
But it's the idiosyncratic individual internal tooling where it's never gonna be viable for, a external software vendor to actually produce that. And that's one of the things that AI is enabling. Right? These are all these things that they can never hope to actually get a product to buy. Now they can build it themselves. But if there is a product available to buy, they're probably still better off buying that product.
I see. Eventually, AI can make, like, a standard standardized, you know, a platform or product, so in in the field. So then Yes. One company can do it. Right? Yeah. Eventually. And and, again, it becomes a case of just like everything else. Right? Eventually, with the right human programmers, you could create a product that would be commercial and and work in the field. And so it's really a question of your relative speed.
Are you able to do that faster than your competitors or not? That's it. You mentioned that so in yeah. So to make a lasting product, so you need to talk to customers closely and, you know, check details. So AI product or AI UX is changing, you know, too. Right? So what what is the common pattern in AI? Also, what kind of, like, element or what is the ideal use case or UI for AI? Do you see? Yeah. Yes. So I think what's interesting is, you know, we've seen AI take off first with just a chat interface, and that's largely because it's simple.
Everyone knows what it is, but it's far from clear to me that it's the right way to do things. I think that there is a a couple there are a number of interesting things happening on the UI side. I think there's a lot of UI that will take advantage of things like voice and maybe even, you know, wearables and vision and things like that because there's so much context that lies in how we say things or what we're looking at that right now with any computer product, you you can't capture that.
Right? You're typing it in, and it's just whatever context you provide. So I think there's a lot of interesting things that will happen on the input side. On the output side, again, I don't think a chat window interface is the best way to provide people with output. Output has it's not like, you know, Salesforce CRM is a chat window interface. It's not like Quicken and QuickBooks for accounting is a chat window interface.
So increasingly, we're gonna see more traditional interfaces that convey information in the right way. But it may very well be that the intake is still gonna be like the voice or the the glasses or something like that, knowing more and more of the context of what you're asking for. Got it. Yeah. And many people are talking about, like, VR or AL. So, yeah, do you see the future? So yeah. I I just curious, like, your thoughts on, like, VR or, like, Google Glass or, like, Apple Vision four.
Yeah. So I think that AR and VR, especially AR, are inevitable at some point in time, but I don't know when. So the fact is that the products aren't good enough. The VR products aren't good enough. They make people feel sick. The AR products aren't good enough. Things like Apple Vision Pro are too heavy, and nobody's gonna wear them for an extended period of time. Maybe the next generation of the Meta Ray Ban glasses will be good enough.
I mean, the first generation was useless. It didn't even project anything. It was just basically a convenient way to listen to podcasts, and you could have just, you know, bought earbuds for, like, one tenth the cost. And now we're you you know, we have cameras built in, and it's improving, but it's still not even close to being what people need. But I do think that they are going to be successful in the long run because our entire world like, it's gonna be so much more valuable if I have AR that's telling me information about the world, letting me see it visually as I go through the world, or VR for the sake of entertainment.
If I wanna watch a program, you know, why should I have to have a 100 inch screen in order to watch it immersively? I should just be able to wear something and have it be super comfortable and just do it. The thing that has held these things back has been the hardware has been insufficient. But with continued innovation, with new technologies, I'm sure we're gonna get there eventually. And I think as an investor, you have met and talked to so many founders and startup founders and so on.
And have you found any interesting ideas or companies using AI in a unique way? I don't know. They could be in the stealth mode, so I don't know if you can share. But if you Yeah. That's that's a little tricky. I mean, one, I can share things that are are publicly known. So one of the one of the companies that we've invested in does something which is interesting, which is to use AI to try to bring people together in real life.
So the company is tackling this issue of loneliness and people being unable to form social connections. And it's like, well, gosh, isn't AI gonna make it worse as people spend all their time with their AI companions and don't even interact with other human beings? Well, that's one way to look at it, but the other way to look at it is for AI to do the work of figuring out who you should actually connect with and then design experiences where you can connect with that person and begin building an invert in an in person, in real life relationship.
