20VC: Why Machine Intelligence Will Eat The World Of Software with Roy Bahat, Head of Bloomberg Beta

Roy Bahat is the head of Bloomberg Beta, a new venture fund backed by Bloomberg L. P. Prior to Bloomberg, Roy was chairman of OUYA, a new kind of game console, where he was the first investor. Before that Roy spent five years leading News Corporation’s IGN Entertainment, an online media company with a monthly audience of 70 million people, a top 10 YouTube channel, and the leading website in its category in almost every market globally. Roy served on the board of Revision3 (acquired by Discovery) and was a board observer at Flixster (acquired by Warner Bros). Before joining News Corp., Roy was in the public sector in the office of New York City mayor Michael Bloomberg and at New York’s 2012 Olympic bid. In Today's Episode You Will Learn: 1.) How Roy made his way into the world of VC from working alongside Mayor Michael Bloomberg? 2.) What do all these definitions within AI mean? What does artificial intelligence include? What is machine learning? What is deep learning? 3.) When we talk about AI are we talking pure AI , with the likes of Watson and DeepMind or are we talking consumer centric software with elements of AI? 4.) With data playing such a huge role in the efficiency of AI, do large incumbents like Google and Facebook not have a massive advantage? How can startups get access to datasets? Is AI not fundamentally an acquihire industry? 5.) How important has open source in allowing and encouraging the progression of the machine intelligence ecosystem? What more can be done to further it's growth? 6.) With the rise of machine intelligence, what does the future of work look like? How will we live in a world where 47% of white collar jobs will be replaced by machines and AI? Items Mentioned In Today's Episode: Roy's Fave Book: Watershed Down Roy's Fave Blog or Newsletter: Media ReDefined by Jason Hirschhorn, Asim Azar, the exponential view As always you can follow The Twenty Minute VC, Harry and Roy on Twitter here! If you would like to see a more colourful side to Harry with many a mojito session, you can follow him on Instagram here! The Twenty Minute VC is brought to you by Leesa, the Warby Parker or TOMS shoes of the mattress industry. Lees have done away with the terrible mattress showroom buying experience by creating a luxury premium foam mattress that is order completely online and ships for free to your doorstep. The 10 inch mattress comes in all sizes and is engineered with 3 unique foam layers for a universal, adaptive feel, including 2 inches of memory foam and 2 inches of a really cool latex foam called Avena, design to keep you cool. All Leesa mattresses are 100% US or UK made and for every 10 mattresses they sell, they donate one to a shelter. Go to Leesa.com/VC and enter the promo code VC75 to get $75 off!

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Harry Stebbings?

  Welcome back to the 20 Minute VC with your host Harry Stebbings, or you can add me on Snapchat @H Stebbings. That way, when new guests come on the show, you can have your input and suggest questions to the guests to answer. Now, to today's guest, I'm super excited to welcome. Joining me today is Roy Bahat, Head of Bloomberg Beta, a new venture fund backed by Bloomberg LP. And prior to leading Bloomberg Beta, Roy was the chairman and first investor at the games console company Uya, which received investment from the likes of Kleiner Perkins and Alibaba, before being acquired by Razer. And before that, Roy spent five years leading News Corporation's IGN Entertainment, an online media company with a monthly audience of 70 million people, a top 10 YouTube channel,

• 0:45Harry Stebbings?

  and the leading website in its category in almost every market globally. And Roy also served on the board of Revision 3, acquired by Discovery, and was a board observer at Flixster, acquired by Warner Brothers. If that was enough, before joining News Corp, Roy was also in the Public Sector Office of New York City Mayor Michael Bloomberg, and at New York's 2012 Olympic bid. However, before we dive into the show with Roy today, the 20 Minute VC is brought to you by Lisa. Lisa are like the Tombs and Warby Parker of the mattress industry. Why? Well, having created an awesome online end-to-end mattress buying experience, replacing the worst-ever experience, let's be honest, of buying and testing a mattress in the showroom, that still is not enough for them, as they like to give back.

