Tim O'Reilly - Founder of O'Reilly Media

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to Tim O'Reilly, the founder of O'Reilly Media, and we discuss his essays on the 21 Technologies for the 21st Century, why it's not a zero-sum game when bringing value to the customer, and how AI adoption is helping to redefine the enterprise. All of this right here, right now on the Modern CTO podcast. Here we go. This is the Modern CTO podcast.

(Joel Beasley at 00:00:35) Tim, hey, how are you, my friend?

(Tim O'Reilly at 00:00:39) I'm good.

(Joel Beasley at 00:00:40) I'm excited. We're recording. Is that cool? We're just gonna hang out and talk and have a good time.

(Tim O'Reilly at 00:00:44) It's fine. That's fine.

(Joel Beasley at 00:00:46) We have so much content to cover today.

(Tim O'Reilly at 00:00:48) We do.

(Joel Beasley at 00:00:49) Your content's amazing, man. I was reading your latest two articles. You're a fantastic writer.

(Tim O'Reilly at 00:00:56) Oh, well, thank you. Appreciate that.

(Joel Beasley at 00:00:58) When did you get started? When did you start writing?

(Tim O'Reilly at 00:01:01) Really, right out of college. I had a couple of projects that are too hard to explain. But then the first thing I actually got asked to write by a friend was a book about Frank Herbert, science fiction writer, author of Dune. And then about the same time, I had a friend who was a programmer who got asked to write a manual, and he didn't know how to write. I didn't know anything about computers.

(Tim O'Reilly at 00:01:25) We sort of—I said, "Look, I'll just back you up. I'll help you with the writing part." And we would go in and talk to the engineers about the project, and I would basically, you know, say nothing. And then afterwards, I'd debrief him. You know, what were they really talking about here? And it actually, in a lot of ways, was the best training. I did some of my best work back then because it was all pattern recognition. I was originally a classicist, and it was a lot like parsing a Greek or Latin manuscript.

(Tim O'Reilly at 00:01:55) You know, you kind of figure out what the words are. You figure out what order they go in. You figure out what's the structure of the conversation. You know, I would often find myself going, "Oh, wait, these two different engineers are really talking about the same thing using different words." And I would sort of be able to restructure the documentation totally based on this idea of—you know, it really goes back to general semantics in a lot of ways.

(Tim O'Reilly at 00:02:22) Korzybski talked about structure, order, and relation, and I would just be looking for the structure of things. And that kind of set up a map-making sensibility for me, I think, with respect to tech, where I would just go, "Oh, these things go together. Oh, wait, there's a dissonance in the pattern here." And so it was just a lot of pattern recognition in the way I kind of got into tech. And, again, obviously, I did eventually learn quite a bit and started doing programming and so on, but I didn't start there.

(Joel Beasley at 00:02:53) So do you have a writing habit? Is this something that you practice on a schedule?

(Tim O'Reilly at 00:02:59) No. Once you do a lot of writing, it's kinda like—yeah, it's kind of like saying, do you have a speaking practice schedule? I do. Well, yeah.

(Tim O'Reilly at 00:03:11) I mean, I guess if you're doing that professionally. But, you know, after I got through that first book on Frank Herbert and a number of computer manuals, you know, I could just sit down and write. That doesn't mean that sometimes I don't have to really cogitate on something, and reworking and rewriting is always kind of a key part of the practice. But just getting the words out is no harder than having a conversation like this.

(Joel Beasley at 00:03:40) I agree. When I did my first book, I didn't realize how much work is in the editing process.

(Tim O'Reilly at 00:03:47) Yeah.

(Joel Beasley at 00:03:47) I think I did between six and eight full draft revisions. Went through it with two different editors. And it was amazing because you're on time four of editing this, and you notice there's a misspelled word or just words out of order because your eyes adjust to it.

(Tim O'Reilly at 00:04:05) But more than that, in my early days when I was writing manuals and didn't really completely understand what I was doing, I felt like I had to work—go rewrite it almost three times. And actually, it was really less of the writing. I guess it was—you know, because often technical writing, at least in the old days, it was really taking an engineering spec and turning it into a manual. It wasn't writing from scratch. I mean, yes, you interviewed the engineers and so on, but there was a body of writing.

(Tim O'Reilly at 00:04:35) And it was just—it was changing it for a different purpose. And, again, I'm old enough that it was controversial in those early days when we wanted to put things in the first person or second person, you know, like "you should do this," because it was always like "the system does XYZ." And the battles we had over "this thing doesn't work." And they'd go, "Yeah, yeah, we know that." And I'd go, "Well, what do you think your user's gonna do? They're gonna beat their heads against the wall and think they're doing something wrong. You've gotta tell them that it doesn't work." And they were like, "No, you can't say that the software doesn't."

(Tim O'Reilly at 00:05:11) And it was such a relief when we started publishing our own books because we could have this conversational, over-the-shoulder—yeah, "this feature is pretty sucky, you don't use it," and so on. Anyway, I digress.

(Joel Beasley at 00:05:30) How did the—it's funny because I did all this research. I read as much content of yours as I could because I wanted to really get an understanding for who you are and how to have the best conversation. And right as I'm getting ready prepping for this interview, I don't know, five minutes before it started, this question just popped in my head. What's with the covers, the sketches? How did that start with the—there's birds, there's different animals. How did that come about?

(Tim O'Reilly at 00:05:57) Yeah. That goes all the way back to 1987, believe it or not. We had done our first seven books, and they were all really little pamphlets. You have to understand, we were a tech writing consulting company. I had this idea that we wanted to do product and, you know, turn some of the manuals—basically, we were a tech writing consulting company. I recognized open standards were happening. People were asking us for the same manual, so I started writing my contract so I'd keep the rights and then be able to modify the manual for someone else.

(Tim O'Reilly at 00:06:37) And then we had a big slump in the tech industry. I guess it was '84. And we decided to try turning some of these manuals that we'd kept the rights to and were selling to—licensing to companies—into standalone books. And we were kind of focused at that point on UNIX, and there was a really good book on UNIX. And I said, "Wow, we'd have to really beef up our manual to compete."

