Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
Unveiling Keith Teare's Journey: Political Activism, Coding, and Tech Investments with AI
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Keith Teare
I'll give you an example using my approach, but just to give you an example, Y Combinator is obviously the biggest incubator on the planet. In the last 10 years, Y Combinator invested in 140 B rounds. So not very many, they do 250 companies every six months. They did 140 B rounds. My algorithm
Keith Teare is an English-American tech entrepreneur. He has founded or co-founded several companies since the early 80s and was a founding shareholder of TechCrunch. He is currently the CEO of SignalRank that he's been working on for a long time behind the scenes. It's an investment platform that uses data intelligence, which is changing venture investing one data point at a time. He uses his background in political science and sociology to analyze market trends and share them with the world in his newsletter That Was The Week and the show with Andrew Keene. Keith has been an influencer in the tech space for four decades and is genuinely a kind person and according to his wife She's Never Met Anyone More Optimistic Than He Is. We talk about his journey and how he thinks, SignalRank and his approach to building a data-only company. So here we go. This podcast is brought to you by GUT, fostering a culture of innovation to build better products, ventures, and cultures. I'm Maria and I enjoy adding value and helping wherever I can, widening my spectrum of thoughts, even if it can sometimes challenge the mainstream. This is why we give data a voice and co-create a collective intelligence involving both people who are not always in the limelight and those who are, in order to learn from each other and spread knowledge and critical thinking. All I ask is to rate the show, leave a review, and share it. It's a fantastic way to help other podcast explorers discover our content. I'd love to know more about you and your preferences to continue producing this kind of content, so type in go.gut.com slash talks. It takes 60 seconds to complete. That's g-o dot g-g-u-t-t dot com slash talks. Now let's get started.
In Palo Alto, right? A few years ago. And I met your dog as well. And you said a few interesting things, and I kept following you, what you're doing. But also, That Was The Week that you started about three years ago. But that also inspired me and my co-host in creating a show in Italian And I'm not just saying this like that, this is just really on top of mind now. I remember you were talking a lot about the creator economy years ago. You were also writing on GetReview, then you switched to Substack. It boosted me also in continuing the podcast. So you were just telling me that you're not that interesting offline. So I just want to make sure that even though you think you're not that interesting, maybe I think you, you have an impact on lots of people and one of them is me. So. I have things I want to talk to you about or ask you about, primarily AI tech investing, but most importantly, how does your brain function? Like, how do you analyze everything? Your background is in political science and sociology. Yeah, and you've been an entrepreneur for like four decades. So if you just go back in time into your upbringing as well, and then your studies later on, and then moving into tech, is there anything that prompted you to get into what you're doing?
I don't know if prompt is the right word, but the decisions I made in my life were all motivated by the same instinct. And the instinct is to escape irrelevance. I grew up in a very, very poor part of the UK. It's a beach town called Scarborough that has a small working class population that serves the town, which is mainly retired people. And I was the oldest kid of six.
that I was destined to work in and the strongest feeling in my head was I had to escape. I had to get out somehow. And so that sense of escaping bad outcomes is a very strong driver for me. You know, when you live in that situation, something bad happens pretty much every week, something bad happens. So you grow up expecting Negative Outcomes. So somewhere you have to find within you optimism. And my wife always tells me she's never met anyone as optimistic as I am. Well, that optimism comes from fear of the opposite. And that is deep in your life, deep in your upbringing. And it really is a sense of survival. It really is an ambition. because that would imply confidence. It's the opposite. It's not wanting to fail and being absolutely sure you will unless you do something. And so that is the key and I can give you a few anecdotes. But in England at that time when you were 10, I was 10 years old. I was like born on the August the 27th. And the cutoff date for a new year in England is September 1st. So I was like the youngest kid always in my year at school. And in England at that time at the age of 11, you did an exam called the 11 plus. And that exam determined whether you would go into the A stream or the B stream for education for the rest of your life. And it was a kind of an IQ test style exam, patterns, shapes, in numbers, just logic, logic and patterns, abstract thinking. And I was only 10 when I took it. And I'd never seen one of these tests before, they didn't give you a practice test or anything. So despite that, I kind of was a borderline pass fail. And I remember when you were a borderline pass fail, the headmaster of the school interviewed you to determine which side you would come down on. And in my interview, they only asked me one question because probably there was a lot of kids and they didn't have a lot of time. What does your father do? And I didn't know what my father did because my father worked for the British Secret Service. in that he was a former coal miner who during the Second World War got trained in Morse code and cryptography and became a kind of a field station communications guy in the Secret Service. So he couldn't talk about what he did. So my answer was, I have no idea what my father does. Fail. They put me in the B stream. And when you're in the B stream, that's the leave school at 15 path.