• 1:29Harry Stebbings?

  Which is why for every 10 mattresses Lisa sells, they donate one to a shelter. These mattresses are also not just any mattresses. They're engineered with three unique layers, including two inches of memory foam and two inches of latex-like foam, designed to keep you cool. Not only are they engineered to perfection, but all Lisa mattresses are made in the UK and US and can be shipped for free to the US, UK, and Canada. They also have a 100-night trial, so what is the risk? Head over to leesa.com/vc and enter the promo code VC75 at checkout to get a cool $75 off. We'd also like to thank Matamark for providing all the data and analysis for the show today. And if you've been living under a rock and have not signed up for Matamark Daily,

• 2:14Harry Stebbings?

  then check it out for the best VC and startup articles, thanks to @thinker on Twitter. However, enough from me, so without further ado, I'm delighted to welcome Roy Bahut, Head of Bloomberg Beta.

• 2:31Speaker 01

  You have now arrived at your destination.

• 2:34Harry Stebbings?

  Roy, a huge welcome to the 20 minute VC. Thank you so much for joining us today.

• 2:38Speaker 02

  You're welcome. Thank you for having me.

• 2:40Harry Stebbings?

  So talk to me, Roy. How did you make your way into this weird, but albeit wonderful world of venture capital?

• 2:46Speaker 02

  You know, I was seven years old and I went to my mom and I said, Mom, I really want to be a venture capitalist when I grow up.

• 2:53Harry Stebbings?

  So did I.

• 2:55Speaker 02

  Yeah, there you go. Figure, what was that, two years ago?

• 2:58Harry Stebbings?

  Very good, very good. It was actually three years ago, Roy.

• 3:01Speaker 02

  Uh, no, I mean, I actually was a big skeptic about technology generally. I got offered my first job uh, in technology, writing HTML in 1994. And I looked at these blue links. And I was right out of high school. I looked at these blue links and I thought, nobody will ever use this thing. So it just goes to show how smart I am. And then it took more than a decade before I realized my folly and that after being in government, you know, I worked in New York City government after nine 11, after working in professional services, after working at a big, you know, Fortune 500 media company, I finally, finally realized, as is common for many, mostly by accident,

• 3:43Speaker 02

  that hey, it turns out startups are the fastest and most reliable way to change the world. And then, you know, I started one um, and helped a bunch of others, invested in that kind of thing. And I had zero intention of becoming a VC. In fact, most of my own experiences with VCs were fairly negative. Um, but then the Bloomberg guys called and said, hey, we want to do a corporate VC and we've worked with you before because, you know, ages ago, I'd worked with them when I was in city government and it looks like you know, something about startups. Would you like it to do it together? And my initial response was a long, rambling rant about the ills of VC and corporate C in particular.

• 4:24Speaker 02

  And then the Bloomberg guys did something very smart, which is they said, okay, okay, that's great that the model is broken and VCs lose money and, you know, they're not actually that helpful and all this stuff, but. If you had to design it in a way that would work, how would you design it? So that was an interesting thought exercise and we invented what became Bloomberg Beta, which is to say, fix the issues with corporate VC by investing just to make money, because it's the only way I know to have a fully transparent relationship with founders where they can really understand your intent. Not that making money is the most important thing, but just that it's legible. And then I thought about all the things I didn't like about VC and then the question was, could you just flip the bit 100% the other way on all those things? So for example, I really didn't like the secrecy of it.

