The a16z Show - a16z Podcast: Seeing into the Future -- Making Decisions, Telling Stories

Episode Date: September 8, 2018

with Steven Johnson (@stevenbjohnson), Chris Dixon (@cdixon), and Sonal Chokshi (@smc90) There's a lot of research and writing out there on "thinking fast" -- the short-term, gut, instinctual decisions we make, biases we have, and heuristics we use -- but what about for "thinking slow" -- the long-term decisions we make that both take longer to deliberate and have longer spans of impact on our lives... and the world? Because we're not only talking about decisions like who to marry (or whether to move) here; we're also talking about decisions that impact future generations in ways we as a species never considered (or could consider) before. But... why bother, if these decisions are so complex, with competing value systems, countless interacting variables, and unforeseeable second- and third-order effects? We can't predict the future, so why try? Well, while there's no crystal ball that allows you to see clearly into the future, we can certainly try to ensure better outcomes than merely flipping a coin, argues author Steven B. Johnson in his new book, Farsighted: How We Make the Decisions That Matter Most. Especially because the hardest choices are the most consequential, he observes, yet we know so little about how to get them right. So in this episode of the a16z Podcast, Johnson shares with a16z crypto general partner Chris Dixon and a16z's Sonal Chokshi specific strategies -- beyond good old-fashioned pro/con lists and post-mortems -- for modeling the deliberative tactics of expert decision-makers (and not just oil-company scenario planners, but also storytellers). The decisions we're talking about here aren't just about individual lives and businesses -- whether launching a new product feature or deciding where to innovate next -- they're also about even bigger and bolder things like how to fix the internet, or what message to send aliens with outcomes spanning centuries far into the future. But that's where the power of story comes in again. The content provided here is for informational purposes only, and does not constitute an offer or solicitation to purchase any investment solution or a recommendation to buy or sell a security; nor it is to be taken as legal, business, investment, or tax advice. In fact, none of the information in this or other content on a16zcrypto.com should be relied on in any manner as advice. Please see https://a16zcrypto.com/disclosures/ for further information. Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Hi, everyone. Welcome to the A6 and Z podcast. I'm Sonal, and I'm here today with Chris Dixon, a general partner on A6 and Z Crypto, and Stephen B. Johnson, who is the author of many books, including where good ideas come from, the PBS series, How We Got to Now, a book on Play called Wonderland, and his latest book is Farsighted, which is how we make the decisions that matter the most. So welcome. Thank you for having me. Could you start just telling us a little bit about the book? Yeah, this is a book that has been a long time in the making, which is appropriate for a book about long-term decision-making. It had a long incubation period. One of the things that occurred to me that got me interested in this topic is that there had been a lot of material written, both in terms of academic studies, but also in terms of kind of popular books. But a disproportionate amount of that was focused on people making gut decisions or instinctual decisions. Just like thinking fast and slow. But also Blink is like that.

It is amazing the amount of processing and all the heuristics we have for making short-term and sensual decisions. But the decisions that really matter the most are slow decisions, are decisions that have a much longer, both time span in terms of how much time you spend deliberating them and then also the time span of their consequences. And I got interested in what the kind of the science is and some of the art in a way. behind those kinds of decisions. Actually, the book partially starts with the great excerpt in Charles Darwin's diaries where he's

trying to decide whether to get married. And it's a beautiful list where he's like, okay, against getting married, I'll give up the clever conversation with men in clubs. My favorite against marriage was less money for books, et cetera. Yeah, yeah. And so he has this list. And, you know, looking at it, it's kind of comical and sweet. in some ways. But that technique of creating a pros and cons list, basically, that was state of the art

in 1837, 1838, and it's still kind of state of the art for most people. That's the one tool they have for making a complicated decision. And actually, we have a lot more tools, and we have a lot more insight about how to make these things. It seems like there's two questions, right? There's a descriptive and the normative question, kind of like, how do people make these decisions and how, or societies or, you know, governments or whoever the actor might be. And then there's a second question of how one should make these decisions, right? I got more and more interested in the second question, right? Like, what are the tools that you can really use to do this in your life?

Can you get better at it? Yeah. And it's a tricky one. I was really grappling, trying to take very seriously, the legitimate objection to a book like this, which is that it is in the nature of complex life decisions, career decisions, should I get married decisions, should I take this job decisions, that each one is unique. right? That's what makes them hard, is that they're made up of all these multiple variables and competing value systems and stuff like that. And it turns out really that a lot of the

science of this and the kind of practice of making a deliberative decision is a set of tricks to get your mind to see the problem or the crossroads or whatever you want to call it in all of its complexity and to not just reduce it down to a series of predictable patterns or cliches or stereotypes. And that's where actually the advice, I think, think is useful. And so that's like the scenario planning where where there's sort of a discipline around what's the upside case, the middle case, the frameworks for forcing yourself to kind of mentally traverse different future paths. Yeah, exactly. Well, one of the big themes of the book that runs throughout it in lots of different ways is the importance of storytelling.