Because if you think about it, relationships and I'm not talking about relation romantic relationships. I'm talking about friendship and getting to know people. It is ultimately a liquidity problem where it's a matching issue where it's difficult to match with the right people. Well, that seems like something AI can do a lot to help with. Yeah. Definitely. Yes. Thank you. So let's talk about AI and learning because you and Neil Hochman often talk about the importance of being an infinite learner.
Right? In a world where AI and technology are evolving so rapidly, how can founders, professionals, and students cultivate the infinite learning mindset? Do you have any habits or Absolutely. So the couple of the couple of habits I think are important. The first is you should be taking in new information on a daily basis. Right? We all have to be learning constantly. You're not gonna learn unless you're taking in new information.
And so that means collecting information from a variety of sources, hopefully reliable ones, and processing at least a certain amount of it per day. And, you know, I, for example, am taking in information all the time from news sources, from social media, from but from more unusual places as well, podcasts about, podcasts about, literary novels or Tolkien or what have you. The whole point is you wanna continue bringing in all this fresh information for your mind to process.
It's just like, giving training data to an LLM. You need that data. So you need to be taking in that data from a variety of sources and getting new and novel information. The next thing that I think you need to do is you need to, as I like to put it, pay attention to weak signals. So what you need to do is not just wait for someone to tell you, hey. Here's the way it is. What you need to do is look for all the things where as you go through the world, you think to yourself, that's funny or that was unexpected because that is where the process of saying the model I have for how the world works might be outdated because I'm seeing things that violate that model.
And being open to those violations and open to exploring those violations is one of the ways that you update that model instead of just ignoring them saying, oh, it doesn't fit with the patterns I know, so therefore, it's not important. It's an exception. I'm like, yeah. I don't know. Maybe that's gonna be more and more the rule in the future. The last thing as far as being an infinite learner is we tell this tell people this all the time.
You have to be willing to let go the lessons of your past success. Oftentimes, people are successful for a reason. They can articulate that reason. They can say, here are the things that I did that made me successful. This is my secret sauce. And I'm like, that was your secret sauce for that particular accomplishment. But the world has changed since then. And whatever made you successful five years ago, ten years ago, twenty years ago, will it make you successful today?
I don't know for sure. You need to be willing to, even though it's something you're very proud of, learn and decide that maybe the lessons of the past no longer apply. That doesn't mean they were the wrong lessons. They were the right lessons during the time that they helped you succeed. But the fact that they helped you succeed twenty years ago doesn't mean you have to be loyal to them now. Yeah. Absolutely.
Yes. And so you mentioned that you see a lot of news and you check the social media, but do you keep the ideas somewhere? Like, I I know you are writing blogs, and it's a long way to keep your memory and ideas. But do you have any other ways to keep your ideas and ways so that you can come back and look at look back at any time? Yes. So there are a couple of things. I actually use a browser extension called Dokkio, d o k k I o, and what it does is it allows me to bookmark it records all the web pages I go to, but it allows me in particular to bookmark web pages and tag them, and then it relates those web pages that I look at to the files and documents that I have in my Google Drive or, you know, in my Slack or what have you.
So it helps tie together all these different streams, the web browser, the tools that I'm using for communicating, the files I'm working on, and it ties them all together in context for me. So that's one thing I do that's that's somewhat unusual. The other thing I do is yeah. Again, this is really primitive, but I actually will if if thoughts strike me, I have a what I call a scratch pad within Google Doc where I just throw in the new thoughts that are interesting to me.
And I may not go back to them immediately, but it's really important to me to record that thought, to put it there both so that I will stop thinking about it, but also so that in the background, I can potentially see how it connects to the other things that I do. And, also, you mentioned about the weekly signals. But with AI generating millions of blog posts and articles, it feels like Internet is getting noisy and noisy there, but do you think human It is.
Creation will become more valuable than creation? So I think people already clearly value more and more in real life interaction, meeting people in person. Right? Post pandemic, of course, for a while, there was a surge in peep and during the pandemic, people said, oh, you know, we're gonna do Zoom. It's so convenient. But after the pandemic, because people missed seeing each other face to face, there was a surge in in person events, more and more people gathering and getting together.