• 5:09Speaker 02

  I didn't like that I would spend weeks angling to get a meeting with somebody through a very trusted relationship, only to end up in the room and realize, oh crap, you don't actually fund this kind of business. So that sort of thing. And then the other thing is, I didn't like the partnership dynamics of most of the venture firms that I encountered. I mean specifically, you develop a wonderful relationship with a partner or partners, and then comes the grand moment of decision and all of a sudden you find yourself staring down a bunch of people who have much less experience in the thing you're doing, haven't really thought about it that much, are clearly not taking your thing with the same seriousness that you're taking it, and are in a position to decide whether to give you money. So I remember, you know, early in the life of Bloomberg Beta, one of my partners came

• 5:53Speaker 02

  to me and he said, I'm going to recommend we do this deal. And I was like, what do you mean recommend? I don't have any opinion on whether we should do that deal or not. If you want to do it, you should just do it. And that's why we have this. Any one person says yes on our team, we do a deal. So anyway, that's how it came to be.

• 6:09Harry Stebbings?

  Oh wow, so at Bloomberg, you don't require unison agreement to make an investment decision?

• 6:13Speaker 02

  Oh, quite the opposite. In fact, just last week, I said no to a company that both me and my partner James were looking at, and six hours later, after he and I talked, he wrote to them and said, I'm overruling Roy and I'm saying yes. So that might have been a little disorienting for the founder, but ultimately, I think it was the right decision for us. Because our view is, look, we invest in startups that try to make the future of work better. And everybody knows the best ideas are highly polarizing. And so we've set up a model to try to harness their polarizing effect to make the best decision, which is to say, we want to say yes when there is one person on our team who is very excited about it.

• 6:52Harry Stebbings?

  And you said there about the future of work, and that's very much correlated at the moment with AI. And so I want to deep dive with you on AI, and it seems kind of like sex at a teenage party.

• 7:02Speaker 02

  In that case, I wouldn't know anything about it, but you can keep going.

• 7:04Harry Stebbings?

  Okay, well, I mean, here I wouldn't either because apparently I'm not according to you.

• 7:08Speaker 02

  You're not there yet.

• 7:10Harry Stebbings?

  But from my knowledge, everyone is talking about it, and no one actually has got a clue what it really is and how to do it and work with it. So I want to start by defining a few key terms with you.

• 7:21Speaker 02

  So let me, let me instead take the cluster of terms because I, so I'm technology curious. I'm not the right person to give you the precise technical definition, but the way we think about it is techniques like deep learning and machine learning are different classes of methods used to generate insight from data. And we can get more specific about whether they're classifiers or doing something else, but those are techniques. And AI has unfortunately come to mean a hodgepodge of things. So we actually try to avoid talking about AI. And the reason is AI has come to mean both the cluster of all those techniques and replicating human intelligence with something that feels humanoid

• 8:03Speaker 02

  and, you know, is out to get you like in Ex Machina or something like that. And so we prefer to talk, describe it as machine intelligence, which a number of people have used that phrase as well. It's less loaded and it signals the fact that what we're really talking about here is computers making judgments that people might be able to make themselves or might not be able to make, but computers are making judgments.

• 8:26Harry Stebbings?

  And so with that theme in mind of computers making judgments, it can most definitely be seen as an enabling technology there. So very much like software with its ability to impact every industry when it first arose. When we talk about machine intelligence, are we talking about pure AI? Are we talking DeepMind and Watson with their heavy IP side? Or are we talking about the next wave of software style?

• 8:48Speaker 02

  It's a great question. Yes, I, I think machine intelligence will eat software while software is eating the world. And here's the reason why. The biggest advance in software in the last few years on the side of how the software actually works as opposed to how it's consumed, where the biggest advance is mobile. But the biggest advance in how the software works is the continuing explosion of the volume of data that is available and slightly more slowly that is in use. And so the, you know, the quote unquote big data thing turns out to have produced a number of very valuable technologies, not that many valuable businesses. And part of it is because the data on its own, as everybody knows, is, you know, it's just information.

• 9:31Speaker 02

  Machine intelligence is the layer that takes all of that data and applies it to things that are useful for us. And so in that sense, I do think it eats absolutely everything. I think whether or not, you know, it can pass the Turing test and be confused for a human being is just a lot less important than whether it does lots of useful things for us. Everything from, you know, beating a human cardiologist in the performance of diagnosing heart issues with an MRI to routing us through traffic, to, you know, turning off dangerous processes at a factory, to dot, dot, dot, dot, dot, name your, you know, name your, name your use case.