Yeah. And in all these different ways, scenario planning is one example. So, and that's usually used in a kind of business context, right? So you're like, okay, we're trying to decide should we start this, launch this new product. Let's generate some scenario plans for what the market is going to do over the next five years. But let's generate multiple ones. Let's not just predict. I had a friend once who worked at a large oil company in their scenario planning group. And, you know, at first it doesn't sound like that interesting.

But it turns out these large oil companies, like whether oil is $30 or $100, you know, it's a lot of money's at stake. And so they had this infrastructure, like thousands of people. It was like the State Department or something, you know. It was like, it was quite. fascinating to hear about and like what if there's a war in this area and oil drops this much and what do we do and like just the level of rigor i never imagined it was as complex as you know as sophisticated as it was well i had some great conversations over the years with with peter schwartz who's here in the bay area and and uh he was one of the pioneers of scenario planning and one model that that he talks about is you do three different narratives one where things get better one where things get worse and one where things get weird I've heard that. Yeah, I love that because I think that all of us kind of intuitively build the, like,

it gets better, it gets worse kind of scenario plan in our head. It's useful to actually walk through it and do it and tell that story. But the weird one is what's cool, because then you're like, what would be the really surprising thing? Well, the funny thing, at least, if you look at history, the weird is often the case. That's well, we're living through it right now, that's for sure. And I think it, and the key part of it is that the predictions don't even have to be right. on some level for it to be a useful exercise, because a lot of this is about recognizing the uncertainty that's involved in any of these kinds of choices. It's creating a mindset that's open

to unpredictable events. So going through narratives where you imagine alternatives, even if they don't actually turn out to be the case, they get you in this state so that when you do encounter an unpredictable future, whatever it happens to be, you're more prepared for it, or you've thought about at least some of those variables. But the other thing I was just going to say, on the storytelling front, one of the places where it kind of came together, there's a lot in the book about collective decisions, like, what do we do about climate change, or what do we do about the potential threat from superintelligence and AI, right?

Something that we think about a lot here. Global, multi-generational type of thing. Yeah. Super long-term decision making, right? And one of the points that I tried to make in the book is, while we have this cliche about our society that we live in this short attention span world and we can't think beyond 140 characters and all that stuff. The fact that we are actively making decisions that involve changes to the environment that might not happen for another 20 or 30 years and we're thinking about what the

planet might look like in 100 years is something that people have not really done before. They've built institutions designed to last for longer periods or they've built pyramids designed to last. But they weren't very good at thinking about, you know, we're doing these things now. What will be the consequences 80 years from now from these choices we're making now? Regardless of what you think about whether we're doing enough from climate change now, the very fact that it's a central political topic is not, that was not the case 100 years ago. It's a sign of progress. And super intelligence is even a better example of it, I think, because the fact that we're having a debate about a problem that is not at all a problem for us now, but that potentially might be a problem in

50 years. That is a skill that human beings didn't used to have. And we have now, when I was talking about this once with Kevin Kelly out here, another Bay Area person, he had this great point where she's like, this is why science fiction is such an important kind of cognitive tool because you run these alternate scenarios of the future and they help us kind of imagine what direction we should be steering in, even if they're made up stories. Don't people actually say that science fiction is the only way to, quote, predict the future

in terms of what you can actually think of for very complex technologies. I feel like I've heard a statistic or an observation to that effect. I mean, I certainly think that you would find more things that ended up happening in fictional accounts than, you know, official people making predictions about the future outside of a fictional context. Yeah. My bias has always been towards history, for example. Like, the only way you're ever going to possibly get a lens on how to predict the future is to read a lot of history and understand how these things work because of such complex systems that you're not going to, you know, have empirical data and polling and everything else to analyze this stuff. I wonder if to what extent our ways

of thinking about these things in academic literature and things like this have been shaped by the kind of the, you know, when you require everything be testable, you also dramatically narrow. The things that can be tested. Yeah. Right. Or the things that can be tested is a subset of the things that are interesting and worth exploring in the world. And you get steered towards those things. I made this decision with my wife to move to Northern California, having lived in Brooklyn, in New York for a long time. And, you know, when you think about a choice like that, there are so many different variables. There are variables about the economics of it.

They're the kids' schools. Do you want to live in a city? Do you want to live near nature? I mean, all these different things. It's incredibly common- On all the second order of things you could never predict. Right. And what will the consequences of the-

serendipitous meeting your kid has that changes in their life? You know, particularly with children, you know you are changing the overall arc of your kid's life by making a choice like that. And that's scary. But to your point, that kind of decision, well, certainly I would say it's one of the most important decisions that I ever really thought about and kind of worked through with my wife. How would you study that in the lab? Right. It's very hard to be like, okay, everybody, we've got 10 of you that are going to move.

Right. And there's another 10 of you that. And there's no double-blind study you could do. You mentioned the book simulations. We have actually some investments in this area. But like the idea that computing is getting powerful enough that you could, you could ask questions like, we want to fix the New York subways and we want to shut down these subways. How does that have, what are all the consequences of that? Or we change interest. You know, there's always been this, I don't know, the Santa Fe kind of. Yeah, yeah. Complexity theory simulation. I think it's still kind of this fringe. I always think about, I have friends who did machine learning in the 80s. And back then it was this kind of rebel fringe group in AI, right? So mainstream AI back then was heuristics-based.