(Tim O'Reilly at 00:07:01) And I had this idea, which actually came from another one of our consulting jobs where we were doing a manual for a financial services firm. It's actually for SWIFT, believe it or not. Society for Worldwide Interbank Funds Transfer software to run SWIFT. And the traders wanted to have a manual they could slip under their keyboard. That was just small. And, you know, this is the days when computer manuals were big binders, often with dot matrix printouts. And so we said, "Oh, let's make the UNIX manual smaller."

(Tim O'Reilly at 00:07:47) And we had written this manual for a CAD/CAM company, and they wanted basically the basics of UNIX and the basics of vi. And we split it into two books, and we had these two little pamphlets. One was called Learning the UNIX Operating System. The other was called Learning the vi Editor. And they had the same cover, just a different title. We had the daughter of one of our employees who was an artist, and she drew these nuts, and we called them Nutshell Handbooks. And later, we did the quick reference books, the In a Nutshell quick reference books, that actually used that nut as a logo for many years.

(Tim O'Reilly at 00:08:12) And then we had seven of these things, and people were complaining. They didn't realize we had seven books. And the guy who was in charge of the program went out and hired a designer who did something, you know, expected. It was geometric and high-tech, and I just said, "This is just not us." And one of our writers went home and talked to her housemate who was a designer at Digital Equipment Corporation, and she said, "Oh, these UNIX program names sound like strange animals."

(Tim O'Reilly at 00:08:41) And she actually literally kind of did seven covers for us. Somebody we did not know. It was really one of the—you know, if you talk about open source as a gift culture, the animals were a gift from somebody who was a stranger. And we looked at it and went, "Those are really weird," because they were some of the really weird covers too. You know, sed and awk with the big lumpish creatures, the one that went on vi with the tarsier with the big eyes and big fingers. And we just said, "Yeah, let's go for it." And, of course, it stuck.

(Joel Beasley at 00:09:07) I love it.

(Tim O'Reilly at 00:09:07) But anyway, that's ancient history. We should move on to more.

(Joel Beasley at 00:09:10) That is actually—that's the most important thing we'll cover today. Anyway, so there's two articles or the essays. So it's a collection of essays, four essays, 21 Technologies for the 21st Century. Are these published yet? Because I got a preview version of them, I believe.

(Tim O'Reilly at 00:09:26) Yeah. You know, they're actually follow-on to an article I wrote called Welcome to the 21st Century, which is out on the oreilly.com site as a public site. These four essays are actually subscriber-only on oreilly.com. And I've written two of them, two still to come. So basically, O'Reilly is now—really the heart of our business is our online learning platform.

(Tim O'Reilly at 00:09:51) And, you know, it's tens of thousands of books, tens of thousands of hours of video. You can look for answers. It's really kind of—there's some traditional learning. In fact, live online training is one of the biggest features of the platform. We also do events that are kind of like "meet the expert," kind of hour-long podcast-ish kind of videos.

(Tim O'Reilly at 00:10:11) I just did one yesterday on the Welcome to the 21st Century, which is really an introduction of the idea of scenario planning. You know, everyone kind of says, "What will happen?" You go, you don't know what will happen, but you know, you can imagine a lot of things that might happen, and then you can ask yourself what are the robust strategies that you can use to adapt to them. And so I kind of took that framework and I kind of said, "Look, here's the way to think about the post-COVID future." And then our VP of marketing said, "Well, could you do some more follow-on that's just about the specific technologies and we can use that on the platform?"

(Tim O'Reilly at 00:10:44) So, no, they're not available outside. Although we do have, you know, millions of active users on our platform, so it has readership there. And what we're really trying to do in those essays, the four essays—and the first one's called Table Stakes—and it's really kind of like, what is the stuff that isn't gonna go away, but that you just have to get good at. And that's everything from communication to math.

(Tim O'Reilly at 00:11:15) And in a lot of ways, these pieces are kind of pointers to where do you learn these skills. And a lot of the point of them are pointers to content in our platform, which is what they're really for. They're kind of guiding people to use it more effectively. But I'm a big believer in pointing to the best resources wherever they are. And so, you know, I go, "Hey, you wanna learn linear algebra? MIT professor Gilbert Strang has these great lectures from the 1990s. They're on MIT OpenCourseWare. That's better than anything we have, so go there."

(Tim O'Reilly at 00:11:57) I've always kinda done that in my career. I remember early in my publishing career, I astonished my competitors by saying, "Well, we're trying to get everybody excited about UNIX and the X Window System. Let's do a bibliography together of the best books." And they were like, "What? You're gonna help promote our books?" I said, "Yeah, because I basically feel like if we do this bibliography and we distribute it to bookstores, they'll give more space to the category and we'll get our share of the category."

(Tim O'Reilly at 00:12:14) And they were like, "What are you doing?" But yeah, sure. But I've always sort of tried to have that approach to business where you're trying to create value for the customer, and it's not a zero-sum game.

(Joel Beasley at 00:12:29) No. I fully agree, and I love your mentality. And I love it, and I actually appreciate it even more having been a couple years into this education business because I have that mentality of, you know, let's get together and do something great and push the industry forward. And not everybody is like that, and everybody's at different stages as well.

(Joel Beasley at 00:12:48) But it's rare. It's rare when I see people that are like that. My sales team was like, "Hey, we're gonna have them on. They're direct competitors." I'm like, they're not a competitor. They're the 800-pound gorilla. They have the best content in the world, and we would point to them for so many things. So I've just—I'm a huge fan of the content and what you guys do and how you've shaped the industry in the last forty, fifty years. So, thank you.

(Tim O'Reilly at 00:13:17) No. You're welcome.

(Joel Beasley at 00:13:18) It's a lot of work.

(Tim O'Reilly at 00:13:19) Yeah. We're trying to be useful. And I think today, you know, our platform is used by many thousands of enterprises. You know, the biggest part of the business is actually enterprise. And that actually is an interesting point because it actually shifts somewhat. And we're gonna watch the shift of technologies within the enterprise. You know, we're at this funny place where our hearts are in emerging technologies, and we've been an advocate for emerging technologies. You know, we first wrote about the World Wide Web when there were 200 websites, and we're like, "This is the best thing since sliced bread." We created the first web portal. You know, we did the first ad-supported website.