But that has now outlasted that initial sort of overhang from the pandemic. People want to get together. I see more events happening, more meetups, more conferences than before the pandemic. And it's partially in reaction to the fact that, hey. People want something real, which is that face to face interaction. Do you I think some people say, you know, San Francisco is the center of AI. Do you do you think so?
I think it is. Do you a It's not the only place where AI is occurring. Obviously, you see, you know, tremendous efforts and a lot of smart people in all sorts of parts of the world. China is an obvious one, but Japan as well. I had a chance to meet the the founders of Sakana AI at one point in time, and there's quite a cluster of AI companies in Japan. It's happening in Korea. It's happening in The Middle East.
So Silicon Valley is not alone, but it is the most prominent. It is the place where people come. And after all, you guys are here. Thank you. Yeah. It was great to meet you in Japan. So and Kazuki met you in San Francisco last time. And so I know the founder of Sakana AI. So, yeah, if you need intro, let me know. And so there are many events happening in your world. But so how do you choose which event you go?
Because ideally, I want to go to many events as much as But I at the same time, I need to develop. Yes. No. And again, the filter that I apply is a question of so if I'm just going to attend an event, not speak at an event, it'll usually be because it is an attempt to learn about something, a new topic that I really wanna get experience with. Because I don't believe that I can learn about some new technology or new thing just by reading about it or reading stories about it, even worse.
I wanna experience it face to face. I want to talk to people who are using it. So there's that. But then the other filter I use is and, again, it's my world, right, where I'm more of a public intellectual. It's am I going to go speak? Am I going to be able to go out there and tell stories and make an impression on a bunch of people at once? Because if I go to an event, I might talk to five people. But if I go speak at the event, all 500 people will see me and hopefully carry away an impression and remember to reach out to me in the future when they're like, oh, something I'm doing really fits with what he's doing.
I see. I think you can meet anybody in the world, but so if you can choose Not quite. You can No. No. No. There are there are plenty of people I haven't been able to meet yet. But so if you can meet anybody in the world, so who will be? Oh, that is a difficult question. Somebody who is act in other words, somebody who's actually alive today that I could meet that I haven't met yet. Mhmm. Or maybe, you know, that yeah.
Yeah. So, I mean Yeah. There are there are all sorts of fascinating historical figures I would love to meet. So for example, Abraham Lincoln, who is the greatest American president, is one person who springs to mind. Benjamin Franklin, who is perhaps the most famous of the founding fathers who never was president, but, you know, was one of the smartest and most well known people in the world. There's so many figures throughout history, fascinating, for example, with Alexander the Great, who, you know, lived this insanely intense life where he conquered the Persian Empire and and the the rest of the known world by the time he was 32.
So there's so many people that would be interesting. Here in in in in real life, I'm trying to think. I I don't know. I've had a chance to meet most of the people I wanna meet in real life, which is nice. Yeah. Nice. Okay. Yeah. This is just from out of my curiosity. But Of course. It seems that from my perspective, you are very successful. And but what what's your next goal or dream you're you're aiming for at this moment, like in five years, ten years?
Excellent question. So, you know, I think that, you know, I've had I've had the kind of success that matters to me. So there are other forms of success that I don't have. I'm not a billionaire. People sometimes say, are you billionaire? Like, no. No. No. I work with billionaires. I am definitely not a billionaire. Nowhere even close. You know, it's always possibly richer and all those things. But to me, what is interesting is can I continue to meet really smart and interesting people who then go on to make the world better?
To me, the most interesting thing is meeting the entrepreneurs or meeting the thinkers that are moving the world forward. So I wanna continue doing more of that over the next five to ten years. I think that beyond that, you know, I think that there are still other things I wanna do. I wanna continue writing more books. Maybe at some point in time, I'll be able to go and do a course of study on something that like, there's areas I don't know enough about, like biology and what have you.