It's like, okay, we're going to win all these things by, you know, literally putting in these rules and teaching computers common sense. And there was this kind of rebel group that said, that will never work. You need to use statistical methods and have the machine learn. Now, fast forward today, like machine learning and AI are synonymous, right? Yeah. It feels like simulations today are this kind of fringe group. Over time, like, it just seems like a far better way to test these really complex things.

Like, what if you could run a simulation? I don't know if you could run a simulation for moving to California, but you could run a simulation for changing interest rates or for closing down a bridge. Those things, I think, are fairly limited today. You could imagine them getting orders of magnitude more come more sophisticated, right? There's so many, so many things to say to that. So the first is it actually gets back to that classic book that David Glearnter wrote in the 70s or 80s. Mirror World.

And that was all about, you know. Oh, really? On that theme post after that. He's one of my dear favorite people. You know, I read that book when I was, I guess, in. just in grad school. It was one of the first tech books where I was like, oh, this is really fascinating. In some ways, my first book was shaped by that.

Mark Andreessen also says there's a huge influence on it. Yeah, yeah. And so we will, I think that is something that's coming. So we should explain. So mirror worlds, the idea is that, as I recall, you kind of have the whole world instrumented with IoT devices and things. And you have, then the mirror world is the computer representation of that. And the two can interact and really interesting.

Yeah. So basically you have all the, every single object in the, in, let's say we're talking about a city, you know, is somehow reporting data on all of its different states. And then, and the computer is just some massive supercomputer, although it was supercomputer in his day. Now it might just be like an iPhone somewhere. He, by the way, today argues it's just streams of information. Right. Yeah, yeah.

What was that thing? It was like life streams or something? Yeah, yeah. He had a live streaming thing, but now he thinks about in the context of streams as like browsers, Twitter, like sources of information that we constantly live in. So you basically have. you know, software that's looking at all that information. And then the idea would be that it would develop enough of a kind of an intelligence that you could say, okay, given the patterns you've seen over the last 10 years with all these different data points,

if we close that bridge or if we, you know, switch this one neighborhood over to commercial development, what would it look like? Press fast forward, it becomes a kind of SimCity kind of simulation, but based on actual data that's coming from the real city. And it was just one of those ideas. I think there's a whole generation of folks who are. I always think of Ender's game in the whole scene where he essentially is playing a simulation. Then you realize in the end, I mean, I'm sure this book's been out for years. Spoiler alert.

Right. But that it's actually the real war that he's fighting in the final simulation. So the other thing about simulations is a big theme of the book. It's one of those kind of ways in which the book connects to storytelling as well because I think the personal version of this for the, should I marry this person or should I move to California. This is actually what novels do, right? And that we don't have the luxury of simulating an alternate version of our lives because we can't do that yet. We probably won't be able to do that for long time, particularly the kind of emotional complexity of choosing to marry someone or something like that. But we do spend, in an order amount of time, reading fictional narratives of other people's lives.

And the idea is that that's part of the almost like evolutionary role of narrative is to run these parallel simulations of other people's lives. Right. And by having that practice. of seeing, oh, it played out this way with this person's life, this way with this other person's lives, and the novel's ability to take you into this psychological kind of gap of what's going on in a person's mind. A great biography will do that too. So reading history, as you said, is a part of that. But it's, in fact, the first draft of this book had just like a ridiculous amount of middle march in it.

Oh, you still had a lot of middlemarch in it for the record. But it was right up front in the first draft. And I think my editor was like, this is great. But like, I don't know if this is what people need. It's interesting how we spend so much time either kind of daydreaming about future events or reading fiction or watching fiction on TV. We spend so much time immersed in things that are not, by definition, not true. They haven't happened or they haven't happened yet. And I think the reason we do that is because there is an incredible adaptive value in running those simulations in our heads because then it prepares us for our experience in the real world.

we're building kind of the emotional like logic space or something and sort of, I don't know, expanding. I always think of that cool. Like I always get this feeling when I read a good book. It's I think someone said it makes the world feel larger, right? And I think it's another way of saying it kind of expands the, you know, the possible like trees of possibility. Yeah. It's like your mental sample space. Yeah.

You just feel like the world is bigger, right? You read history and you just feel like it's big. Or you read a novel and you feel like the emotional world is bigger, right? And there's just sort of more possibilities. and it's interesting. So you're saying that's sort of almost like an evolutionary need to do that to sort of adapt to be more emotionally sophisticated. There's a great essay by Tubi and Cosmades, I believe names are pronounced, about the kind of evolutionary function of storytelling. And one of the things that they talk about is precisely at this point that we spend an ordinate amount of time thinking about things that are not true.

and that would seem to be actually a waste of time. But in fact, there's a whole range of different ways in which things are not true. There's the she said it was true, but it's not true. Or like, this might happen and thus might be true, but it's not true now. Or, you know, I wish this were true. And our brain is incredibly good at bouncing back and forth between all those kind of hypotheticals and half-truths. And I don't mean this in a kind of fake news kind of way. Like, this is actually a really good skill.

the ability to conjure up things that have not happened yet, but that might, is one of the things that human beings do better than any other stages on the planet as far as we know. Right. And also to do it in a, like Aristotle said, the point of tragedy was that you could experience it with an emotional distance.