Luckily for me in British politics, Harold Wilson was elected as Prime Minister in 1964. I was born in 1954. So the year I was put in the B stream, he became the Prime Minister and he brought in a Dean age 15 exam that let you try again. And I got straight A's when I was 15. And I went back into the A stream, which is the stream where you can go to university. The minute I showed up in the new school, age 16, immediately you're doing university applications for two years later. My teacher told me not to apply because I had no chance of going. Before they even knew me, just because of where I came from. So, so you can tell either you become a rebel or you You basically carry out the life they tell you is your life. So I became a rebel. I was a rebel. And I was a rebel all the way through. That's why I studied political science and sociology. I wanted to analyze why everything was so bad.
Wow okay so a few things here where did this strength come from at a really young age to be like no I'm not gonna take this like what's given to me for an answer or the only option?
Yeah you know I don't know the answers to that but my clue is this I don't think it was strength I think it was escaping. So just to give you a sense of it, what did I do? We had a, in my, where I lived, there was a library, a public library. I used to go to the library every day and get books. And I would stay in my room and read. And it was more like escaping into a different world that was under control. And in your thoughts, you can be anywhere. And Your Life Can Be Anything. So it was really not as much strength. Although, like all kids, I reacted against parental negativity, like my children do today. It's like I see the other side now. And I did have a kind of that rebel streak. I can remember a few little things like in 1968, the British Army was sent into Northern Ireland. and I remember watching on BBC Irish Catholics protesting for civil rights because they didn't have the equal votes or anything in those days and I remember seeing a policeman hitting old ladies with a baton and instinctively I supported the old ladies not the police and you know that meant you were on the side of Irish Catholics, Republicans against the British state And I grew up being a strong believer in Irish independence. And my dad was in the Secret Service, was on the other side, obviously. I ended up founding and leading something called the Irish Freedom Movement, which was a campaign for British young people to support Irish independence. And that kind of repeats over and over. I led campaigns against Margaret Thatcher's invasion of the Falkland Islands. against the war in Iraq. I grew up wanting the underdog to win.
When I was younger I thought politics could change the world. Obviously at some point it can, but when I was still very young Ronald Reagan became the US President and Margaret Thatcher became the Prime Minister of the UK. and it became very clear that politics was not a path that you could control the outcomes very easily. And I learned to code when I was about 21. I bought some of the first computers. I was very curious about computers and I learned to code in order to do digital campaigns to raise money for political causes. So they were kind of joined together and as soon as I could code all the people I knew started asking me to do things for them. Could you do this? Could you do that? And it ended up when I was in my late 20s, I started writing systems. I wrote the A&R music system for Warner Brothers Music. And I wrote a skills passporting system for mobile oil that determined whether somebody could go on to an oil platform with a skill to do a job. And if they could, it issued the passport that they could do this job. So I started to understand that you could impact outcomes through tech. In both of those cases, fairly small outcomes. But because of my political past, the scope of my ambition was to affect big things. I wanted to impact big things. And so I transitioned really from being, I was always fairly entrepreneurial. If I tell you when I was just finished my undergraduate degree and I had started a PhD, I didn't have any income, but I managed to persuade a mortgage company to give me a 100% mortgage to buy a house.
in Birmingham and I rented out four of the bedrooms. It was a five bedroom house to pay the mortgage to my friends actually. And I live pretty much for free by owning this house. So I always had this kind of survival. How do I make things work kind of sensibility. So tech tech during the 1980s, I kind of made more money than I'd ever seen in my life. because I could write databases and install networks for people and teach them how to use them. So I actually became not poor for the first time in the 1980s. And by 1994, honestly, I don't think I would like myself if I met the 1994 version of myself. I was kind of complacent, lazy, Why? Because I had enough income, I could come to Italy on vacations if I wanted to, and did often four or five times a year, I'd be going to vacation somewhere. And that was a dramatic change for me. I only traveled outside the UK for the first time when I was 32 years old. So I, you know, it's like, it changed a lot who I was. I discovered food that wasn't really bad English food. You know, I so I started to understand bigger things and in 1994 I said to my brother who was my business partner. I'm super lazy. I'm totally complacent. I want you to fire me. You can you can own this business. I give you the whole business as long as you will pay me for six months. I'm going to try something new. And I started EasyNet, which was with a partner, David Rowe. We met through database related work and we, we used a credit card, $35,000 to buy the equipment needed to start an internet service provider, which in 1994, there wasn't, there was not no such thing. Windows didn't yet have the internet. And so I figured out how to make Windows. Be An Internet Client. And we built the software stack to make that possible. It was all open source software. And between June 94 and August 94, I bought, learned, and wrote the software for EasyNet in two months. And EasyNet opened in August 94, 10 pounds a month for internet with dial-up modems. First one in Europe for consumers. And within 15 months, it was a public company on the London Stock Exchange. So this is all, you know, I think it's a common theme, escaping limited contexts to try to get to bigger contexts.