• 10:14Harry Stebbings?

  And when thinking of the hugely important role that data plays in enabling AI to really have its effect on life and software itself, one very striking issue for me is the incumbency advantage that large players like Google and Facebook have with their massive data sets. So surely there's an instant advantage for them compared to all the little startups with no data sets at all. I mean, how do you view this?

• 10:36Speaker 02

  Yeah, so I think that's true. My partner, Siobhan Zillis, who two and a half years ago was the first one on our team who said, hey, guys, machine intelligence is a big deal, let's start paying attention. And she began by doing this survey that she's now updated annually of more than 2,000 machine intelligence startups. And one of her early conclusions was that at least today, access to the data is the limiting factor, which means, by the way, not only do companies like Facebook and Google have an advantage, but pretty much every old industrial incumbent, if it has a way of getting at its data and making use of it, also has an advantage. In a way, you know, it swings the pendulum back toward the bigger organizations. And so does that mean?

• 11:19Speaker 02

  That startups have no chance? No, there's plenty, there are plenty of opportunities for startups. And I'll give you the warning and the opportunity. The warning is if your startup is, hey, I've got a black box that does machine intelligence. Let me try to apply it to a bunch of different situations. Then I think the going is very hard. You know, you probably have to start with some consulting services. It's, it's just difficult to find examples where that's worked. That said, if your startup is, let me take some pretty well-understood AI techniques and apply them to a new user use case where I can get access to the data and then create a virtuous cycles where you get more data, which gives you a better experience, which gets you more users, which gets you more data, then it can work.

• 12:01Speaker 02

  So I'll give you an example of a company like that is, one of our portfolio companies is Textio, which analyzes business texts. They start with job descriptions and we've all used a spell check and remember, you know, that magical feeling the first time you saw like the little red line under a misspelled word. This is like that, only instead of checking spelling and grammar, it predicts the results of a business document while you're writing it. So it'll look at a job description and say, if you change these words, you'll get more better applicants or you'll get a more balanced gender mix of applicants. And that idea of being able to predict the effects of something while you're working on it is a really powerful one. You know, imagine a journalist, imagine you being able to, you know, to, to see how well

• 12:45Speaker 02

  a podcast episode will do based on, you know, the text of your introduction, you know, that kind of thing. It's, it could be really exciting.

• 12:51Harry Stebbings?

  And for anyone listening, please do invent that. That would solve many, many hours of hard work.

• 12:56Speaker 02

  Although I think you're off to a pretty good start here, so let's not complain too, too much.

• 12:59Harry Stebbings?

  Yeah, no, that is true. Thank you very much for that, Roy. So then are there easily accessible repositories of data that startups can get access to?

• 13:07Speaker 02

  Well, sure, there are. I mean, there's the competitions, you know, like ImageNet. And then Kaggle, which is another of our portfolio companies, publishes some of its data sets in public. And so sure, the issue is if you're looking at the same data everybody else is looking at. Then you really do have to be smarter than everybody else for it to work, which if you're doing a Kaggle competition is kind of the point. You're competing to see who can come up with the best algorithm to describe some effect on a particular data set. But if you're trying to build a startup, then it's good to think that you might have some data that might one day be your own.

• 13:40Harry Stebbings?

  And how important then do you think open source will be in the progression of AI as this next generation software?