Yeah. Right. So you can go, that's the other value of narrative, right, is you can go and you can experience and look at the logic without, so you can go and think about tragedy and how to deal with it without actually being overwhelmed by the emotion of it, right?

And so you're involved, but not so involved that you can't sort of parse it and understand it, right? And that's a great point. And the other thing I just last point on simulations, we were talking about how it's hard to simulate these types of decisions in the lab. But the one place in which we actually have seen a lot of good research into how to successfully make complex, deliberative decisions is another kind of simulation, which is mock trials and jury decisions, right? And that gets you into group decisions, which of course is a really important thing, particularly in the business world.

So like what are the key, I guess, components both to the group composition and also to the process to determine, you know, to get to the right answer? So the biggest one, which is something that's true of innovation as well, not just decision making, is, you know, diversity. It's the classic slogan of like diversity trumps ability, which is you take groups of high IQ individuals who are all from the same. say, academic background or economic background and have them

make a complicated group decision. And then you take a group of actually lower IQ people but who come

from diverse fields, professions, fields of expertise or economic fields, whatever, cultural background, that group will

outperform the allegedly smarter group. Is that because the more diverse group will traverse more future

paths of the tree, of the tree of possibilities? So the assumption was always

the diverse group just brings more perspectives to the table, right? So they have different, you know, it's a complicated, multi-variable problem. That's the way of the earlier framework. Is that like good, bad, weird? Like, they'll just simply bring up and explore more possibilities because of their more diverse experiences. There's no doubt that it's part of it, right? When you're dealing, what makes a complex decision complex is that it has multiple variables

operating on kind of different scales or different, you know, and it's a convergence or different things. But you're saying it's more nuanced than that. It also turns out that just. Just the presence of difference in a group makes the kind of initial kind of insiders more open to new ideas. If you have kind of an insider group, a homogeneous group, and you bring in folks who bring some kind of difference, even if they don't say anything, the insider group gets more kind of original. They challenge their assumptions internally more. So there are exercises you can do to bring out the kind of hidden knowledge that the diverse group,

has. The technical term for it is hidden profiles. And so when you put a bunch of people together and they're trying to solve a problem, come with the decision, there's a body of kind of shared knowledge that the group has. This is the pool of things that everybody knows about this decision that's obvious. For the group to be effective, you've got to get the hidden pieces of information that only one member knows, but that add to the puzzle, right? And for some reason, psychologically, when you put groups together, they tend to just talk about the shared stuff. Like there's a human that, you know, kind of desire to kind of be like, well, we all agree on this. And so some of the exercises and practices that people talk about are trying to expose that hidden information. And one of

them is to just assign people roles and say, you are the expert on this, you are the expert on this, you're the expert in this. Just arbitrarily. So they say, okay, my job is to go and be the expert on this and therefore I'll more likely surface hidden knowledge. Yeah, it diversifies the actual information that's shared, not just like the profile of people. I have a question about this because I found that fascinating that you can. and essentially define expertise as a way to go against this problem of seeking common ground. But then later, you talk about this difference between the classic phrase of foxes and hedgehogs and how actually it's not hedgehogs that are deep experts in a single thing that perform well in those scenarios,

but foxes that are more diverse in their expertise. So I couldn't reconcile those two pieces of information. It's a great question. So just to clarify it, so comes out of this famous study that Philip Tetlock did. And he wrote super forecasting. Yeah, yeah. And expert political judgment. And he did one of the most amazing kind of long-term studies of people making predictions about things.

And it turned out kind of famously that all the experts are like worse than a dart-throwing chimp at predicting the future. And the more famous you got, the worse you were at kind of figure. But he did find the subset of people who were pretty good, you know, significantly better than average at predicting kind of long-term events, which of course is incredibly important for making decisions. because you're thinking about what's going to happen. You can't make the choice if you don't have a forecast of some kind. And what he found with those people, he described them in the classic Fox versus Hedgehog, which is, you know, that Hedgehog knows one big thing, has one big ideology, one big explanation for the world. The Fox knows many little things.

And it's a kind of monolithic thinker, but has lots of distributed knowledge. And so the reason why that I think is in sync with what we were talking about before is in that situation, you're talking about individuals. So it's a Fox and a hedgehog. And what the Fox does is simulate a diverse group, right? He or she has a lot of different eclectic interests. And so inside his or her head. They're like 10 people in their head.