Yeah. Because up until then, my company that did databases was called Brent, which is a district of London, Brent Computer Services. And I did that from the early 80s to the mid 90s. So 12 years, maybe. And I was politically active the whole time through that as well. I was, I was running campaigns. I wrote a book. What if I have it here? Yeah. I wrote this book, which is uh and it's about racism in England um this was 19 published in 1988 so I was I was leading campaigns doing speeches writing books and the database stuff on this you know as to make money but I found the database stuff super easy so most of my time was spent on politics and the database stuff paid for that But by 94, when I started EasyNet, I became 100% focused on on the work. I stopped doing politics as a thing that took a lot of my time. And I was the chief technology officer, which is kind of weird with my background, but I understood networking, I understood the software side. So I was the chief technology officer. And for the first time in my life, I thought I was doing okay before, but for the first time in my life, I discovered that equity in a startup has value, even though the startup is still very young. A friend of mine offered to put 250,000 pounds into EasyNet. And that was super important because we couldn't raise venture capital. There was very little venture capital in London at the time. And insofar as there was venture capital, it wasn't available to people with my regional accent. England's very, very class aware. And when you have a regional accent, you're clearly from the lower part of society. And so when we met venture capitalists, within five minutes, you knew they were not going to give you money. Because you just were not one of them. The pattern recognition didn't work for them. So we got money from a friend. That friend ended up making 50 million pounds from 250,000. So he did super well. And we IPO'd. And after the IPO, companies become very boring after an IPO because what happens is they don't want to innovate anymore. They just want to grow revenue. So I actually left EasyNet a year after the IPO. And in order to leave that, to sell my shares, which I actually probably didn't have to, but I was naive and I was told I had to, so I did. So I sold, I owned 20% of EasyNet. I sold it for 5 million pounds. A year later, the same shares would have been worth 200 million pounds.
Well, there was a buyer, an institutional buyer called Hall-Govett, which is a big private equity house. I should have been clever enough to realize if they wanted to buy them, I shouldn't want to sell them. But I did sell them and I moved to the US. And I mean, from my point of view, I had more money than I could have ever dreamed of. I could buy a house. which I'd never been able to do before. I did buy a flat in London actually earlier, but this was the first time I bought a house and I was two years married by then. I got married in 92. We only had our first kids in 2001. So between 96 when I left the UK in 2001, I was in Silicon Valley doing the next level up, which was again escaping. I thought the UK was too constraining. the experience of trying to raise money and failing and having to get money from a friend and then do a very fast IPO to all of that was dealing with constraint. And so I moved to Palo Alto really to try to be in the place where there were the least constraints.
But the UK risk appetite is extremely narrow. So there are great venture capitalists there now, people like Saul Klein at Local Globe or Reshma Sohoni at Seedcamp and many, many others. So I do think that the UK is a healthier ecosystem than it's ever been. But it isn't Palo Alto. Palo Alto is many steps above that.
So you still think that where the tech is and where the next, I don't know if we call them now unicorns, deca-corns or whatever-corns, is it still there for the next 10 years?
Yeah, it hasn't been only here for quite a while now. I think the percentage of unicorns coming out of Silicon Valley is now at or around about 50%. China is second, and London and the UK is third, I think. So clearly, things are flattening and leveling. But if you look at the recent developments in AI, Think about 90% of all the AI investments this year came from Silicon Valley. Nine zero. So it's still much earlier than everybody else. It's prepared to put money into new things. People call that a bubble. I actually think a bubble is the word to describe an opportunity. When there's an opportunity, money is attracted. Usually Too Much Money. But the reason it's attractive is because of the opportunity. And so the narrative really shouldn't be about the bubble. It should be about the opportunity. And AI clearly is an opportunity.
Yeah, it is. It is. And that's all we hear about, I guess, today. But I want to go back to when we met a few years ago. You said, I'm paraphrasing here, but you said what mobile phones did to the internet is that it made it global. It kind of democratized the impact where tech used to be for the elites. Where is AI or what is AI doing to the world today?
Yeah, well, so when I when I said that, I said it from two points of view. The first is that billions of people on the planet can have a smartphone, and so have access to everything that the smartphone can deliver to them. The second reason I said it is people who write software have only two points of distribution. Google App Store and Apple App Store. And through those two, where you can upload software today, by tomorrow, 4-5 billion people theoretically can use it. So the second globalizing thing is the infrastructure for software distribution is fully globalized. Now, AI benefits from that. because AI, at least from the point of view of the user, is just an app on a phone or an app on another device. Or, you know, with this new device that was announced this week, the PIN, it's software running on a device that doesn't have a screen. So AI, to me, couldn't really be distributed if it was not for what happened with mobile. But the second part is it couldn't even be created without developments in computational power. And there I'm thinking of CPUs and GPUs and things called ASICs, basically dedicated chips for AI. And when you put together mobile, the cloud, and the hardware platforms that have evolved,
Large language models, which is what everyone thinks about today as AI, become possible. And a large language model is basically best thought of as brute force intelligence. Brute force meaning you just give it everything. It can use everything you give it. And just because of the weight of what it has access to, its training makes it have access to everything the human race can give it. And so suddenly, you know, if you take any discipline, let's just take a discipline like physics. Suddenly, there are very few teachers on the planet better than the AI at physics. And that's true for any specialism. For any specialism, the AI will be better than probably the top 5% of experts in the field. So suddenly, the human impact in leveraging knowledge So as to enhance human capability, it's never been, it's never been this good. And of course, it's going to transform everything just just like the calculator transform math, AI is going to transform every single discipline, and it's going to do it quick, quickly, it won't be slow.