• 13:46Speaker 02

  I think that when you think about how the future of work is done generally, not just in software, open source is probably the single most important new method that we as a, as a working culture have invented, period. And let me describe why. One of the things I said earlier is we didn't like that VCs are secretive and so we try to be the most transparent VC possible. You know, an early way we did that is we said, we started writing up our kind of manual for how we would run the fund. You know, what investments would we make? How would we ask diligence questions? All this stuff. And then we decided to publish it. And then we thought, but if we're going to publish it, it's going to change a lot. And we're going to want to invite comments from the community. And it turns out if you want to have a business document that changes a lot in public and collaborates

• 14:31Speaker 02

  with outsiders and insiders to evolve it, open source actually tends to be a great way to do that. And so we published our manual to GitHub and released it under an open source license. And so not just in software and not just in machine intelligence, I think this, the open source idea is a very powerful model for interacting, as powerful as the line managed corporation. There are a lot of misunderstandings about it, but, but it can really work. And specifically in machine intelligence, yes, I think that whenever developers have attention on an issue, the quality of software in that issue tends to go up relatively quickly. And there's no question that's going to be the case for machine intelligence.

• 15:12Harry Stebbings?

  And going back briefly there, you mentioned your investment in TechDio and you clearly have a desire to further invest in the space. However, many warn me of the aqua hire status of the industry. So I'd love to hear your thoughts on this as an investor and why the large standalone AI companies can be built in this space.

• 15:30Speaker 02

  Well, it's sort of like asking, can you build a big standalone software company? In the sense that the software has to be applied to something useful in order to be valuable. So, you know, is Amazon a commerce company that uses software or is it a software company? If what you're talking about is, I guess I'd call them method only companies, you know, some technology that you think can be applied to a wide variety of situations and you're not really building around any one of them, then yes, then I do think that is aqua hire terrain. And I also think that machine learning as a technique, now to get specific on a technique, I mean, it just seems like every startup and every big company has all of a sudden, you know, crossed out data scientist in all of its most wanted

• 16:11Speaker 02

  job requisitions and it replaced it with machine learning. It's just this talent area where everybody knows it's important and is gasping for expertise. And there aren't even that many people who know how to apply the techniques all that well, let alone invent new ones.

• 16:24Harry Stebbings?

  And speaking of the hype there, how's the recent excitement around machine intelligence been for you as an early investor, an early supporter in the space? How's that affected you in your role as a VC? And have you seen valuations soar because of it and increase in competition and deals? What's it been like for you?

• 16:42Speaker 02

  I mean, look, as with anything where you are standing in front of a steamroller as it, you know, comes by, there's a dangerous aspect to it and a powerful aspect to it. The powerful aspect is lots more people are thinking about these techniques and that means lots more founders. And, you know, the world of startups is constrained by one variable, which is the quality, the quantity of extraordinary founders. And so more people doing this is great. On the other hand, you know, it's not so fun when you work on something and really I credit, you know, my partner Siobhan, as I mentioned, and also James Cham, who is another of our partners with, with opening my eyes to it. That's not so fun when you work on something and then all of a sudden the rest of the world thinks it's almost a cliche.

• 17:23Speaker 02

  You know, AI, we funded the first Slack bot company that was funded, which is Howdy. And then, you know, the next Y Combinator class comes and you have many, many, many chatbot companies. You know, those moments of intense competition tend to produce the biggest winners. It's funny, you know, Peter Thiel has this whole thing, of course, about picking markets in which there's no competition. But Facebook, you know, his most successful single investment ever started in the year when I think there were 26 other college social networks started that year. And so the competition isn't bad, it's just, it's a signal of opportunity there. But it can be a little bit maddening and, you know, what you try to do is just ignore it, is have your beliefs and ignore the way that the fads in startup company creation, you know, oscillate around your beliefs.

• 18:07Harry Stebbings?

  Now, this was by no means part of the schedule, but you mentioned your investment in the Slack bot company, Howdy. So I have to ask, what use case do you think will be a sustainable million dollar business built on the platform through the medium? What will that first use case be or look like?

• 18:22Speaker 02

  Isn't that the million dollar, literally, question? I have no idea and I don't want to have an idea about it, by which I mean to say I could blubiate and tell you, oh, I think it'll be task management or document retrieval or analytics or sales. But sort of like the early days of the software industry, it is too soon to see exactly how this will shake out. I mean, I think we tend to be obsessed with first in the technology industry and first loses to best every time. You know, the first iOS app companies were not the best iOS app companies, you know, platform after platform. And I think that's exactly what we're seeing here. What I'm excited by is the world is now a grand natural experiment where founders are just building all kinds of stuff.