(Tim O'Reilly at 00:13:57) You know, same thing with open source. I can be in the meeting where the term was voted on by open source leaders and free software leaders, and they said, "Yeah, we're gonna use this new term." I did all the PR for it. You know, so I kind of been an advocate for emerging technologies, but our core base for the heart of our business is the enterprise. And so we're really at this sort of coal face of enterprises trying to understand, "Oh, wait. Okay. This is the way we used to do things. This is how we are gonna have to do things."

(Tim O'Reilly at 00:14:27) And you see this—some of that shows through. And if you read the second of the two articles, which is called The New Normal, it's really just about things that are kind of old hat to a lot of us. You know, agile software development. And you go, but guess what? There's big lagging industries that are still doing waterfall. And I'm very—my nose is right up to one of them, government, because I've been—you know, my wife started a nonprofit called Code for America and now another one called United States Digital Response, which is helping government respond to COVID.

(Tim O'Reilly at 00:14:57) And it's crazy-ass stuff that they're still doing. You know? So yeah. And there's a lot of laggard industries that are suddenly waking up and going, "Wow, you know, cloud has been around, but we're really just trying to understand."

(Tim O'Reilly at 00:15:11) What does it really mean when everybody has to do cloud? You know, what does cloud actually mean? Does it mean just using AWS or Azure or Google Cloud and doing everything the old way? No. It means you actually have to adopt new workflows.

(Tim O'Reilly at 00:15:27) You have to adopt new practices for software development. And so I'm kind of talking a little bit about that. Same thing with AI. How far along are we with big data and machine learning? You know, these are things that are penetrating our society in new ways.

(Tim O'Reilly at 00:15:45) And so I'm trying to kind of talk to that enterprise audience and say, okay, here's how you have to think about it. Because there's this great quote that I've always loved from a guy named Edward Schlossberg, who is a museum designer by career. But he said the skill of writing is to create a context in which other people can think. And when I think about something like my advocacy around open source software was around, hey, look.

(Tim O'Reilly at 00:16:15) You guys have kind of bought into this free software narrative that it's about Linux, and Linux is this threat to the established order. And yeah, that's real. But you realize that you're all using free software. Did you know, like, if you—because this was at that point, we're five years into the internet boom, dot-com. And I said, you all have email.

(Tim O'Reilly at 00:16:36) Right? You all have a website. You realize that the web was put into the public domain. You realize that the DNS, which lets you be newyorktimes.com instead of—I don't know what their IP address is anymore. I used to be able to rattle it off, you know, for that narrative.

(Tim O'Reilly at 00:16:53) You probably have an Apache web server and that. And I kind of had these guys up on stage. I had like ten guys to go, you know, they all built shit that you're all using. Stop pretending that it's this—you know, that was shaping the way that people thought. And you see that throughout these articles.

(Tim O'Reilly at 00:17:12) I'm really trying to tell people you gotta think differently. You know, like, the good example is, what does it really mean? How do you think in a sort of cloud and algorithmic world? You have to realize that a lot of these software development disciplines are fundamentally now management disciplines, because the programmers are actually doing management work. You know, the workers—you go to Amazon, who takes your order?

(Tim O'Reilly at 00:17:38) It's a bot, you know, effectively. I mean, what do we think of as a standalone bot? It's a program. And the job of the programmers at a place like Amazon or Google or Facebook is they go in every day and they kind of go—just like a manager. You know, they're kind of walking around, looking at their workers and going, oh yeah, this guy could shape up a little bit, you know.

(Tim O'Reilly at 00:17:56) Oh, you know, we just got some new data from customers that they're abandoning the site when this worker tries to serve them. Yeah, we gotta get this worker in shape up. You know, maybe we're gonna fire them and replace them.

(Tim O'Reilly at 00:18:14) You know, like, think about Google. They fired their entire robotic workforce for doing translation and replaced it with a new AI-based workforce. You know? And they did that based on data from customers. So this is all part of the management of a company now.

(Tim O'Reilly at 00:18:31) It's not just—and of course, that says a lot for, like, what the role of the CTO and CIO are in a company. You know, it's not just, oh, you're over here in a corner. And I talk a little bit about that in terms of security. I had this great conversation with the CIO of Goldman Sachs, and he said—you know, I'm not gonna mention names—but he kind of said that two years ago, the CEO didn't have any idea who this guy is, their chief information security officer. Today, he's his best friend.

(Tim O'Reilly at 00:19:01) Yeah. You know? Because all of a sudden, people go, holy crap, this is absolutely mission critical to the business. And that's a lot of what the piece, "The New Normal," is about.

(Tim O'Reilly at 00:19:10) You know, what are the things that we're kind of like, oh yeah, they're over there? And now, no, they're relevant for every enterprise. And if you don't understand them, you're at a competitive disadvantage. And then the third installment, which I'm just beginning working on, is really about emerging technology areas that are, you know, potentially the next Silicon Valley.

(Tim O'Reilly at 00:19:39) And I do think that we've kind of gone down a bit of a rat hole with just consumer and mobile and, you know, we're gonna do one more advertising-based startup. And yeah, we've got a bunch of stuff that's sort of touching the physical world on logistics and—you know, I talked some about some of that in "The New Normal." But I think there are some emerging areas, you know, agriculture. You know, like, the fact that—and I talk a little bit also in this piece, "Welcome to the 21st Century," on the scenario planning stuff. How do you, you know, watch trends?

(Tim O'Reilly at 00:20:14) And you get some little random data point, and you go, wow, that's interesting. And you start watching. One of the ones that just blew my socks off when I learned it a few years ago was that the Netherlands is the world's second largest agricultural exporter. How do you have a tiny country be such a big player in agriculture?

(Tim O'Reilly at 00:20:31) And you go, oh, you go do a little homework and you go, holy shit, they are so far ahead of the rest of the world in managed agriculture. You know, verticals fall. Export? What?

(Tim O'Reilly at 00:20:42) What did they export? You know, I didn't go that deep. Well, no, I mean, but you know, a lot of it is vegetables, but they're growing them in massive greenhouses. They're doing vertical farming.

(Tim O'Reilly at 00:20:54) It's all software controlled. Really? And then you kind of go, oh, and a couple years later, there's all these startups now, you know, coming out of Silicon Valley in ag. You know? Yeah.