Question, going back to what you were saying about education, you know, when you were at school, and you had set expectations by somebody else and the society and, and how the system worked. In hindsight, do you do you wish you had an AI teacher? Like, how can you see traditional education changing?
Well, if you combine all the things I had then, which which basically was a typewriter, a public library and school. That was probably a slow version of the human race, collective knowledge being delivered to me and my ability to consume it and then develop who I am by using it when I wrote things. So that process of gathering together knowledge, making it available, And then people taking advantage of it to make themselves. That is kind of somewhat eternal in human history and consistent. So the thing that changes is the tools to do it with and the better the tools get humans are super good at taking in and learning from information. And my children all speak fluent Spanish. Why do they speak Spanish? Because from the age of less than one year old, we put them into Spanish-speaking nursery environments. We put them into elementary school for six years that was only Spanish, no English. And they taught themselves English, because it's easy, because everyone around them was speaking English. But Spanish is natural to them. So that ability of a human being to take in and become something is a huge. So the better you can package up the things and give them the, you know, what do children do most, they ask questions, often irritatingly, so that, hey, what, what is this? Hey, what does that do? Hey, what about this? Well, suddenly they can ask as many questions as they want of an AI that never goes to sleep and knows a lot. Now, you know, there's dangers in that because as we both know, the AI is not perfect yet, probably never will be perfect. So they have to also have critical thinking. They have not just assume that what they're being told is right. The same with the teacher, by the way. You shouldn't assume the teacher's always right. So they need critical thinking as well. But now the ability to access and engage with things is going to be fantastic.
You touched on a point that I usually like to talk about this critical thinking here. As we're recording this, I mean, this week custom GPTs have been released, right? And I was telling you offline that I tried to do gut AI and like you also tried to do signal rank AI and so on. I didn't want AI to just give the answer but to make the user think so they get the right answer but it just gets it because I think for me at least the way I learn but the way I also teach is I don't just give the answer I don't make it hard I would say but I would try to get the others to work a little bit hard even though if they get tired after just to never forget because it's so easy to forget as well
But I didn't manage to do that on custom GPT just yet.
Words and timings
ButIdidn'tmanagetodothatoncustomGPTjustyet.
Keith Teare
So I started and failed to do the SignalRank AI version this weekend. I started and failed three times. Every time I failed, I learned something about why I failed. And so the most recent one I did is way better than the previous ones. It still isn't perfect. It's all to do with What Do You Give It To Learn From? And what parameters do you put around it for using what you give it? And to what extent do you let it think for itself versus training only on your data will result in very different outcomes. So yeah, there's a lot of trial and error like with any tool. I once bought myself a synthesizer because I thought I could hear music in my head and I wanted a way to make music and I'm not musical. So I bought a synthesizer and in that case, I failed. I think I've bought a synthesizer maybe three times in my life. Every single time I ended up selling it cause I couldn't figure it out. With ChatGPT, I'm figuring it out little by little.
Yeah. Well, it's your area, right? So you understand it. Yeah. It's, it's, it's actually changing the way we work with software anyway, before we would. We're not using a used, we're not used to working with, we're used to just giving a command and kind of knowing what to expect and what to get.
So, yeah, I, that's a good way of thinking. So my primary use of ChatGPT this year at SignalRank, I, I invented what's called a heuristic model for predicting outcomes in venture capital. And to do that, I had to write a programming language called SQL. And back in the 80s, I actually did learn SQL. I worked with the very first versions, but I haven't used it for 30-40 years. So I had kind of a basic memory of it, and I spent maybe, before ChatGPT, I maybe spent six months writing my own sequel using a product called Snowflake, which is a public company here. And I'd probably say I was like, you know, a B minus level of capability in writing sequel. Then ChatGPT came out, and I started sharing the code I'd written with it and explaining my purpose and it upgraded my code to A++ and I could just copy and paste it and run it and test it and then if there was an error go back and say no that look this is an error and it would say oh yeah of course I forgot this and what it is like to me actually is an expert in SQL that's way better than I am But like a professor hasn't been coding itself for many years. So sometimes gets it wrong. And if you point out where it gets it wrong, it fixes it. So it's highly iterative. I can tell you SignalRank's code base could not have been written by me and I'm the only human to have touched it.
