“You should read things that actually make you excited and make you want to engage with the world. For that, what you choose almost doesn’t matter, but you should choose something that allows you to create your own worldview.”
– Rohit Krishnan
About Rohit Krishnan
Rohit is Investment Director at leading global venture capital firm Unbound, focusing on software and fintech. He was previously VP at Eight Roads Ventures and leader of McKinsey’s growth tech practice. He is author of the Strange Loop Canon on Substack.
What you will learn
- How to keep across useful information as a venture capitalist or investor (01:33)
- What are the best practices to start your information day (03:49)
- How to deal with disconfirming evidence and avoiding confirmation bias (08:33)
- Are routine and structure important to thriving on overload? (10:54)
- Why we should use explicit structure mental models more frequently (17:22)
- Why searching for dynamic loops is crucial to business (21:28)
- What are some information tracks that aspiring venture capitalists should follow? (24:26)
- Why having a worldview is the first step to thriving on overload (25:53)
Rohit Krishnan: Absolute pleasure. Thanks for having me, Ross.
Ross: You are an investor, I think it’s fair to call the thinker as well. But as an investor, you have so many opportunities in such a fast-changing world. How is it that you keep across all of the change and all the things you need to be aware of?
Rohit: It’s a great question. I would say that the type of investor that I am, I work in Venture Capital, which means primarily, I work with companies that are at rather early in the stage of development, Series-A, Series-B sort of companies. In many cases, what you’re trying to buy into is a little bit of what the vision of the future is likely to be for those companies. Two things are important that they’re slightly contradictory or orthogonal, I would say, thing number one is that you need to have a pretty open mind to see all of the things that are going on in the world around you, technological development, social development, consumer developments, business developments, etc. so that you can look at them and see where the puck is going to as the phrase goes.
The second is that you need to try and dive usually rather deep into individual niches so you can use those macro themes and understand how those impact specific points like open source software, banking infrastructure, or what have you. The way that I’ve at least approached some of these conversations is to mainly focus on things that I’m most interested in, which perhaps is the only enduring way to do this, quite frankly. Because initially, at least I had a very focused or frameworked point of view, where you would start from, this is what the world is going to look like, therefore, these are the sectors that are going to grow, therefore, these are the technologies, etc. But what I realized is that that form of thinking is like you trying to analyze and come up with every startup’s mission statement from the top down, which is an incredibly hard job. It’s much easier to focus on things that you’re most curious and passionate about, things where you spend your time regardless because that’s the only thing that can give you a little bit of an edge.
Ross: Why don’t we go down into the weeds now and maybe we can come back to some of those themes? What are your information habits? What’s the first information you touch on in the morning? How do you touch it? How do you start your information day? Where do you go? How do you do that?
Rohit: It’s a good question. It keeps changing because I feel the information habits that I have are rarely the same information habits that stay over a period of time. If we were having this conversation last year, I would probably have said, books that I’m reading, articles that I have bookmarked that I’m going back to, maybe some podcasts that I’m listening to. Recently, I’ve been doing a lot more of that through Twitter. I find it both delightfully serendipitous, and a wonderfully random collection of things that comes across which is great because I need a little bit of randomness in my feed.
Currently, I think a large portion of my information diet is self-selected books that I’m reading. At any point, there are several books on Kindle, that I’m at least reading, going back and forth, some fiction, some nonfiction that I like. Any article that I like, I bookmark on to Pocket because that just helps me go back to it when I get the time. Those come from a variety of normal websites and things that I track. I have a bookmarks folder that I go through or have built over several years at this point, from aggregators like arts and literature daily to anything else occasionally, the usual New Yorker-type stuff or other long forms. Changes over the last year, a large number of sub stacks have been added to it, as you would imagine which feed into the information diet.