(Tim O'Reilly at 00:21:05) And so that's one example. Another: climate. I do—you know, I'm pretty serious. I'm seriously concerned about climate, and I think, you know, responding to the climate crisis is, in fact, the greatest opportunity of the 21st century. You know, Elon Musk is just the first and most obvious climate billionaire. You know what I mean?

(Tim O'Reilly at 00:21:26) You know, Tesla was a climate play. Solar, space. There's a bunch of climate billionaires in China, you know, around solar panels and the like. And yeah, it really kind of—it's, again, this narrative mismatch. The narrative is, oh, climate is—you know, changing, responding to the climate crisis is a cost, and we can't afford it.

(Tim O'Reilly at 00:21:46) I got a—it's a freaking opportunity. It's a huge opportunity. And just like, you know, open source was a huge opportunity and the internet was a huge opportunity, climate is going to be the most amazing economic driver if we can—if we can basically get the politicians out of the way and have them stop propping up old industries and start investing in new ones. Another one that I think—we've kind of seen with coronavirus, you know, genetics and its role in health care, drug, you know, AI and drug discovery—all these kinds of things are, you know, we're on the cusp of some pretty serious new industries that will suck up the skills of the kind of people who used to be going to the consumer internet.

(Joel Beasley at 00:22:33) So you have a lot of ideas around the future in AI. It seems like you spent a lot of time thinking about this. Have you thought about, like, the universal basic income concepts or the freedom dividend stuff? Do you have thoughts on that?

(Tim O'Reilly at 00:22:47) I do. I think that there's clearly a lot of value, and there are some natural experiments. Think about Alaska. You know, when I talked to David Autor, famous labor economist at MIT, you know, he basically said, yeah, oh yeah, you want two natural experiments: Norway, you know, what they do with their oil dividend, and Saudi Arabia. And in Saudi Arabia, they basically take all the oil revenue and they basically use this to support an idle, you know, aristocratic class, and everybody else is screwed. And in Norway, they distribute very in a very egalitarian way, and it's led to a pretty good society. You know? So I do think that there's a role for it.

(Tim O'Reilly at 00:23:33) I do think, though, if I had to make a choice between UBI and, say, you know, investing in, you know, like, when you think about what's the national purpose—and I do think that in the US, we have this idea of the pie. Despite our rhetoric about opportunity, our basic action is the pie is limited. We have to do all this remedial stuff. And why do we have to do remedial stuff?

(Tim O'Reilly at 00:24:05) Because we basically design the economy to reward elites. You know, it's systemic. Just like we have systemic racism, we have systemic inequality baked into the system. And, you know, so like, if we got rid of the systemic inequality, we wouldn't need UBI, you know, because it's kind of a patch to a broken system. Whereas what we really have to do is we have to, you know, focus on why are we rewarding, you know, bigness and homogeneity?

(Tim O'Reilly at 00:24:37) Why are we totally ignoring the supplier side of the economy? We're just basically like, as long as some big company can give people stuff cheaper, it's fine if they put all these people out of work. It's fine if they drive down wages because they have so much market power. You know, these are things you go, no. If you actually are looking at how would we balance this economy, you would say, oh, let's start out by making sure that we have an economy that pays people well doing things that we want them to do.

(Tim O'Reilly at 00:25:09) And again, you look at something like climate in that context and you go, oh yeah, we're gonna have so much work to go around. We just don't have enough money to go around because we've basically been, you know, funneling it to elites who use it in this Ponzi market of stock buybacks. You know? And it's like trillions of dollars just going up in smoke, effectively, while we ignore investment in the operating economy.

(Tim O'Reilly at 00:25:36) That's a new term. Oren Cass used that recently, and I love it because I always say the real economy and it sort of nagged at me. And he said, no, the operating economy. You know, we have a financial economy, and, you know, people kind of conflate the stock market with, um, you know, the operating economy. And I actually was writing a piece—I've written a little bit about this in a couple of pieces I wrote about Google and Amazon earlier, last year and the year before. About—I wrote a critique of the blitzscaling idea from Silicon Valley. But I also was working on a piece right before the COVID hit. I was all ready to publish it, and I thought nobody's gonna give a damn right now. It's called "Why Is Jeff Bezos the Richest Man in the World?" And the answer is surprising.

(Tim O'Reilly at 00:26:23) You know, people just don't really quite understand that there are two economies that overlap, and those two economies are the operating economy and the financial economy, which is a betting economy. And the reason why Jeff Bezos is the richest man in the world is because people are betting on him. You know, if you look at, say, Larry and Sergey, if they had kept Google a private company—I've done the math based on the public filings. I haven't updated it for the last year or so. But if you do the math based on public finance, if Larry and Sergey had kept Google private and never become a public company and had just kept their share of the profits, they'd be richer than they are today.

(Tim O'Reilly at 00:27:04) So Google's a win in the operating economy. Amazon—if Jeff Bezos had just kept his share of the profits, about a year ago, he would have been worth about $4 billion, not $160 billion. He wouldn't be the world's richest man. He's the world's richest man because people give Amazon hundreds of dollars in stock market value for every dollar profit they get. You know, whereas Google and Apple and companies like that get about $30.

(Tim O'Reilly at 00:27:33) And even there, you kind of go, even 30, that's like—we're basically, you know, do we really think that the Google of 30 years from now is gonna be 30 times the size that it is today? Probably not. You know? It's like this betting machine on constantly improving profits, which means that when companies get to a certain size, they have to start screwing their workers.

(Tim O'Reilly at 00:27:57) They have to start screwing their customers. You know, Google—you know, what I was writing about last year was about the way that Google has increasingly—used to be this enabler for this rich ecosystem of businesses on the web, and category after category, they go, actually, better for our users if we just offer—you know, never mind sending them to TripAdvisor. We'll just have Google Travel. You know?

(Tim O'Reilly at 00:28:22) And that would be fine if Google said, okay, we're gonna introduce a travel site, and it's gonna have to go through all our same algorithms and compete with TripAdvisor and Expedia. But no, they just pin it at the top and go, you know, we are the winner. You know, that's something I've been writing about. I'm starting to work with the economist Mariana Mazzucato on a project to document what I'm calling algorithmic rents.