But really the most important thing to me is just meeting smart and interesting people. Yes. And I think a lot of younger people in the younger generation are afraid of AI taking jobs, I think. Yeah. If you were starting career today, how would you learn AI? What would you do? Though there's so there are several things. The first is I would very enthusiastically adopt AI because the people who adopt AI and get the most out of it are going to outcompete the ones who don't, the laggards.
I think the other thing I would do is I would look towards, learning the kinds of skills that AI does not necessarily provide. So I feel very lucky that when I was studying at Stanford University, I didn't just study my academic studies. I also studied public speaking and taught public speaking. I also studied improvisational comedy and and did improv comedy for many years. All these things contribute to to my being able to really connect with people and learn and listen and and accomplish things face to face, which as we've discussed, has just become more and more important.
The other thing, of course, is that AI is very good at dealing with things where there's patterns already, but has more difficulty creating things that are truly novel and new, which is why it's like, great. Continue to take these things in and develop your ability to create, to push in directions, to be curious about things that don't make sense, to really go and continue to be the driving force behind learning more, doing more, making the world better.
Totally. Yes. And so our audience are, like, founders or writers, creators. Do you have any advice to them, like, feel like, who feel pressure to use AI about but Yes. Don't know where to I do. So and, again, this is, an important sort of nuance around how you use AI. I think that, of course, everyone should use AI, but you shouldn't necessarily just sort of say, I need to use AI to do the most important thing.
Right? The most important thing may remain your act of creation, having that insight, having that novel thing. Like, when I write, I still write myself. I may use AI to help me, research it, but I'm still gonna write on my own. And oftentimes what I'll do is I'll write something. And then just to see, I'll ask AI to write it for me, not with my not with what I've written, but just, you know, on its own.
And, you know, sometimes I learn and I see things that I'm like, wow. I wish I'd thought of that. But, usually, there's a bunch of things where I'm like, well, I'm glad I did this and AI did not. It will be interesting to see what happens when in the future we are able to have personalized AI that really understands all of our own context. Because that's the advantage we all have right now when it comes to AI.
Our personal context is so rich and so, so massive that it's impossible when it comes to something personalized to for AI to compete. You can't go to ChatGPT and say, ChatGPT, you know, rank my favorite classmates for the third grade because ChatGPT doesn't know that. That's super personal. And so I think that that is something that you need to do. You need to think about how AI is able to help you with the things that are impersonal or generic, but really still dive deep into doing the work yourself for all those things that make you unique.
And before asking the last question, and since you have been working with for over, I don't know, ten years or more, I guess? Almost fifteen. It's almost fifteen It's years been a while. So what stand out to you about LEAD and what what did you learn from LEAD? Or how's the experience working with LEAD? So it's an incredible experience, and there are an assortment of things. So many things I've learned. I'll try to distill them down to the most important lessons and the most important things I've learned.
The first and what the thing that Reed is most famous for is sort of saying, you know, life is a team sport. And so it's all about the network and calling upon the resources that you know, not just trying to do everything yourself. And so Reed really does learn so much from the people around him. He's always looking, always has curiosity. He wants to learn from anyone. It doesn't matter. I mean, obviously, if Reid had to wait to learn from people who knew more than him, it would be difficult.
He knows so much. But he's always looking who from to from whoever he's sitting across from to learn from them. And that, I think, is a a really important lesson. I think that one of the things I've really benefited from is Reid thinks big. Right? He thinks ambitiously. He thinks about how can I make things bigger? How can I make things have a greater impact? I think in this world, many of us, maybe for rational reasons, think cautiously.
Try to think about, well, what's the safest thing I could do? Reid never thinks that. Reid thinks, what's the most impactful thing I could do? What's the biggest thing I could do? So that is another huge thing. And then finally, there is a clarity to Reed's thought that is based on following a couple of principles. The one of the principles I always remember is if you're thinking about doing something, the primary reason for doing it, the primary motivation has to be enough to justify it.
You cannot get there with a bunch of secondary motivations. And so if you're like, should I do this? Should I not? And you say, well, there's this and this and this and this, and therefore, I should do it. No. That number one reason needs to be enough. Now, of course, you wanna add those other reasons as well, but having that criteria really helps you think about the decisions you need to make much more clearly.