I've written the whole software myself for SignalRank. Now, I will say, Single Rank has a data science team. They have a totally different project. It's not what I described as a heuristic model. They have a machine learning model, the job of which is to be better than me. And they're miles ahead of me in their capabilities. But their model is worse than mine right now. My Model's Better.
But it's also it has to do with the approach. I mean, you said you have 10 years, it's not just number of years, obviously, it's your experience. And what you know, I don't know how to explain. It's just not a number in that case. Yeah, this is hard to just, like, I don't know, do in like a year or two, right?
Yes, it is. The first the first tests I did, I've still got the spreadsheets were in 2013. And in 2013, I wrote spreadsheets that tried to test some assumptions I had about venture capital. Huge spreadsheets that Google had trouble running because they were too big, because I needed lots of data to test on. So I had spreadsheets with maybe 70,000 rows. which doesn't seem like very many if you're using databases but for a spreadsheet that's a lot of rows and then lots of calculations happening against that and 2013 I kind of knew that my thesis was right but it took some trial and error I only did the first investments using the thesis in 2016
By 2019, those investments made a lot of money, 20 times the investment in three years. And then I knew it worked and I only started SignalRank in 2021. So from 2013 to 2021 was this idea was in my head. I was testing it.
And I'm one of those people that unless I believe in something, I can't do it. I can't sell stuff I know is not right. I can only really get passionate when I know something is right or works. And the goal of SignalRank is to make venture capital available to everyone as an investment type. Most people in the world can't invest in venture capital because it's too risky. Actually, what I do is I make it not risky, which seems a weird statement. But basically, you're using statistics and math to take away the risk. And when you do that, you're left with the good stuff. You know, if you if so, the goal of single rank is to remove risk. And what remains is way less risky.
How did you come up with this idea? Like what is it just an overnight thing where you had or like while showering or I don't know you had this because it's as you said it's a big statement to do actually.
It's it well it started with a with a question. I wrote an essay in 2013. It's called the essay is on medium is called This Is Not Silicon Valley. So if anyone Googles This Is Not Silicon Valley and Keith Teare is the name it'll come up in Google. And that essay, I wrote it on the island of Hainan in China. I was in China doing two speeches, one in Guangzhou and one in Shenzhen. And there was five days in between where I had nothing to do. And Hainan is like the Hawaii of China. So it's one hour flight from Guangzhou. So I said, okay, I'm gonna, I had five nights on hotels.com free hotel. because of all the stays and I spent my five nights and went to Hainan and I wrote This Is Not Silicon Valley during those five days and it has a pencil drawing. I had the first iPad where you could use a pencil. I did this pencil drawing called the Valley of Death and on the left hand side was all the seed investments. Then there was all the death of the companies that didn't make it. And on the right hand side was all the growth investments by well-known brands today, people like Tiger Global, that were writing very large checks to companies that were successful. But in the middle was death. And the problem there was a lot of good companies die. Companies that shouldn't die do die. And they're not really given enough time And at that time, there was this idea in the air called Lean Startup, which Eric Ries coined, that basically said, you can build a startup for nothing, and it can succeed very fast. And I felt deep in my soul that that was wrong. Because the key part of a startup is people. And Eric Ries was right that the cost of computing came down. but people are not free. And so actually a good startup needs two things. It needs time and people. And that's the one thing lean startup didn't allow. So the only startups that got funded were the successful lean startups got funded. But anything that had a bigger vision that wanted to do something big couldn't get the money to do it anymore. And so I felt like this was broken, that from the point of view of innovation and entrepreneurs, you needed time and people. So this is not Silicon Valley was my statement that short-termism was taking over from long-term thinking. And so then I said, okay, how do you get, by the way, at that time, there were lots of unicorns beginning to be formed. By 2013, the word unicorn, I don't think it had been coined yet, but maybe it had. And so the question I asked is, how could more founders understand how to build a unicorn and benefit from it? And how could the early investors who invest in these companies not lose their investment in this valley of death? And so I had lots of questions in my mind about the very structure of venture capital. And that started my brain thinking. And the first thing I asked is, are there any patterns that help you understand very, very early, which companies are likely to be unicorns? And so the math of SignalRank is answering the question, as early as possible, what are the chances of this company becoming a unicorn? And I can answer that question about three times better than the average performance in the market. I can answer it at the seed round, the A round, the B round, all funding rounds for those who aren't familiar with venture capital. At the B round, I can predict a unicorn one out of three times.
Way better than any font. Yeah. And I have a kind of a way of doing that. The reason it works is because of the, as you were saying, that experience you kind of understand the patterns, you have to think abstractly. But if you think abstractly, the patterns that give a signal that this is a potential future unicorn very early, those patterns are discoverable.