(Tim O'Reilly at 00:28:47) You control the algorithm. You control the screen real estate. You're one of these companies. You can extract profits just by deciding, you know, you're the first result. Anyway, so there's a bunch of stuff like that that I'm kind of digging in on.

(Tim O'Reilly at 00:29:02) And, you know, when you think about, is that how true is that in the app? So I'm trying to get more and more data on that so I can make that case with more rigor.

(Joel Beasley at 00:29:08) I've been getting more involved, upping my education and my skill level in, like, investing and how the market works. Best resources I found there, Ray Dalio.

(Tim O'Reilly at 00:29:23) He's got—yeah. Ray is great.

(Joel Beasley at 00:29:25) Yeah. He's got some amazing videos that when you understand how futures work and how the stock, you know, the financial market, as you're referring to it—or the financial economy—when you understand how that works, it's almost, like, odd. Like, it feels weird to understand how much money there is and how it works. But I wanna bring that back to talk a little bit about AlphaTech Ventures and—yeah, what are you—given your knowledge and your passion for the tech industry and in media and writing, what are you guys investing in?

(Tim O'Reilly at 00:30:00) Well, let me explain a little bit what AlphaTech Ventures is. It's been around, uh, 12, 13, 14 years. Originally, we were one of the first seed firms, you know, kind of somewhere between angels and—it used to be you just kind of have angels and then you do an A round. And we were kind of like, no, we're sort of halfway in between.

(Tim O'Reilly at 00:30:21) And we were investing in things that were sort of out of the O'Reilly network of people that we knew. You know, a good example would be Planet, which is the microsatellite company. The company was formed at Foo Camp. Yeah. When the founders met some investors and, you know, we were one of the first investors. Or Fastly, which is a public company now, you know, came out of our—Arthur Bergman, who was the founder, was one of the people who kind of came to me and said, oh, you've got to do a conference for this, for our tribe.

(Tim O'Reilly at 00:30:56) It was all the people who are doing, you know, operations and making the stuff work under the covers. And so we were investing in people out of our network. In, starting about—if I don't know if it's four years ago—my partner, Bryce Roberts, had this idea for something he was calling NDVC. And it was really a reaction to the kind of unicorn fever in Silicon Valley. And I was writing stuff to critique it, you know, just like, you know, increasingly entrepreneurs are more like actors in, you know, in Hollywood. You know, they get hired for productions by VCs.

(Tim O'Reilly at 00:31:33) You know, the goal is not really a company. It's an exit. And you could see it. You know, you go talk to people at a cocktail party and you say, how's your company doing? And they say, oh, we just raised our Series B round, you know, at this valuation with this much money.

(Tim O'Reilly at 00:31:46) I go, that's not your business. You know, at least I don't think it is. The fact is, for many people, it is their business, and people weren't focused again on the operational economy. And so with Indie VC, Bryce recognized that there was this underserved market, which is companies that are focused on revenue and profits instead of just growth and a financial market exit.

(Tim O'Reilly at 00:32:14) And, you know, obviously, O'Reilly was that kind of company. I started with $500 and we just grew over a long period of time. We were funded by our customers. And so he built a process where he had company supply. And we've ended up, you know, in the Indie VC portfolio, you know, getting companies from all over the country, women and people of color as founders because, hey.

(Tim O'Reilly at 00:32:40) You know, when you're looking at companies in the operating economy as opposed to in the Silicon Valley betting economy, the people are like, hey. You know, I have this market I'm serving and there's people who love what I do, and they're giving me money. And I just need some money to grow the business further. And it's kind of funny because it's been a great investment thesis, but you really see how the system is sort of rigged against it. You know, in a typical VC firm, you actually, you know, every time you get a new financing, you get to write up your portfolio and you get to tell your LPs.

(Tim O'Reilly at 00:33:18) Yeah. Yeah. Your LPs that, you know, look, your investment's more valuable now. And we have some companies where you go, we gave them $500,000, and now they have, you know, $5 million worth of revenue, and they're profitable. And they don't need any more money, so there's no write-up to be had.

(Tim O'Reilly at 00:33:38) You know? But it's clearly a great business, and, you know, it's kind of interesting. Many of them will go on to be operating businesses. We believe some of them will end up getting acquired because they're great businesses. And, you know, there's obviously there's lots of stories from out there in the industry.

(Tim O'Reilly at 00:33:57) And one of my favorites is the guys who did RXBAR. You know? The, um... Yeah. Yeah. It was two guys, and they're both children of immigrants who were all in the food services business.

(Tim O'Reilly at 00:34:11) And they'd gone off to, you know, business school, and they were talking in the kitchen of one of the two founders. And they were talking about pitching, you know, to raise money. And the dad said, you know, shut up and go out and sell some bars. And so they went out and they took his advice and they basically, you know, they pooled a little bit of money they had and they made some bars and they figured out they could sell through the CrossFit network. And they ended up raising no money.

(Tim O'Reilly at 00:34:48) And they ended up selling for $500 or $700 million and they got all of it because, like, holy shit. You know, the customers funded them. Yeah. You know, I mean, other than a little bit of family seed funding. And that's the story of Indie VC is to build these businesses that are actually in the operating economy, that their outputs... You know, like when I started O'Reilly, I thought, you know, I'm doing stuff that customers will pay me for.

(Tim O'Reilly at 00:35:14) And most of our businesses have been in that category. You know, we're selling books or we're selling events or we're selling subscriptions to this online service. And the funding is real operating revenue and profit. You know? And we've never had to take money from outside.

(Tim O'Reilly at 00:35:32) I mean, we've had a number of equity exits because we've spun stuff out and sold it. We sold GNN. We did some early work in collaborative filtering. We bought a company that was doing collaborative filtering kiosks and we spun that out into a company and sold that. You know?

(Tim O'Reilly at 00:35:48) So a bunch of things that we did over the years where we had sort of equity infusions that came from things like that. You know, and then we started doing VC. You know, our first big win was Blogger, which Evan Williams. You know, he was actually a former employee, and we funded it. And we stood by him right through the, you know, where it was down to him sleeping under a desk, keeping the servers going.

(Tim O'Reilly at 00:36:16) And then, as the market came back after the nuclear winter, we actually helped him negotiate and sell to Google.