• 19:08Speaker 02

  And as the founder of Howdy pointed out to me, Ben Brown, who's extraordinary, he said, you know, it's a different kind of developer that's building these tools. It's a developer that's more interested in. Human oriented design in the the the experience of it as opposed to the technology. And I think that's exciting because we'll have a wider range of businesses. I mean, one of the fascinating things to me about the startup world is how the nature of companies that get started is so similar to the communities that are starting companies. When you have a bunch of young people starting companies, of course, you know, food delivery and, you know, you know, tech, urban transportation, those tend to be things that young people care a lot about.

• 19:50Speaker 02

  I'm excited for when more people who are founders have been in the working world for, you know, thirty plus years and identify new needs, let alone, by the way, we don't invest in the in the space of things that serve people's personal lives. But if I did, I think services for the elderly are the number one untapped opportunity. And just wait till we have more sixty five year old tech founders because they'll see the opportunities in ways that, you know, the the stereotypical tech founder does not.

• 20:16Harry Stebbings?

  Oh, Roy, what a perfect lead into my next question because talking of people that've been in work for thirty years or so, I want to touch on the future of work with you now. And in a recent study, I think they stated that forty seven percent of white collar jobs would be taken away by AI and machines themselves. So where do you stand on this and where humans as a race will operate in the future?

• 20:36Speaker 02

  Well, look, the good news version of it is if you'd been standing here at the dawn of the industrial age and said ninety seven percent of people work in farms, all those jobs will go away, you know, you'd be right. But it happened slowly enough and enough new jobs were created that people satisfied their needs, which I think to be doing meaningful work that's necessary to subsist, you know, to make enough income in order to be able to live and to progress and grow. Like, I think those are the core human work related needs. The only question with machine intelligence is because it is happening more quickly, will it happen so fast that it will be disruptive to the point of creating a crisis? You know, the number one job for men in the United States is truck driver, you know, let alone the self driving car, the self driving truck.

• 21:21Speaker 02

  Could put millions of people out of work. And my view is, A, that we probably will need to reconsider the relationship between work and subsistence at some point in the next 10 to 20 years. And I've been studying policies like the universal basic income in order to figure that out and see what we ought to do about it as a technology industry, because if we're going to cause all this, we have to have a view on where society should go and try to build a constructive society.

• 21:51Harry Stebbings?

  And do you think that is a sustainable and effective way to sustain ourselves in society?

• 21:56Speaker 02

  I don't know yet. I think it's very hard for us to imagine what that future will look like. So many assumptions will change. And so I think grappling with the questions now and elevating those questions is probably the best we can do. And my view on the nature, on why work matters to people is it matters to people when it is necessary. The nobility of being a farmer when people have to farm in order to eat is extraordinary. You know, the nobility of doing something that becomes unnecessary can be compromised. And we need to figure out how we make sure that people are doing things that feel noble to them. I do think one logical conclusion of this might be what I call the human corner theory, which is that what jobs will human beings do?

• 22:40Speaker 02

  Well, they'll do the jobs when other people want to buy something specifically because it was made by a person. And we have a word for this, which is handmade. And we might imagine a world where, you know, the short story written by a person is intrinsically more valuable to us because it was written by a person than the short story written by a machine. And we live in our human corner, you know, serving one another and having our material needs provided for by a broader economy that is largely run by computers. And I don't know whether that'll be a doomsday scenario or a utopia, but that, I think, is the biggest question that we, as a technology industry, face for the next 20 years.

• 23:16Harry Stebbings?

  And what a brilliant way to conclude our longer format questions. And as you know, we normally now have a quickfire round, but you stated your desire for us to ask questions that are really what we want to know. So we're going to call this round the honesty box round. But first, let's start with one of my favorites and a must for the show, which is your favorite book and why.