I used to read a lot more news, now I read it very less. It’s only for the main headline, so to speak, occasional scanning, but I do read a lot more in-depth pieces, essay-type pieces, and opinion pieces to get to it. Occasionally I listen to podcasts, not too many, just because I just don’t have enough time to listen to that many podcasts. I read the transcripts of several podcasts where the ones that come out because it’s just much faster and easier, depending on which ones are most interesting to me. It’s a little bit of a combination, and which part of this I spend more time on depends at any point in where I want to focus.
If at times, I feel like I have been over-indexing on reading about tech news, for example, then I just reduce that for some time and over-index on something else that is valuable. Some of these things are easier because I have to keep up with tech stuff for work anyway which means even if I reduce it, I will have enough of a buffer that I know what’s going on. Whereas some of the other ones like which books that I’m reading are what are the items or ideas or sectors that I’m most interested in, or segments of the world that I want to learn more about or history that I’m looking at reading. Those are usually a function of what I’m most interested in at any given point in time. That was a bit of a rambling answer but hopefully, it gives you the main components in there.
Ross: Yes, the thing is we have lots of information sources. You in your role need lots of information sources. Listeners are interested on what is the scope of that.
Rohit: Yes, I was just going to say about the variety of information sources, at least one of the things I found is that I try and look at it a little bit like applying the Shannon entropy model. I want the next piece of information that I’m reading to surprise me as much as possible. Because otherwise, there are 100,000 articles say roughly similar stuff and have slight variants of the same-ish pieces of information, which don’t help me that much, neither for the job nor personal satisfaction.
If I’m reading, here are the five trends that are going to change the future, and the first trend is AI, I know what the rest of it is going to be like, there’s no point in me reading that. If I’m reading something highly specific about the nitty gritty weeds or a scientific paper about how transformer models are changing, that could be helpful but there’s a point at which there are diminishing returns to me learning more about it. You gotta ride that curiosity wave so your information density is high enough and ideally, you want to keep getting surprised by any new piece of information that you’re bumping up against. If it’s not surprising, it better at least be entertaining, because if it’s neither of those things, you should probably put it aside and move on.
Ross: Yes, looking for surprises is critical. The more we can sensitize ourselves to things that are surprising, one of the other advantages as well is that it helps us to be more familiar with our potential confirmation biases. If we say oh, that’s surprising because it’s disconfirming evidence, that really should be paid more attention to.
Rohit: No, I completely agree. One of the things that I’ve been thinking more about recently, at least is that it’s very common, I would say, to think about pieces of information that come at you, you analyze it, and then you fight against the confirmation bias things, and especially to then judge the article that you read, or the book that you read, or the blog that you read, and say oh, yes, that was pointless. We do this on an ongoing basis so that we can curate information sources. Something that I force myself to do is that, if it is surprising to me, or if it provides a new and interesting lens to view the world with, I think of that as time well spent, regardless of whether or not I believe in its conclusion, which I find to be reasonably helpful as a habit.
It happens all the time, when you’re reading about tech, it happens a lot, if you’re reading about anything in social sciences, it happens a hell of a lot, where you might accept some of the premises you like the argumentation that they have taken or the syllogisms that they put in or the rhetoric, but you might not agree with the conclusion for a variety of reasons. It’s a change that I’ve been trying to do over for a few years. I think I’m getting better at it. It helps make me far less annoyed at the stuff that I read on the internet, which is one of the main goals, isn’t it?
Ross: One of the formative experiences of my life, I was Global Director of Capital Markets at Thomson Financial, very long time ago, and one of the favorite analysts, the salespeople go out, the client says, I love this particular analyst, I disagree with everything he says but he makes me think. That’s valuable.
Rohit: That’s the best thing that you want in an information source or in any person, in most conversations for that matter. It has to be either surprising and informative, or entertaining; ideally, both.
Ross: Yes. Do you have any schedules for your scanning or spending deeper time simulating information? Or is there a lack of structure to that?