(Joel Beasley at 00:36:25) So entrepreneurship, in one of your articles, you linked... and it really caught me off guard because it was counterintuitive about entrepreneurship actually declining. And you had a reference, and I went in and read the reference, and I was actually really surprised about that. Can you talk a little bit about that?

(Tim O'Reilly at 00:36:42) Well, there's actually a fascinating article that I... I've just read the summary. I have not read the study. Some people describe it as Veblen entrepreneurship. You know, Thorstein Veblen is the nineteenth century economist, or maybe it was even early twentieth century, but the idea of conspicuous consumption, you know, where people are basically doing things for display.

(Tim O'Reilly at 00:37:11) And it's like, you know, there's a way that a lot of our VC has become about... is a kind of status thing about it rather than, you know, investing in real needs. And I think that the reason why... well, there's two reasons. Probably the biggest reason why entrepreneurship is declining, and I'm quite certain of it, is the focus in antitrust on consumer benefit. Because the bottom line of our current thinking is that if the consumer wins, it's all good. And it's very hard, particularly in the area of digital goods, which often have no price to say, hey.

(Tim O'Reilly at 00:37:53) Like, hey. We're getting all this free stuff. It's great. What's there to complain about? And what they're missing is the fact that almost all markets have two sides.

(Tim O'Reilly at 00:38:04) You know? And so when, you know, whether it's Walmart or Amazon who's squeezing the suppliers, all the people who are on that supply side make less money. They have less money to pay people. They have less money to invest in R&D. They have less money to come up with new products.

(Tim O'Reilly at 00:38:23) And I really see this in the web, you know, where Google... you know, Google's original mission was find the best content on the web and send people to it. And now more than half of all searches end on Google. Google just gives you the answer. And they do that either with their own products, which they've developed internally, or they do it by making a single decision. We're gonna partner... you know, we're gonna do a five-year contract for, you know, music lyrics with so-and-so.

(Tim O'Reilly at 00:38:56) You know? And so, effectively, they're building a centrally planned economy with Google as the central planner, and it only looks like it's still a free market. So here was this revolution in free market economics, you know, literally where, if you think about Google search, it was this brilliant innovation, you know, in which the coordinating function is this sort of massive number of sources of data. And, you know, in traditional economic theory, price and price signaling is sort of the primary coordinator of the invisible hand. And here in Google search, price was not one of the factors.

(Tim O'Reilly at 00:39:37) You didn't get to bid your way to the top of Google results. I mean, yes, price kind of tried to worm its way in in the form of SEO, you know, and price certainly played a role in the parallel market that they did for advertising. The genius of the original design was search is really kind of a pure play that's not affected by people paying for anything. And then we're running the same search through another search engine, an auction engine, which will produce ads and will produce the money. And so they had this incredible invisible hand matching up the suppliers' information with the consumers' information that was a beautiful advance.

(Tim O'Reilly at 00:40:16) And then they basically said, well, actually, you know, we can... we can act... and particularly as screen real estate was reduced on mobile, they said, well, we actually have to get better at just giving people the answer. And what it does is it basically takes away the opportunity from the supplier side. And so when, you know, when Google says we're gonna give you Google travel results and we're not gonna send you anymore to TripAdvisor or we're not gonna send you anymore to Yelp or we're not gonna send you anymore to this website. You know, they're basically starving businesses of traffic. And I think that happens across the economy.

(Tim O'Reilly at 00:40:55) You know, when Amazon's... you know, and again, it was bad enough, you know, when your supermarket or Walmart says, oh, yeah. We're gonna do, you know, our house brand product and compete. But that was sort of pretty granular. You know? It's like, oh, yeah.

(Tim O'Reilly at 00:41:10) It's pretty clear that batteries or toilet paper is a commodity. But when Amazon goes, wow, somebody came up with a really great new idea. We can see it in the Amazon marketplace, and we're just gonna go make that product. That's sort of rifle shot at a particular small company. Yeah.

(Tim O'Reilly at 00:41:29) That's just, you know, it's really what Tim Wu calls the curse of bigness. Actually, I think that goes back to Louis Brandeis. You know, we've optimized our economy for size, and now we're reaping, you know, the whirlwind, which is, yeah. And there's a lot of people who don't have income. They've been encouraged to borrow.

(Tim O'Reilly at 00:41:50) You know? I said, well, you know, we've encouraged companies to press down wages in order to drive up profits. I write quite a bit in my book, WTF, which came out a couple of years ago, and in some of the other things, you know, I just wrote a piece for Rockefeller on AI, you know, sort of AI safety. And my... you know, the piece for Rockefeller was called "We've Already Let the Genie Out of the Bottle."

(Tim O'Reilly at 00:42:13) You know, if the idea is, you know, the worry of people like Elon Musk and Nick Bostrom is, you know, we're gonna inadvertently give the AI the wrong instruction. You know, the classic Nick Bostrom was the paperclip maximizer. You know, the AI is supposed to create as many paperclips as possible, and it eventually realizes that humans are in the way and becomes hostile to humanity. I go, we've already built one of those. It's called our financial markets.

(Tim O'Reilly at 00:42:42) The explicit instruction is, you know, maximize corporate profits. People are a cost to be eliminated. You know? I mean, you couldn't have a more clear illustration of that, and I kind of feel like these algorithmic systems are proto-AIs. You know, Charlie Stross, the science fiction writer, calls corporations slow AIs, and it's a great way to think about it.

(Tim O'Reilly at 00:43:06) Anyway. But, you know, and, you know, so I guess I feel like all these lessons from tech... and one of the great opportunities for us is to take those lessons and apply them to the broader economy. So my kind of cutting-edge learning is about lessons from tech for the economy. I like to say, you know, tech economies are kind of like the fruit flies of economic research, you know, because, you know, they grow up and they go through their entire life cycle very quickly. Yeah.

(Tim O'Reilly at 00:43:39) But, you know, you can kind of see that boom and bust and you can kind of go, oh, you know, yeah, you know, the PC software was this sort of booming entrepreneurial economy. And then as Microsoft got bigger and they took over all the niches and they put everybody out of business, you know, everybody had to go find a new market. They did. They went and found the web, and that was good. And you could say, oh, yeah.