That's one of the reasons why a lot of the people who really are able to have these big breakthrough ideas, one of their defining characteristics is that they have a lot of hobbies. Oh, that's so true. I used to give the tours at Xerox Park for our. all the visitors. And actually one of the big talking points was when we had like these big muckety mucks coming through was how like there'd be a material science expert and he'd be an ex, the world's expert in like goat raising. Or there'd be someone else who's the father of information theory for computers and he's like a world class surfer. Yeah. They all had one specific like music,

whatever. Yeah. There's a funny connection actually to Wonderland, my last book, which is all about the importance of play and driving innovation. And so much of kind of hobby work is people at play. And also has a post on, I mean, a classic post on this on like the things that the smartest people do on the weekend is what the rest of the world will be doing 10 years later. I remember reading that. Yeah, I mean, the way I was thinking about it is there's so many things in life, especially the workplace, are governed over. You basically have a one to two year horizon, right? Yeah. So like, and that's particularly because business people, almost by definition, right, if you're working at a public company or something, they're moving by quarter by year.

And so where are the places in the world where you actually, people actually, smart people have a 10 year plus horizon, right? And it's like probably academia? And then my model would be sort of technical people on the weekends, right? Nights and weekends, right? Like this is, I think there's more than, it's more than a coincidence that's so many of these, you know, Wozniak and jobs and just was a whole bunch of the internet, early internet, and all these other things started off as these like homebrew clubs and weekend clubs and things like that, right? Because it's just simply time horizon, right?

I mean, I think it relates to your book, but like, yeah. So much of what we've done, what we do in the business world and just the whole kind of system, right, is structured around a relatively short time horizon. I think about it in terms of like what we do in our job. Like one of our big advantages, right, is the fact that we are able to take a longer-term perspective just based on like where our capital comes from and all the other kinds of things. And that just lets you do invest in a whole bunch of things that you just, other people just simply can't because they're under a different set of incentives. Well, that's, I mean, one of the, one of the great things that I got out of actually deciding to move to California is spending a bunch of time with the folks at the Long Now Foundation, you know, which you're really trying to encourage not, it's not 10 years. It's, you know, a thousand years, you know, whatever.

It's a 10,000 year class. Basically to be as long as to last as long in the future as civilization is old. I tell us some people about that. That's incredibly idiotic waste of time. Why would you want to look like? There's so many of the problems we have now come from not having taken that kind of time horizon. And in fact, one of the other rifts in the book, I started thinking about like, okay, if we are now capable of thinking on longer time scales, if we're thinking about climate change on a 100-year scale, if we're thinking about superintelligence on a 50 or 100-year scale. Like, what's the longest decision that one could contemplate?

And actually, Sandra Rose, who runs the interval for long now. He heard me talking about this, and he said, oh, we're working on this project with this group called Medi, which is a group that is debating whether or two and what they should, if they decide to, send as a targeted message to planets that are likely to support life. Now we've identified these plants, whatever. And it's similar to superintelligence, and it's a surprisingly controversial project. And there are a bunch of people, including the late Stephen Hawking, who think it's a terrible idea. And if you've read the three-body problem, it's the worst idea ever. Exactly, yeah, three-body problem. And I'm sure, all your listeners have read that is.

It just provokes them? By definition, they are going to be more advanced than we are, which is a whole complicated reason why that is. But they will be. And in course of human history, every encounter between a more advanced civilization, and a less advanced civilization is, and this is, by the way, rooted in the Drake equation and in the dark forest analogy. Well, yeah, and the dark forest idea, right, is that then, therefore, the best strategy is to be silent. You hunt silently, or you don't hunt.

And that's the end of the, was it a Fermi's paradox because of everybody. Fairme's paradox, exactly. It brings all these concepts together. What I just love about it is just because of the speed of light and the distance you have to travel to these planets, this is a decision that by definition, can't have a consequence for at least, you know, 5,000 to 50,000 years, depending on the planet you're targeting, maybe 100,000 years. And so the idea that humans are walking about be like, all right, I think we're going to decide to communicate with these aliens living on this other planet, and we'll get the results back in 100,000 years.

Just the fact that we're capable of thinking that is pretty amazing. You know, I find something kind of self-indulgent, not self-indulgent about, but something that I think is very confusing about making decisions in this framework is that, you know, we can't predict 10,000 years ahead. But nor can we predict immediate second and third order effects of things we build today. So my question is, I mean, this sounds like a terrible question asked given the book is about making better decisions. But why bother making a good decision? Why don't we just sort of let it work itself out in a series of complex little tiny events? You're saying why bother because there's, you can't do it. You can't predict the future? I mean, we don't know how things are going to play out.

Well, the question is, can you get better at it? I think that was the thing that, I think that's one of thinks it's important about Tetlock's work, which is that first book was about people being comically bad at it. But he did carve out this zone. Some people actually have a strategy that works and seems to be better than just flipping a coin or, you know, just making it up. And so I think that, you know, it's not, there's definitely not a crystal ball for this and there's not an applied strategy that works in all situations. But I do think you can kind of nudge it. And because decisions are, I mean, that is kind of the definition of wisdom is like you make the right choices in life, right?

I have a question, too. So we talked a little bit about the fox and the hedgehogs. One of the things you mentioned in your book is the role of extreme perspectives versus mainstream. And I thought that'd be really interesting because we think about that a lot, like where ideas come from on the fringes. Well, it all kind of revolves in the story about the high line in New York, right? The now iconic park that was an old abandoned rail line.