Rohit: No, is the answer. Most of it is because I’ve worked very badly in putting schedules on myself unless it’s superimposed by someone else. I’m very good at doing what I want to do because whatever is at the top of the priority list gets done if it’s stuff that I like whether it’s reading a book, writing an essay, or whatever. I’m very bad at saying things like 9:30 to 10:30 is the time for me to listen to X, or do X because I just find it hard to switch my brain into that rhythm. I treat my information diet roughly the same way that I treat my normal diet, where if I don’t want to eat cakes, pastries, and chips all the time, the only way that I can solve that, or the best way that I’ve found to solve that is just to not have it around me, and just make me interested in eating something else, which is what I focus on, on the information stuff as well. Try not to go into silly rabbit holes, and focus on the stuff that is interesting to you at any given point in time.
I will say something that has significantly changed that is I have two young kids, and especially after my second son was born, the number of hours I have as free time drops off precipitously as one would imagine, as a consequence, you’re forced to become far more productive in the hours that you have left, which means that there is an automatic forcing function at any given point to say, Hey, do you want to watch that Netflix show? Or read this book? Or do you want to be doing something else? It’s an implicit calculation that goes on, which is far more pleasing to me than a schedule.
Ross: We all have to find what works for us. Certainly, it sounds like it works for you.
Rohit: Yes, I’ve tried to make chaos my friend, I find it to be the one thing that works reasonably well for me.
Ross: Probably the first thing that made me interested in your work was you do have a channel, Strange Loop Canon.
Ross: I’m a lifelong fan of Gödel, Escher, Bach, and Douglas Hofstadter’s work and the concept of strange loops and their implications. There’s a bit of a broader piece around mental models. Perhaps in whatever way you’d like, I like to say, what is it that draws you to this concept of strange loops? What relevance does it have to how you build your mental models?
Rohit: I first read Gödel, Escher, Bach when I was very early in college. It’s one of those books that just blows you out of the water in the sense that it fits together so many disparate streams of human knowledge and tries to link them together in a pleasing narrative. It’s phenomenal. It’s one of the few books that at least during that time, I read and reread, and I just loved, and it stayed with me for a long period of time, which is why several years later when I wanted to write about at least a broad swath of topics around tech, culture, economics, industry, I wanted to use something that at least was very helpful to me.
You asked about mental models, two of the things that attract me to the idea of strange loops in the first place is that strange loops are self-referential, which I think we all are. That’s one of the core properties of consciousness, it was what Hofstadter wrote about. I find that idea to be interesting, and fascinating, because if you think about the economy overall, as a complex adaptive system, then strange loops effectively are all around us, things that we do impact the output that we want, which impacts the next input that we want to give to a particular system, and this goes on and on forever. That, to me is incredibly interesting and a useful way at least to look at the world and the different ways in which causal arrows intersect. That’s one part of the answer.
The second part on mental models, I would say, I’m a little bit more like” Mungerian” (refering to Charlie Munger-ed.) in my thoughts. We talked earlier about if you’re reading an article or a book, and it helps you think in a particular fashion, that’s its true gift as opposed to its conclusion, or, as opposed to its facts. Mental models are basically just that, it is a way for you to think ideally relatively quickly, but even otherwise, about topics that you don’t usually have huge amounts of domain expertise in, or huge amounts of knowledge in that you haven’t spent 30 years owning, you can use your mental models to use a series of them, ideally, to figure out does this thing even make sense, which is what its primary use is.
I think of the combination of these two as trying to create a little bit like an incredulity filter, most things we see, and we are like, Oh, my God, how can the world work like that? Or how can this particular thing work like that? The mental models have been at least putting several pieces of thought pathways that at least you can look at and say, Ah, I see there is one pathway by which this particular line of argumentation makes a lot of sense, or there is one pathway by which this decision makes a little bit of sense. That’s at least to me, a bigger point about mental models. It’s not about learning frameworks, it’s about developing frameworks repeatedly so that you have a shorthand and still remember what you thought about before? Because you don’t want to start thinking everything from the first principal onwards every single time, that’d be quite inefficient, I would imagine.