• 23:33Speaker 02

  Watershed Down. You know, it's the story of these rabbits who go on a long journey to save themselves. And the reason I love it is, A, It's a tale about unexpected leadership. You know, the leader in the novel is not the strongest, not the smartest, but a person who cares about the right answer and has extraordinary judgment. And because it reminds me of how all these settings that we human beings live in are so different from one another. But you know, sometimes you transpose it into this silly tale about bunnies. And you realize it's all kind of the same and are, you know, we're fundamentally much more similar to each other than we are different. And that to me is a really inspiring message.

• 24:14Harry Stebbings?

  Now, what is it that you most dislike about VCs still to this day?

• 24:18Speaker 02

  Well, first of all, I'll just add, I think there are many wonderful VCs. I just think it's that the typical VC is disastrous. What has yet to change is that we still believe at the very earliest stages that we know more than we can possibly know. We see these as little businesses when I call it quantum startup physics versus Newtonian startup physics. You know, a business that is pre-product market fit. What is it doing? What market is it in? What is it working on? You know, these, we don't have language for these very, very young businesses to understand their complexity and their ambiguity and their unknowability. So I think a little more awe for the unknown.

• 25:01Speaker 02

  You know, for all of the people who invested in one business that only succeeded because it turned into something completely different. You know, I think respect for the unknown is the biggest thing that I think we're, um, we're missing as an industry. And curiosity is the way that I think you build more respect for the unknown.

• 25:15Harry Stebbings?

  And other than Bloomberg Beta, who do you think really does that best?

• 25:19Speaker 02

  So one thing I'd say is there are many VCs who are very private and not doing the blogging, marketing, tweeting game, who have extraordinary curiosity and are wonderful human beings and are just, you know, somewhat less retweeted. And, you know, to me, that quality of being true to yourself and being curious is the thing that produces trust. Because when you fake expertise you don't have, or assume you know something you don't know, you know, that's what puts distance between you and a founder because the founders are the ones who actually have to figure this out. That's the hard work.

• 25:51Harry Stebbings?

  And then what do you do when a founder messes up or doesn't live up to expectations? What's your role then as an investor and maybe lead investor in that situation?

• 25:59Speaker 02

  Yeah, so and we have many cases where we're the only investor in a company. So here's the thing I'd say, the founders are my customers. So it depends when you say messing up. Every founder struggles performance wise with some aspect of building an early stage company. And so our orientation then is to help them and work very hard. And if we thought that there was a better alternative than the founder for the business, we'd tell the person that. And then if they said, no, no, no, thanks, that's nice, that's your opinion, but I'm going to keep going. Generally, we would back them to keep going. Um, in fact, every case, we back them to keep going. Um, and so we're very straightforward about our views, but ultimately the founder is our customer and we work for the founder. There is one other kind of way though, where, which is, you know, when you hire an investor,

• 26:43Speaker 02

  it's a shotgun wedding often because you've known the person maybe for a month or two. You know, ideally for longer, but you know, sometimes a month or two in a funding process is a lot. And then you're married and you can't get divorced or it's very, very difficult to get divorced. And so the bigger issues tend to come up when around trust. And this has only happened one or two times, but where a founder and turned out had misrepresented something to us that was very material, you know, manipulated us. And we're very easy to manipulate because we're so transparent. I mean, when I give people a term sheet, I say, go shop it. It's not going to explode. I only want you to work with us because we're the best choice for you, not because I've put you under some time pressure. Um, and I think that you can end up in these situations where, you know, people think

• 27:26Speaker 02

  that they have to play a certain game and then you end up, you know, having a trust break. And so the trust breaks are very difficult for us. And candidly, we're still trying to figure out what you do in those situations. Fortunately, they've been so few and far between that the vast majority of experiences we've had with founders are just making something as brutally difficult. We're behind you, let's help make you better and go. And, you know, we'll roll dice and see if it works. And if it works, we'll all be thrilled. And if it doesn't, we'll help you do the next thing.