The thing that's going to my mind is, you're taking qualitative data primarily, and making it quantitative with numbers. How is that working? In this case, because you're having good predictions. I mean, I mean, there's a fund here in the UK, I forgot the name, but they were supposed to raise their third fund, and they're not doing it. And they invested heavily in about 50 startups already.
Yeah. Well, I'll give you an example using my approach. And I will answer your question. But just to give an example, why Combinator is obviously the biggest incubator on the planet. In the last 10 years, Y Combinator invested in 140 B rounds. So not very many. They do 250 companies every six months. They did 140 B rounds. My algorithm predicted 124 of them. Now, they spent $400 million on those 140 rounds and the value
Today is 2.5 billion. Now, here's the interesting thing. My algorithm also predicted another 170 B rounds that they did not do. Okay, those would have produced the same outcome. Roughly $500 million would have been needed and 2.5 billion would have come out. So they're really good, but they missed 50% of the opportunity.
Yeah, I just... Anyone else investing or it's based on... Other people are investing. But I'm focused in on the decision makers at YC. Yeah. What did they do? And what would they have done if they listened to the signal rank algorithm?
And the answer is if they listened to the single rank algorithm, they would have doubled their money. Yeah. And now, and they're, they're really good. So, so we wouldn't have improved the ones they did do. So it's finding the ones they didn't do where we made a difference. And that, that analysis could be done for any investor. I could do that for any investor. I'm just choosing Y Combinator. Now, how do we do it? How do we turn qualitative data into quantitative? You're probably familiar with this idea of Moneyball in sports. Moneyball is using data to build teams. What we are is Moneyball for investors.
Everything is launched, except we only write checks at Series B. But we score everything. And the reason we score everything is you can't tell how good a Series B is without information about the prior round. So a Series B does not stand alone all by itself. This company has a history. Now the company's history is not available as public knowledge. You often don't know where the team went to college or There are many venture investors try to score the company through the team and various other things. We don't. We score investors, only investors. Now an investor has a history and the history resulted in outcomes. And the history is either repetitive or not. You know, it's like, like in music, a band can have one hit record or it can have a hundred. The One That Has 100, you can predict the next one, the chance of a hit is high. So we do that with investors. We look at their history, very specific to the stage. So we have scores for pre-seed investing, seed investing, A round investing, B round investing. We don't mix these scores. It's how good are you at this stage? We also score the partners in the fund. The Week.
And so that creates a filtering effect, like a league table, like in soccer, the number one, number two, number three investor at pre-seed in the last five years. And the definition of good is measurable. It's how many investments did you make? How many of them produced positive outcomes versus negative outcomes? How positive was the outcome on a scale of one to a hundred? How Many Unicorns Did You Make? And How Many Did You Not Make? As a relative score. Many, many things. We have about 200 different data points for each investor. And then we create a league table. Now, if you then look at a funding round, there may be five investors in a funding round. And I can say, well, what is their collective score? Now I have a round score. And when the company is doing the B round, I'm probably have three round scores, one for seed, one for A and now one for B. So that gives me a company score. So from investor score, I get to round score, I get to company score. The correlation between a company score and the future outcome is very high.
Every day it gets new information about new funding rounds and it calculates a threshold score equal to this round being in the top 5% of all rounds. And if this score beats the threshold at the B round, that company has a one in three chance of being a unicorn. And the threshold changes every day according to the data.
One of three is so good. Yeah. You probably can't improve on that easily. I've tried many ways to improve on it, and they're almost none of them ever result in improvement. So I think we'll see. Probably we can improve on it. The machine learning guys have a better chance than me, because you need to see more variables and how they interact than the heuristic can deal with. So probably they will eventually get there. Now just stand back a little bit. Venture capital, everyone knows, is what's called a power law game. I'm sorry if I sound like I'm preaching here. The power law says this, if you make 100 investments, one of them will make you 100 times your money and return the entire investments. So it's gambling. Venture capital is gambling.
It's very different to say in the public markets you can invest in an index like the FTSE 100 or the S&P 500 and you can get the average results of the market into your bank account. In venture capital historically there's been no way to index venture capital because of this widespread of outcomes. If you would have invested in every single B round in the last 10 years you would have had to have made 19,000 investments and You would have gotten 300% gain So it would have been a good decision if you could do it, but nobody can invest in 19,000 B rounds What single rank has done is created a subset of B rounds? that in five years the average outcome is five times your money and The Average. And in order to achieve the average, you only need to make 10 investments. 10 investments removes the randomness. And so instead of the power law where you've got to pick one in 100, we can pick 10 and deliver five times your money in five years on average. Of course, that's not a promise. Markets change. But based on history, that's what we would have done. and we can pick one in three unicorns based on history. So what we've created is the possibility of indexing venture capital for a normal consumer. It's better than a savings account.
Well, so now you know that we're very early because the full end game for SignalRank is to own lots of companies' BRounds, hundreds, and for SignalRank to be listed on the Stock Exchange so that you can buy a SignalRank share as the equivalent of buying all these BRounds.