Ross: Part of the base of my question is using the concept of reflexivity in mental models of thinking about the world, so part of that can be expressed in systems dynamics, systems modeling, and so on. Do you use systems modeling or systems thinking tools explicitly?
Rohit: For work, explicitly, almost never, but implicitly, quite a lot. Let me rephrase that. Explicitly, I do do it when I want to try to understand something better. If you want to understand a little bit more about the economy, inflation is a big topic, you want to understand more about inflation, it’s helpful, at least it was for me when I was learning about it to draw different boxes about whether it’s consumer demand, consumer demand for multiple types of topics, these come from individual firms, the firms make products, to do that, you have to make an investment, to make the investment, you have to get financing, to get the financing, you have to get a particular rate, there’s a demand for money, you draw this out, and you link the arrows between these pieces, it helps you create a better model of what exactly is going on in the world, which is really helpful.
That’s one way this explicit part of it is helpful. I feel people underuse it a lot. I genuinely don’t find enough people using it in their life, they want to find out how something works. But then, once you have a few pieces, once you’ve done this a little bit in some areas, at least, if you want to employ it in your work, it is much harder to do that explicitly. Just because these are about you learning to ask the right questions as opposed to trying to get a quantitative answer at the end of it.
When I was trying to learn how to do investing in startups, you could do all of this frameworky analysis, but that’s not going to help you. I’ll give you an example. When I first started, I had a spreadsheet, I still do, of 20 different metrics that I thought were most important. Some were qualitative, like how good is the management team, and some were quantitative, like how fast they’re growing, because I was like, I’m just going to pop it there, I have some sense of which ones I think are important but if I look back a couple of years, a few years, I will at least know the companies that I thought were going to succeed, did they succeed or did they fail? And use that as an answer. A few years in, a lot of the things that I thought were important were just like, it didn’t matter. A lot of the things I didn’t think were important, did matter. But the process of actually doing this explicitly helps hone that.
When somebody is new in the industry, you can have these conversations like, oh, yes, I want to look at that company and I think they are interesting because their product is interesting. However, they’re not growing fast enough. At the same time, their competitors are producing these different things that could create a compliment, and they’re burning too much. But at the same time, you might look at a different company that is growing slower but has a higher burn but might be more interesting, there is no if loop that allows you to solve this, because your mental models then come in and say that in this particular situation, you should overweight product, or in this particular situation, you should overweight go-to-market or you should overweight what the broader market is going to do to you. That’s the learning that you need to get to, and that’s the ideal state you get to, or rather, you continue to try to improve on as you want to get to a decision-making process.
Explicit structuring is highly useful and dramatically underused in learning how to think about new pieces of topics. Part of the reason is that doing it just feels like a little bit of a waste when you’re doing it just because there’s no reason for you to be doing it. Do you know what I mean? Nobody’s asking you for it. It’s not a deliverable. It doesn’t help you say things to someone else very easily so you got to do it for yourself a little bit, at least in large parts of our economy anyway.
Ross: I do think it would be valuable for people to be thinking more systemically. One example is that you’ve got your business model canvas and a whole bunch of other similar tools, none of which have any reference to any self-reinforcing loops in them. We do have Jim Collins’ Flywheel and Kees van der Heijden business idea from wayback, which were implicitly looking at what are the self-reinforcing loops that underlie our business model? But I think that’s a massive gap in how most people are explicitly thinking about business models that they just don’t look at what are the self-reinforcing loops in there.
Rohit: Yes, I’d like to pick on something you said. A large part of the conversation is about business models, and how to understand them at least from an education point of view is highly static stuff. You write down a comprehensive list of things that you can think about in all different boxes, it becomes a box-filling exercise very quickly. But which boxes are important is far more interesting than did you fill the box in the first place? That is a much harder thing to make people understand. This is one of those areas where if you’ve done it a lot, there is tacit learning that you get in identifying which parts of it are most interesting, or most important, which is what the training process is for. Part of that is identifying the actual dynamic loops that you talked about.