One of the West Side Highway. Yeah, one of the great urban parks created in the 21st century. And for, you know, 20 years, it was an abandoned rail line. and I saw a public nuisance and so on. And so one thing that the book argues is there's a stage in decision making in the early stage, which one should consciously kind of seek out to do, which is to diversify your options, right? And folks who've looked at one of the key predictors of a failed decision is it was a whether or not decision. There was just one alternative, like, should we do this or not?

In a company. In a company. Yeah. But I think it applies to a lot of things. when you just have one option on the table, those decisions are more likely to end up in a kind of failure of one form or another. So part of the strategies that you should, when you're at that early stage, let's do this versus this versus this, multiply your options.

In the case of the high line, for 20 years, the debate about the high line was, basically, should we tear it down or not? And it was really even agree that we should tear it down was just who's going to pay for it. It was like it's a rail line that's nobody using. Industrial rail is not coming back to downtown Manhattan, whatever. And so it was just stuck in this kind of weather or not form. And then this interesting bunch of folks who, to your kind of point about extreme positions, who were not part of the official decision-making process of what to do.

That was a city. It was a debate between the rail lines and, you know, and stuff. But then you had an artist and a photographer and a writer who had kind of gotten attached to this. idea that maybe you could do something with this space. And it was this kind of marginal set of folks who were not part of the official conversation about what to do with this, who added a second option or, you know, said, listen, what if we kept it and turned it into a park? That would be amazing. Because our politics are so contentious and polarized, there's this kind of default, you know, anti-extremism now. Like we want to get out, you know, we get rid of

this extremists. But in a society, there's a certain level of extremism that's really important. So sometimes ideas that are important and that need to happen come into the mainstream from the margins. So it's trying to get what I call it's the optimal extremism. Like how do you, and it's a tricky one. I don't have actually a clear recipe for this, but it's, I think, when you're making a decision, are you, you know, are you bringing in those fringe voices to at least have a seat at the table? Yeah, one thing I, so relating to the internet, like one thing I think it's so potentially great about the internet is you have all of these niche communities, you know, subredits and, you know, crowdfunding. Like, you know, we were investors in Oculus. And I don't think Oculus would have ever gotten initially funded had it not been for the crowdfunding.

I mean, there's obviously been, you know, bad things on the internet as well. But I think for the most part, I believe, has allowed some of these kinds of more interesting and potentially positive fringe groups to get together, whatever. that will continue, you know, as the internet has become more and more centralized and as the topic that we both have talked about before. You wrote a really interesting article for the New York Times last year about something I spent a lot of time on. It was that kind of adaptation of your work, actually. Well, that's awesome. It's actually great to hear. Much better version of it.

So, you know, I think this, the issue we're just talking about of sort of the centralization of the Internet and how do we make sure that the Internet stays interesting and, diverse and I think good for small businesses and creators and all sorts of other people right and it is an issue that I think a bunch of people are talking about right I mean you see it discussed and

when people talk about these issues like demonization de-platforming you see people talk about it in terms of regulation should these platforms be more regulated are we headed to an internet that's similar to TV where you have like four channels that control everything you know Google Facebook Amazon etc and then you wrote about

There's this kind of, you know, fringe movement that is trying to kind of through technology principles and innovations create alternatives infrastructure. Yeah, there was a direct connection actually between Farsighted this book and that piece for the Times Magazine. And really the thing that began at all was Walter Isaacson wrote an op-ed, I think, in the Atlantic saying the Internet is broken, you know, and we need to fix it. It has these problems. And he kind of listed a bunch of problems, which I thought were reasonable. And so I sent him a note and I said, you know, I liked what you wrote. How would we go about fixing it? Like what would be the decision-making body that would decide these are the fixes and we're going to apply them. And he wrote back and he said, you're right, it would be impossible in this polarized age. You know, we can't do it. And I thought, that's incredibly depressing. Right. That's not a good answer. If we're just stuck with the infrastructure we have, then really that's really depressing, right? So I slowly kind of dug. through the writing about it.

And about halfway through it, I began to think that some of the blockchain models and some of the token economy stuff that you've written about as a way of creating sustainable business models for open protocols, basically, which is what we really kind of need. I think one of the reasons that piece worked

is that there were a million pieces written about the blockchain, but I didn't actually set out to write a piece about the blockchain. I set out to write a piece about how would we fix this problem? And I got organically led towards the blockchain. Meanwhile, as that was happening,

all the crazy ICO scams were happening. It was like the best and the worst of the online culture exploding all around me. I think Walter, I think I read the same thing, Walter Isaac is the thing. I think he articulates very well the negative side of it. I think the positive side, I would argue two things.

Like one is just the nature, like the architecture, specifically the internet protocol being very present. professionally designed as a dumb layer. That's in a good way, right? So that you can reinvent. The internet is reinvented if the nodes on the internet upgrade themselves, right? Yep. And so I think of internet architecture as the intersection of incentives and technology design, right?