From from, like, working with him for Mhmm. Fifteen years, have you seen any changes in him, like a thought process, how he deal with problems or how he treat people, or was he the same all the time? So pretty similar pretty similar the whole time. I would say that, if I were to come up with any differences at all, it would probably be along the lines of just the different tools that he uses now. Right?
He continues to seek out, new ways to to learn. And, of course, part of that, it comes with success and things like that. Like now, for example, he has this wonderful guy, Parth Patel, who's sort of his technical adviser on using AI, somebody who uses AI all the time and shows Reid the the cutting edge applications. That's an example of how you can leverage a person, and Reid has sort of brought people like that into his life to to help him.
That wasn't as much the case fifteen years ago. Yeah. Actually, we welcomed, Passpartout in the past interview. He's incredible, isn't he? He is. Did you use ReadAI for sure? Did you use it? So I was so we have we've used ReadAI a couple of times, though, primarily, for events where Read couldn't go. And so as a chance for people to hear from Read, they always enjoy that. But, you know, I think that there will be more in the future.
Right? I I have a friend, for example, who's an entrepreneur. He uses AI and has his own AI board of advisers, and he has Reid, and he has me and Steve Jobs and Clay Christensen and all these other people. Now, again, I don't know how accurate AI is at simulating all of us, but he actually gets a lot out of talking to all of them. I I hope he I find it very flattering that he was willing to include me in that board of directors.
I'm like, you know, I I don't think I belong on that list, but I'm glad to be included. Yeah. Yeah. And I I hope they don't conflict with each other in the board meeting. Well, sometimes they do disagree, but that's okay. Right? It's good to have a board of directors that sometimes disagrees and has different inputs because, your goal is not just to do whatever they say as a consensus. Your goal is to take all the different advice and make the best decision you can.
Yes. Yep. And we are happy to welcome, you know, Read AI, so to this Glasp Talk. Yeah. Fantastic. In the future. Yeah. In the future. We'll find way to make that happen. Yeah. Yeah. Yeah. And so from your perspective, so what is, you know, Reader's primary reason to, you know, keep doing what he's doing? Oh, it's very simple. Reed believes that we all have an obligation or responsibility to help make humanity better.
Is that the best of patience? He wants he wants the world he wants humanity as a whole to do better. And so AI is one of those tools, and it's helping humanity do better by achieving its dreams. It's also helping humanity do better by, you know, for example, maybe finding cures for diseases or allowing us to live longer, healthier lives. And so that's his primary motivation. And of course, it just happens to be that if you create that kind of value in the world, well, fortunately enough, you're usually able to make some money along the way.
Beautiful. Yeah. Yeah. So this is a time is running up. So this is the last question. Since Glasp is a platform where people share what they are reading, learning as their digital legacy. So we want to ask this question. So what legacy or impact do you want to leave behind for future generations? Well, is an interesting one because not too long ago, just last year, had my, no. You're you've been a year and a half ago, I had my fiftieth birthday.
And for my fiftieth birthday, I actually used a product called tribute, tribute.co, where you can go and collect video tributes from people saying nice things about you. And I wanted to do that because usually you have to wait until you're dead for everyone to say those nice things or maybe if you win an award or something like that. But, you know, why not do it for your birthday? And what I was really pleased with in terms of the legacy was the number of people who talked about, how I was how I showed them kindness and how I tried to help them.
And I think that, you know, obviously, it's really cool to have books and things like that and create a, create a word that hopefully will persist for many decades and people will be using, out into the future. But I also want people to remember and say, you know what? He was somebody who really was kind, and he was somebody who tried to help me and and made my life better. And I love hearing those things, most of all.
Thank you. And definitely you are already. Yeah. Thank you. But thank you so much for taking time, and we learned a lot from you today. It was my pleasure. I'm so glad I could take part in this and now I've to run back and check out your Parth episode because I can never get enough of hearing Parth speak.