Yes, the product is not great today, but look at the velocity by which they’re improving so if we wait a couple of years, they are going to be good, and if they managed to do this much with the okay product, think what they can do with something great, which is an example of a line of thought, that is very hard to just make someone drawing a diagram, but you see it a few times, suddenly, the pattern recognition helps. To your point, the dynamic loops are super important and it’s also important to use… I was going to say stories or narratives, but that’s perhaps not the right word, but to really try to verbalize what that future state is likely to be and how you’re going to get from here to there. Because in describing that or in drawing that out, you’re forced to define some of those loops in the first place. Because it is very difficult for you to describe a situation a year or two years down the line when companies are doing X without having some sense of how they get to that X from where they are today.
Ross: Yes. There are many investors and people that stopped seeing that would be well served by a little bit more exposure to some of these ideas of dynamic loops and their impact and thinking through those, bringing those into the explicit and implicit models of startups.
Rohit: I fully agree with that.
Ross: To round out, I’m sure many people look to you for advice on all sorts of issues, based on your success and your experience, but to just tap your quick advice, somebody who’s an aspiring VC, and tech VC in particular, you’ve got to follow all of these emerging trends and technology, there are all of the overlays in the economy, as well as look at all of the emerging sectors, markets, competitions, there’s a lot, so what’s your advice on how someone should keep on top of that, and thrive?
Rohit: I’m going to make a meta point first, which is that my biggest piece of advice, when people ask me this question, is to ask the question back about do you want to be a VC? I’m not sure that enough people understand what the job is, or understand why they want to do it beyond a general feeling that it is an interesting job that they would like. It’s okay, but I just want to make that meta point. Because it’s linked to my core piece of advice on getting the right amounts of information in doing it, which is that coming back to the original point, focus on what you’re most interested in, what you’re most curious about, and lean into that.
The big difference that I find in good VCs, or good investors, good anything, is that they like what they do. I don’t mean this in jobs, your passion kind of sense, I mean it in the sense that they would be people who would be reading up about fintech anyway, as opposed to having to read about fintech for the job. That distinction is crucial. Because ultimately, a large part of at least my job is about reading all of these, a bunch of different sources, and trying to know what’s going on in the world and talking to a large number of entrepreneurs and other VCs to figure out what they think is going on in the world and how their company is doing. You’re constantly synthesizing information that is coming your way because you have a worldview.
I’ll give you an example. I ask this in interviews and stuff, a sector thesis, what do you think is interesting? What do you think is likely to happen? It doesn’t matter what the answer is, but they need to have a point of view. That point of view will evolve. You talk to 200 companies and people, it will change and evolve, but you’re only really going to evolve it if you have a worldview first, which only comes if you’re interested in it first.
Beyond that, in terms of information intake, there are the normal sources, the TechCrunch, Crunchbase type of stuff, where you get the immediate information about fundraisings and stuff like that’s going on in the world, broader world stuff about trends, etc., it’s probably extraordinarily well covered, but I would say there’s a vast variety of Substacks that you can read, that actually helps in that, whether it’s broad stuff like bakkies are boring, or The Generalist, or any number of these. In your information curiosity diet, you should have at least some of this how do you think about the world kind of stuff in there. That is very much idiosyncratic. You should read things that actually make you excited about the world and think about the world and make you want to engage with the world. For that, what you choose almost doesn’t matter, but you should choose something that allows you to create your own worldview.
Ross: That’s fantastic. Hopefully, we’ll get more VCs that follow your advice. Thank you so much for your time Rohit. Fabulous conversation
Rohit: Indeed. This was wonderful. Thanks for having me on, Ross, and congratulations on the book.
Ross: Thank you. Take care, bye.