I Know When I'm Planning to Do It. I don't know if my plan will survive the real world, but the plan is between 2027 and 2028. Okay, so we raise money today. We have a pool of shares we sell.
Yeah. Going back to like where VC funds get their money from anyway. So, uh, but that's, um, an interesting model because I think you can create lots of other companies from what you're doing.
uh kind of alternative businesses and so on just from the data you have like how can you improve your like basically some funds okay you're not performing well we can help you you know things like that yeah yeah no we can partner with there's a there's a thing in seed stage venture capital called an opportunity fund and an opportunity fund or a continuity fund that's when a fund creates a special pool of capital only for investing more in the companies it already invested in And typically those funds are very inefficient because there's a strong incentive to invest in more companies than you should in your portfolio. It's called adverse selection. We want to partner with those funds to provide the capital. So instead of them raising an opportunity fund, we will supply the capital like any fund investor would, but only in the companies that score highly.
So you create the competition amongst them as well. I mean, this is changing the model anyway. I mean, I've been listening to you for a while, right? And you've been saying that venture capital is changing anyway. And now with AI, it's going to change even more. So where you might not need to raise money or probably in the way you used to. So what's your take on this? And where is SignalRank in this change?
Yeah, it's interesting. So Chamath Palihapitiya did a speech this week, saying that it is not impossible that startups will become one or two people with an AI and will figure out how to make large amounts of value without building teams or infrastructure. He said that if that happens, there'll be no longer a reason for venture capitalists. And David Hornick, who is another venture capitalist who was at August Capital and now runs his own fund, disagreed and said, you know, if that ever happens, it's decades away. And what is required now is infrastructure, all the different layers required to make AI work. And those are real businesses and venture capitalists are going to be still required for a long time. So there's a debate around that. Single rank is basically a statistical analytics company. You could think of us as a fintech because we've created an asset class. The asset class is B round investments through partnerships with seed investors. We basically give them the money and they put it into their companies. So we're a fintech that uses AI to create statistical analysis on likely outcomes. And we're, you know, the heuristic is really, really good at correlating to outcomes. So you could think of us as a digital venture capitalist. We're not data driven. We're data only. We don't use data as a supplement to human decision making. The humans serve the data. The data says make this investment, the humans make it happen. And so The idea of a digital venture capitalist being able to rank likely outcomes, I don't think that goes away. I think it just gets better and better. Today, however, our ability to be good relies on scoring actual venture capitalists. If there were no venture capitalists, we wouldn't be able to score them. So in the future, we'll be scoring algorithms.
Is Signal, I mean, I guess Signal, right, is global, right?
Words and timings
IsSignal,Imean,IguessSignal,right,isglobal,right?
Keith Teare
We made already this year, we only started investing in March, I should tell you that. We only started selling our shares for investment purposes in March. And so the first money came in, in March, and we started investing in June, actually. We've done nine investments in B rounds so far, led by people like Sequoia Capital, Accel, Andreessen Horowitz, I mean, top, top B rounds. So we solved the access problem. Because knowing the good companies is only half the story, you have to be able to put money into them. And so we solved the access problem. And I think we'll end up in the next 10 years, we'll probably do 500 investments in B rounds as part of this index. And no single investment will be able to really change the results. It's going to be the collective of all of them.
sorry i forgot that's global and that's global yeah oh that's why i was saying it so in the first uh nine uh i think six are in the u.s and three are not um i have a question here because i did a podcast also with bernard moon and uh you know they're started in like asia australia the u.s and they're now in Saudi Arabia. So in Saudi Arabia, other than getting football players and investing heavily in culture, entertainment and so on. There's movement also in the startup world, but also it's a young population. And it's an active population. And now, the six GCC countries introduced a unified visa. for the six countries. And that's going to affect business, not just tourism. So what's your take on the region, actually?
I think we're going to see the largest change in geographical wealth in the next 50 years that we've seen for a long time. Saudi Arabia, you know, is kind of contradictory because it's a modernizing economy. but it's starting from a set of structures that you couldn't normally describe as modern. So it's very modernizing. The neon project in the desert is, it's either crazy or spectacular depending on your point of view and probably both. But now the other part of the world that is very different but also one needs to understand it is Africa. especially in Nigeria. Nigeria is going to become one of the biggest countries in the world in the next 50 years and wealthiest. It has natural resources and minerals as lots of Africa does. And then of course the one we all know about is Asia where 66% of the world's middle class will be in Asia by 2050. And the middle class are the people who spend money en masse. So You know, we're living in a time when the global footprint, if you think of it as weighing scales by country, the US is going to go down to weigh less. Asia already is up, weighing a lot more. And India and the Middle East and Africa are all going to come up as well. So many things could thwart this. For example, the decline of America's ability to be the world's policeman will lead to military conflicts that would not have happened earlier in history. I'm not saying it was good that America was the world policeman, but it did make a difference to decision making, that they were so powerful. And that decision making now can factor in that America is less prepared to act. And I think that it's a dangerous era, an era of potentially huge wealth coexist with the danger of potentially huge conflict. And so managing that is non trivial. And it actually would involve America understanding that it is not going to be number one forever and managing its future relationship to the world in a peaceful way now. That's probably the biggest factor and Saudi Arabia is you know in a good position from that point of view because it's friendly with America and with China.