(Tim O'Reilly at 00:43:58) That's... But I think, you know, we've sort of optimized for the wrong things. And, again, my part of my advocacy is around the idea that markets in general are not a natural phenomenon. They're actually engineered by the rules. And, you know, tax policy and tax incentives and so on are just as much an algorithmic intervention as whatever Facebook does to its newsfeed.

(Joel Beasley at 00:44:26) You've got some good deep thoughts, my friend. Yeah. I, um, I want to share with you one of my deep thoughts about learning. Okay. So, did you look into, like, a little bit about Leaderbits or the podcast? Are you familiar with that?

(Tim O'Reilly at 00:44:42) I have not.

(Joel Beasley at 00:44:42) Oh, it's fine. 100%. So I'll give you, like, the thirty-second backstory. Yeah. So I'm an engineer for seventeen years. Kept seeing people get farther because they had relationships. I was fairly introverted. And so I started writing, turned a blog into a book. And then before I published the book, I started interviewing different CTOs about their experience. That created the podcast.

(Joel Beasley at 00:45:10) Then that grew and got very popular. And then we ended up getting people calling us up saying, hey. I heard some advice the CTO gave on your podcast, and I want to get my 100 leaders in 10 different countries to actually, like, implement that advice and take action on it. And so we said, okay. We built, like, an MVP. I got our first customer for, like, $50,000 a year, you know, sent it out to their people.

(Joel Beasley at 00:45:32) They became the customer, and they thought it was really cool because they watched this, like, five-minute video. It was a leadership challenge, and they went and took action and actually had an experience with their team so they remember that experience. Right? And that was really profound for them. This...

(Tim O'Reilly at 00:45:47) That's awesome.

(Joel Beasley at 00:45:47) Yeah. This concept of, you know, let's not learn 20 great things. Let's, like, focus on one great thing, a small story, two or three minutes, and then actual steps that you can use to go take action on it. So then we got into that business, ended up raising some venture capital and growing. And then, so the... I guess the takeaway from us was I was reading your content about the two pyramids.

(Joel Beasley at 00:46:12) And I thought... and that showed me that you spend a lot of time thinking about learning. And I've been... I'm new. Like, you've done this forty years. I'm new. I'm two years into this.

(Joel Beasley at 00:46:24) But I came up with that... the way you articulated that is like something I knew to be true based off experience, but I didn't have words for it. You know what I mean? Yeah. And so I was very excited to see that come out and then have those words for it because that really is important, this concept of people who are newer need this highly structured concept. And people who are high performance, they almost need to go reference it as they need it, to go pick it out.

(Joel Beasley at 00:46:52) And we were coming to that conclusion after making these, like, leadership challenges, but then some people needed, like, programs that were differently structured and other people wanted, you know, to go in and, like, grab a challenge on, like, team communication. And so coming to those conclusions separately, I always think is interesting because I tend to seek truth, and I find that when two separate entities come to the same conclusions, there's some good truth.

(Tim O'Reilly at 00:47:14) I'm totally with you. Yeah. I think that fundamental idea... and it's sort of obvious. You know, you think about, you know, say, learning to read. You know, you need structured work until you can read and then you can read anything. You know? And so our thesis in our platform, where we offer both structured training, you know, like, hey, you want to learn to program in Python. You can go to a live training. You know, you want to learn to use TensorFlow. Here's a live training or here's a structured book. But then when you want an answer, bang.

(Tim O'Reilly at 00:47:46) You know, it's like you want to just go look it up. You know, you can have a question. And obviously there are companies like Stack Overflow that fill that role. But we're actually doing a bunch of pretty powerful engineering on our product to actually make it a much better answers platform for that second use case, because that's really—we look at our logs of what people do on the platform.

(Tim O'Reilly at 00:48:10) And, you know, there's a set of people who do linear learning, and there's a set of people who just go on looking for answers. And I think most of the online learning world is focused on the structured part of it, you know, all the way up to—look at someone like Coursera. You know, we're going to give you a mini degree, or a micro degree.

(Joel Beasley at 00:48:30) So in reference to moving industries forward, right, regardless of competition or brand or anything like that, the reason why I was sharing that concept of the challenge with you—just littlest amount of information to get them to go take action—it works really well in management. You also see companies like DataCamp that have the interfaces where you do the challenges. But we just focus in leadership team management. That's our little area. But the reason why I share this with you is—I'll send you a couple samples.

(Joel Beasley at 00:48:58) But I think if you guys made some challenge content, you'd be blown away by how people respond to that. It's actually very, very fascinating, the different type of experience when you challenge someone to actually go perform an action with their team and then come back and journal about it really quick and you give them those steps. It creates such a profound difference. And, you know, we're smaller. We're trying to grow.

(Joel Beasley at 00:49:22) We have two lines of business, and right now our podcast line of business is doing better than our leadership line of business. And so we are focusing more on that right now. So I was thinking, you know, because we're focusing more on that right now, why not share this innovation with you? Maybe you guys can amplify this concept.

(Tim O'Reilly at 00:49:43) It's really worth trying. Obviously, you look at—there are definitely sort of challenge businesses that are more technical. So like HackerRank, you know, you're going, "Okay, we're going to go do bug bounties." That's a kind of challenge. Yeah.

(Tim O'Reilly at 00:49:58) Holberton School, you know, the way they teach actually is they basically don't do courses. They give you projects and a bunch of resources, which I love. You know, they're kind of like—even in programming, it's like, okay, your job is to do Hello World. Write a program that does Hello World, and you go, "What the hell is that?" And, you know, they go, "Well, here's a bunch of resources about programming. Go study it and come back." You know, it's like, as opposed to somebody holding your hand the whole way. I think that's pretty interesting. And Kaggle, you know?

(Joel Beasley at 00:50:35) I haven't heard of that.

(Tim O'Reilly at 00:50:37) Oh, it's owned by Google now, but that's basically data science competitions. And, you know, they really were challenges where companies literally would say—or you think about the Netflix algorithm challenge, which was early in that.

(Joel Beasley at 00:50:49) Oh, Zillow had one too. It's actually interesting.

(Tim O'Reilly at 00:50:53) Exactly. So it is an interesting idea. I really will take a look and try to think about how that fits.

(Joel Beasley at 00:51:01) Yeah. Did you see the Facebook one too, where they tried to do the spot the deepfakes? Yeah, that's going to—the deepfake stuff is something that I think about quite a bit because it's kind of like, I don't know how you solve that problem.