• 27:51Harry Stebbings?

  And then two more, your favorite blog or newsletter?

• 27:54Speaker 02

  Media redefined is the thing I read every day. Jason Hirschhorn's newsletter about the media industry, which I read both because of its content and because Jason as a personality is just amazing. I should also disclose we were an investor, but I was reading it for years before we invested. And so, uh, I am, uh, I'm a big fan of that. And then, um, There is a newsletter by Azim Azhar that's called, I think it's called The Exponential, and it is all about technologies that are the exponential view, all about technologies that are having exponential effects and obviously focuses on machine intelligence. And it's just, he finds these things that I don't know where he finds, they're just terrific and he's got a warm tenor to him.

• 28:37Speaker 02

  You know, this whole idea of human beings helping us to distill and simplify the world is, you know, despite all my earlier comments about machine intelligence, is a really powerful and wonderful thing.

• 28:46Harry Stebbings?

  And then finally, your most recent investment and why you said yes.

• 28:50Speaker 02

  So, our most recent investment is not disclosed. You know, the reason I said yes, so this is actually very easy because we went through, we saw some Twitter conversation about, oh, checklist of the things you invest in. And then we realized checklists are actually not the right way to think about it because it's not like, oh, if you have all these 10 things, we'll invest. And so, we went through the effort, you can see it on our website, of actually writing down our decision criteria for investing. And what we look for is, there are a bunch of things that are tick boxes, you know, do we trust the founders, are the deal terms fair, you know, is it in scope for us? But then really what you're looking for is one reason to believe the startup could potentially be an outlier. And sometimes it's the founder's experience, sometimes it's early engagement data.

• 29:32Speaker 02

  In the case of this last investment, it was just an insight on a media market that was so different from anything I'd ever heard that I just thought that, we got to see how that turns out. I almost just wanted to invest just to see the next, turn the next page in the book.

• 29:45Harry Stebbings?

  Well, Roy, thank you so much for joining us on the show today. It's been-

• 29:48Speaker 02

  Thank you, it's great that you're doing-

• 29:49Harry Stebbings?

  I was very excited for the show and it's absolutely lived up to our expectations. So, thank you so much again.

• 29:54Speaker 02

  There we go.

• 29:55Harry Stebbings?

  Thanks so much, Roy.

• 29:56Speaker 02

  I'll take that as if I can do that once this week, I'm happy. So, thank you for that, I'll talk to you later.

• 30:02Harry Stebbings?

  What a fantastic deep dive that was on all things AI. And if you have not had a chance to read Siobhan's piece on Medium, then check it out by searching for Siobhan Zilis. And if you are loving the show, then you can follow me on Snapchat, at H Stebbings, that's Stebbings with two Bs, where you'll have the ability to send me questions that you'd like to ask guests on the show. But before we leave you today- I'd like to thank Leesa. Despite the show only being twenty minutes, it takes a huge amount of work and none of that would be possible without a good night's sleep, which has made all the more effective since I got my Leesa mattress. And Leesa mattresses really are crafted with the user-first approach, with three unique layers including two inches of memory foam and two inches of latex light foam that ensures the user stays cool throughout the night.

• 30:47Harry Stebbings?

  And not only is the Leesa mattress itself optimized for the user, but so is the buying experience with a completely online experience, meaning you you don't have to engage in that terrible mattress showroom buying Bonanza. And not only do you avoid that, but they also deliver to your door to anyone in the UK, US and Canada. And for every ten they sell, they give one to a shelter. Awesome stuff. And to get yours, all you have to do is head over to leesa.com/vc. That's L E E S A dot com forward slash V C and enter the promo code VC seventy five to get seventy five dollars off your order. As always, I'm so grateful for all your support and I cannot wait to bring you Friday's episode with Jonathan Abrams of Nuzzle and Friendster.