Well you know I'm getting old so at some point My wife is from Cape Town in South Africa and I have my eye on a nice little beach cottage where I can peacefully decline but not for a probably I've probably got a good 10 years in me still and after that we'll see.
So you're not doing any political activism at the moment. You're just like on That Was The Week, like Andrew always cuts you off when you try to go into a political debate or something. So I'm sure you still have a keen interest in politics because this shapes the way you have your thoughts and ideas and so on.
Well, I, you know, my belief is that political activism is only fruitful if large numbers of people want to make change. I don't believe real change comes in private rooms. I think real change happens on the street when you could see it even in, if you go back in American history, the civil rights protests would not have been as successful if it wasn't for the fact that millions of people were marching. And the politicians felt compelled To Make Changes. And of course, they always filter down the change to the minimum they can get away with to try to appease the population. So a political activist is only as powerful as the underlying sentiment. And the problem we have today is we've become individualized to such an extent that most people's perspective is very personal to themselves and not analyzing social structures and the need for change anymore.
I think there's a kind of a dance that happens between innovation and freedom. I think, you know, these are deep questions. But the ultimate question is, You know, in the early part of the 20th century, it was a popular point of view was that in order to make societal change, you need a revolution. And people focused on revolutions, not economics. However, a real reading of history tells you that economics and innovation produce enough wealth to free people from slavery to work. And so my current belief is that innovation and economics, especially AI, and the ability for machines to do jobs will lead to the largest amount of wealth the world has ever seen. And the problem will change from producing wealth to distributing wealth. How does this wealth get into the hands of everyone? Sam Altman actually has a separate project called WorldCoin that tries to answer that question.
So I think that the production of wealth and the distribution of wealth are correlated. The question of whether you need a revolution for that wealth to be distributed evenly or somewhat evenly across the world according to people's needs
open question but I actually think that we're closer to utopia in that regard than we've ever been and utopia removes the biggest reason wars happen is nations fighting for ownership of resources and so if the human race can start seeing a future based on wealth production they will also start seeing troublesome nations who want to fight wars as a problem. We may get to that Star Trek moment where we're just human beings and we're all in it together.
Yeah, I mean, I admire your optimism. I don't know how you do it. But I mean, it's good to stay positive. But the distribution of wealth in that sense, I mean, that's really utopia for me.
Well, well, I think the raw material, I would say never try to solve problems that where the capability to solve them does not already exist. So because then you have to do R&D to figure out how to solve the problem. But if the if the capability to solve the problem already exists, then all you have to do is decide to do it. So a bit like Uber, Uber didn't have to invent anything. Everything already existed. You just have to decide to make Uber. So I think that's true of distribution of wealth. We already have the computational power to understand every human being on the planet existence and to allocate through blockchains, credits to them. We already have the political will to avoid a world in which a small number of oligarchs own everything, which is the other possible future.
Well, it comes down to Silvio Berlusconi versus Keith Teare. Okay. And I think most people, even though I'm not very interesting, I do think I'm better than Silvio Berlusconi when it comes to... Yeah, I would vote for you for sure, yeah.
So question here, you spoke about your instincts at the beginning and I always ask this question at some point usually on the podcast is, does your gut or you have some impulses that drive your decisions?
I do. It's probably a struggle for me to identify what they are anymore because they're all kind of intuitive now. I don't know how conscious I am about them. One of my impulses is, is not to be bored. The Week.
So the first impulse is, is not to be bored. The second is to produce an outcome that makes things better than the current outcome. But I wouldn't claim to be like this charitable, benevolent, nice person who only thinks about goodness, because the puzzle probably is the stronger instinct. And my ethics would Mean I Wouldn't Focus On Bad Things. But I think that's a secondary driver. The primary driver is solving the puzzle.
yeah all right so i think i have lots of things to ask but i think we're gonna stop here just being conscious of time thank you so much uh for this and we can and we can definitely do a second one just figure out what you want to do it and when you want to do it and we'll do one and and thank you for giving me the opportunity to talk i mean and my wife always tells me i like talking
Thank you so much for listening. If you've derived value from the show, you can subscribe on platforms like Spotify, Apple podcasts, or wherever you listen to your podcasts. Your feedback is incredibly important to us. So please consider rating the show or leaving a review. It's a fantastic way to help other podcast explorers discover our content. To gain more insights, visit our website at www.ggutt.com. This is www.ggutt.com and see you next time.