(Joel Beasley at 00:51:14) I actually saw some—there were some, I think, government programs out there trying to get people to come up with some of these identity deepfake problems. But do you have any thoughts on that? And then we can wrap up.

(Tim O'Reilly at 00:51:27) Well, I think probably the first thought I would have is that for some class of problems, there's sort of a PageRank kind of approach, you know, which is how authoritative is the source. That being said, you know, all of the traditional media that used to be authoritative is so jumping on the page view bandwagon that they're not—journalism doesn't work the way it used to. I still remember kind of getting schooled in journalism early in the days of blogging where I was talking with John Markoff, famous New York Times reporter on tech, and he gave me a tidbit of news. And I said, "Well, God, this is from John Markoff. How authoritative is that?" And I said, "Can I blog this?" And he said, "I'd prefer if you got a second source. You know, it's like, don't trust me. Go find somebody else who tells you the same thing."

(Tim O'Reilly at 00:52:13) And here's some people you could talk to, you know? And I was like, "Whoa, that's how journalism used to work." Yeah. And now it's like, "Man, this is a hot story. Jump on it." Yeah. So that won't work in quite the way it used to. But I still find that there's a lot of—I would bet that there's some work that you can do, you know, like, if you think about, say, here's somebody saying something. You go, "Well, we have a corpus of everything else they've ever said. This is clearly not consistent with that."

(Joel Beasley at 00:53:00) That's interesting.

(Tim O'Reilly at 00:53:01) Yeah. Raise the red flag. Or, you know, there's a lot of things you can do with—if you have a lot of somebody's words, there's small word choice differences. Yeah. I still remember a beautiful piece I heard once on the radio about Agatha Christie. Somebody had done analysis of all of her novels.

(Tim O'Reilly at 00:53:21) And there's a point in her career when she was quite old, when her language became much more specific. And it was—she ended up having Alzheimer's, and she kept writing. And it was clear that she knew something was going on because the novel, or the point at which this happens in the analysis—it was her ninetieth novel—was about a detective who's losing his memory. Oh. Yeah.

(Tim O'Reilly at 00:53:47) And it's just, like, so wild. Yeah. Yeah. And the courage to go, "I'm losing it and I'm going to keep going," you know? But the point is, you kind of go, "Okay, we had 90 books to see how she wrote and something is different."

(Tim O'Reilly at 00:54:03) And so I kind of feel like there are some techniques, you know, and they—people in literature have kind of applied some of these. "Yeah, you know, we're kind of looking at this play, and we think this might not be Shakespeare because..." Or paintings, you know, same kind of thing. You know, it's like there's a lot of fakes there, and you go, "No." It's not easy, and it's particularly not easy in real time.

(Joel Beasley at 00:54:25) And then even if you have the real stuff, maybe it's not faked as far as the deepfake—you put on a mask with someone else—but you have the out of context stuff too that happens.

(Tim O'Reilly at 00:54:36) Yeah. Exactly. And, you know, the fact that you can create a person saying things that they didn't actually say, I bet that there are some checks on that that people who have enough data could do if they really want to. You know, a Google or Facebook could probably go, "No, actually, you know, that doesn't use the words that Nancy Pelosi actually uses," or "The timing is slightly off."

(Tim O'Reilly at 00:55:02) Or there's probably all kinds of small bits of signal. Yeah. You look at just how the internet has figured a lot of this stuff out. There's lots of small signal that was not obvious at all in what Google did originally. And I think, again, that doesn't mean it'll get solved in time for the fall election.

(Tim O'Reilly at 00:55:24) You know, it's like any other exploit. You know, we have to get faster at responding. And until we get—we also, there's probably some social principles that could be applied, like, you know, if it's a critical issue, be more aggressive. Yeah.

(Tim O'Reilly at 00:55:44) I still remember—this is a very trivial example—I used to build a, in the early days of web scraping, we built this tool for looking at the computer book market by scraping Amazon. And we were looking at all the categories and, you know, what was our relative standing against our competitors. And we were using a tree map, which is a kind of visualization of categories. And I still remember that my metric was, you know, okay, we had a—the smallest level, like, it was sort of a five or six level deep. I said, "Look, if there's five titles with the same keywords in the title, we'll call it a category. You know?"

(Tim O'Reilly at 00:56:22) But if there's one bestseller, we'll call it a category. Because I know that if there's one bestseller, there will be more soon because all the other publishers are going to swarm on it. You know? So I had this heuristic, which turned out to be quite true. You know, as soon as there's one book that kind of blew out in some new thing, it would become a category, and I could call it before it actually happened. And so in a similar way, you could kind of just say, "Okay, we know this is about the election that's coming up. You know, we're going to basically put a much higher standard of rigor on it." And, again, one of the things that I think people are really confused about when they think about social media is this is all about the algorithms somehow suppressing speech.

(Tim O'Reilly at 00:57:14) And it isn't. It's all about what do you promote. You know, I mean, the fact is there's no obligation. You know, it's like people can have completely free speech, and you just don't have to be promoted. You know, right now, I mean, Facebook is being held to the standard—there's some random speaker in Hyde Park and the London police are supposed to round up a crowd for them, you know, because they're saying something inflammatory, and this will be great, you know, because we'll get a mob going. It's like insane. You know, it's like, the fact is, if people are doing stuff that's controversial, like, if they get organic following, great.

(Tim O'Reilly at 00:57:51) But don't let—yeah, there's a lot of signal in there. Anyway, lots—

(Joel Beasley at 00:57:56) Who do you know if I want to do a special episode on that concept? Who do you know that would be good to talk to about that?

(Tim O'Reilly at 00:58:03) Part of the question is who would talk to you about it. Let me do a little homework, and let's take that offline.

(Joel Beasley at 00:58:12) Yeah. If it pops up, great. Matt, Tim, we did it, my friend. We made a podcast. How do you feel?

(Tim O'Reilly at 00:58:19) Thanks. Thanks a lot.

(Joel Beasley at 00:58:20) This is great. Talk soon, Tim. Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you would like to hear discussed on the podcast, either add me on LinkedIn or send me an email: [email protected].

(Joel Beasley at 00:58:43) Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.