So you have to create better kind of software that runs on those nodes, and then you have to provide the right incentives, right? And one of the fascinating things about the Bitcoin White Paper is it's essentially, you know, eight pages of incentives. And if you do the incentives, right, the internet is able to sort of heal itself or upgrade itself, I should say. change itself. And then the question people are looking at is, can you take those, that interesting incentives design, and can you apply it for things that are more useful than simply solving cryptographic puzzles like Bitcoin, right? And incentivize new behavior. So that's the, the other thing I would think about is so many of the models we use and are hardware-based and including, like, I know, I've read all your books and like the people you talk about, right, just by definition,

are building usually physical things, right? Because that's what they were doing pre- 20 years ago, right? And you think about, like, you know, once you build, like, the combustible engine, you basically built it. I mean, you can improve it. You know, you build a car. You basically build it. Whereas software is fundamentally different.

This is Mark Andreessen point. Software reads the world. Like, he's always talking. He just thinks people fundamentally misunderstand software and keep applying these old, you know, physical kind of models of how innovative. Carlotta Perez and, like, all these, you know, which are great frameworks. But they're all kind of based on how hardware cycles were. Yeah.

I guess the one thing that I would kind of bring out is, that I actually didn't get to in that crypto piece in the Times was the importance of governance structures inside of these crypto protocols and platforms. And there's always been some level of governance involved in software in the sense that you had a corporation or you had a standards body that was deciding what the actual software package should be or what features should be included. But now, really, for the first time, the governance is actually built into the code If you think about decision-making, that is in a sense what, you know, you have governance.

Like we have embedded in this code a set of rules governing like what we collectively are going to decide for the future of this platform. And the fact that that's now being built into the software. Well, the idea, right, is the point of this movement, right, is to decentralize to take the power away from an individual. And therefore, you have to think about, well, then how do these systems upgrade themselves and govern themselves and who gets to decide, who gets a voice and all these. questions, right? Because in the old model, you just said, okay, the CEO. Right now it's like, well, there is no CEO. So how do you figure it out? For masses of people to decide and coordinate activity at unprecedented scale. So this has been great. We've been talking about decision making and how it plays out in crypto and creative innovation. And also then even in personal lives like Darwin or even

novels and literature like Middle March. But what are some concrete takeaways or advice, not just for how to think about decision-making and being far-sighted, but for what both people and companies, big or small, could do. So, for instance, one of my favorite kind of tricks in the book is this thing that Gary Klein came up with, which is a technique also to deal with kind of the dangers of group think and making a, you know, let's say a work decision where you've got your team and you've decided, we are going to launch this project, this product, and we're all really excited about it. And so he created this kind of technique.

which you call as a premortem. And I love this idea. So a postmortem, obviously, the patient is dead. You're trying to figure out what caused the patient's death. A premortem is this idea is going to die a spectacularly horrible death in the future. Tell the story of how that death happened, right? In five years, this will turn out to have been a bad decision. Tell us why.

And that exercise, again, it's like scenario planning. it's a kind of negative scenario planning, even if it ends up not being true, the exercise of forcing your brain to come up with a story. As opposed to just saying, hey, guys, do you see any flaws with this plan? Do you guys do that when you talk through deals?

Yeah, no, so I think a good investor discipline is to do something similar to that. We know where you kind of, and frankly, an entrepreneur, like I think that one of the myths around entrepreneurship is that they're, that they're, I mean, the risk takers, that said entrepreneurs do take risks,

but good entrepreneurs are very good at, doing premortems, ordering the risks, and then systematically trying to mitigate them, right? I mean, now that's not to say that they don't take big risks, but you certainly don't want to take unnecessary risks, right? So I think what your good entrepreneur is doing is constantly thinking about all the different scenarios, how they'll go wrong, you know, kind of rank ordering them, taking a bunch of risks, but saying, hey, the big, you know, so my key risk and this is, you know, sort of like this type of business is all going to be about financing risk, and this one will all be about, How will it go wrong and you see it enough of it?

And of course, it's a very rough and imperfect science. But it feels like it seems like you get better over time. Yeah. The original patent that Google filed for the self-driving car projects included in it is this thing they call the bad events table. Basically, it's like at any moment in the car, as the car is driving, it's creating this bad events table. And the bad events are, range from I'm going to, you know, dent the, um, you know, right side mirror by accident, you know, just scraping against this car to, I'm going to collide with these two pedestrians and, you know, they're going to die. And there's like 15 bad

events that can potentially happen given the circumstance on the road. And not only do they kind of list the bad events, but then the software is calculating both likelihood of the event happening and then the magnitude of the risk, right? So two pedestrians die, very high magnitude, but if it's very low probability, you kind of measure it. And I think of that as, in a sense, the car is doing that at the speed of instinct. But in a way, that's a kind of table that would be really nice to put next to a pros and cons table. You know, what are all the terrible things that could happen? Yeah.

And let's rank them with probability and with magnitude and just to see it. I think about this all the time, actually, in terms of how people make pros and cons list and how they're so flat variable-wise. And if you've gone through any statistical training, the first thing you learn in any linear model is how to weight your algorithm and you weight the variables. And I always think about that. Like, well, I'm going to give this like a move to California at 10x weight and my move back from New York. You'll give something else 2x and you multiply all those probabilities and those weights to come up with your decision.

I think that's a very good way of thinking about it. You know, pros and cons tables date back to this famous letter that Ben Franklin writes to Joseph Priestley, who coincidentally was the hero of my book, The Invention of Air, but he's like explaining this technique he has, which is basically a pros and cons list, and he calls it moral algebra. What gets lost in the conventional way that people do pros and cons lists is Franklin had a kind of weighting mechanism

where he basically said, okay, create your list of pros and cons, and then if you find ones that are comparable kind of magnitude on one side and the other, cross them out. We would do it differently now, but it was a way of assessing,

okay, these two things are kind of minor, and I got one on one side, one of the other, so I'm going to cancel them out. That's great. I think some of those exercises are really important. important. I think cultivating a wide range of interest and influences is a really important thing to do, both in terms of innovation and creativity, but also in terms of decision-making. And I think it's very important to stop and say, okay, what would the alternate scenarios be? What if it gets better? What if it gets worse? What if it gets weird? And the other thing about the kind of diversity point, I think that's going to become increasingly important. The diversity is actually going to be also machine intelligence, too, right? And increasingly, part of that,

intellectual cognitive diversity is going to involve machine intelligence. And so it's going to be, you know, not just, you know, making sure you have a physicist and a poet on, you know, in your kind of posse that's helping you make this decision. But we're going to see more and more people making decisions. For instance, you know, there's a lot of interesting research into, in the legal world, bail decisions that normally a judge would make a decision, okay, this person should be let out on bail for this amount or not let out on bail, whatever. And there's some evidence now that machine learning can actually make those decisions more effective. it's not that we want to hand over the process to the machines entirely, but the idea that you would be assisted in making a choice like that, I think is going to be something we'll see more and more of.

I mean, I think we're already seeing hybrids of that play out, like with hedge funds, with quant strategies, et cetera, but you're saying something even more. You're saying it's like a partner in decision-making. Yeah, it's a collaborative model. My friend Ken Goldberg is at Berkeley in the robotics program there. He talks about inclusive intelligence, right? Right.

The idea that it's not just about, you know, just human intelligence versus artificial intelligence, but actually this kind of dialogue that you can have with the machine. You might say, I think I should release this person on, you know, a very low bail. And the machine comes back with, while looking at all comparable case studies, I think he actually, you know, shouldn't be released at all. Right. And at that point, you're like, okay, that's interesting. I'm going to question my assumptions here and think about what I might have missed. And you don't, you might not change your mind.

But having that extra voice in the long run will probably be better for us. Right. It feels like crowd intelligence on a whole massive different scale. Were there any qualities of people that you've seen? One of the things that you put in the book was that one of the key factors is an openness to experience as a real great predictor of very good decision-making prediction, et cetera. And I thought that was fascinating because I thought of immigrants. It's like a defining quality of immigration and what brings people to different places. It's one of the big five personality traits. It's openness to experience.

It's another phrase for curiosity. Gotcha. And I love the word curiosity. But openness to experiences, it's a slightly different way of thinking about it, that you are walking through like looking for, you know, this, I'm open to this thing that I've stumbled across and I want to learn more. And Tetlock's predictors, the super forecasters that we talked about, they were, they had that personality trade in spades in general.

So it's a wonderful thing. And it's related, I think, to another quality, which is empathy, right? Which is also, by the way, one of the very things that fiction helps with. Exactly, exactly. So when you get into the world of kind of personal decision-making, novels, in a sense, train the kind of empathy systems in the brain because you're sitting there, like, projecting your own mind into the mind of another. Listening to their inner monologue, their kind of consciousness in a way that almost no other art form can do as well as a novel can. And so that exercise of just what would that other person think?

What would their response be? In so many decisions that we have to make, you have to run those simulations in your head, right? Because your decisions have consequences to other people's lives. And if you aren't able to make those projections, you're going to be missing some of the key variables. That's great. And then finally, what do you make of all those folks that have like these lists of tips and advice? Like when they think about, like, Jeff Bezos does this. And Elon must do that.

I think you might have written about this in your book about how Jeff Bezos. believes that you should get to 70% certainty. Yeah. I like that technique, which is to say, don't wait for 100% certainty. Because a lot of the challenge with these complex decisions is they're just, you cannot, by definition, be fully certain about it. So the question is, where do you stop the deliberation process? So you don't just freeze and not do anything.

And by measuring your certainty levels over time, taking a step out of the process, say, like, okay, how certain am I really about this? Yeah. I think that's a really good action. So I think those little, you know, I definitely. definitely included them. I tried it with this book to try and hit the sweet spot of like, these are kind of interesting tools that have been useful and that have some science behind them, but also then to just look at the kind of broad history and some of the science about the way that people make decisions.

And somewhere have it kind of be a mix of those two things. I think it's great. And especially because we as Homo sapiens are very unique in being able to actually have the luxury of doing this. Well, thank you, Stephen, for joining the A6 and Z podcast. He is the author of the new book, Just Out, far-sighted, how we make the decisions that matter the most. Thank you. Thank you very much. It was great talking to you, I love it. Thank you.