AI transcript
0:00:03 Hi, everyone. Welcome to the A6NZ Podcast. I’m Sonal.
0:00:07 Today, we have one of our book launch episodes with Safi Bakal, author of “Loone Shots.”
0:00:14 The subtitle is, “Had a Nurture the Crazy Ideas that Win Wars, Cure Diseases, and Transform Industries.”
0:00:20 So joining me to co-host this episode is A6NZ General Partner for Bio, V.J. Punday,
0:00:23 who also happened to be a Miller Fellow with the author back in their grad school days.
0:00:27 Speaking of, Safi has an academic background in theoretical physics,
0:00:33 co-founded a biotech startup developing new drugs for cancer, and then led its IPO and served as its
0:00:37 CEO for over a decade. So he has a really interesting vantage point that unites perspectives from
0:00:43 academia, startups, and big company innovation. By the way, you can find other episodes touching
0:00:47 on some of these themes from different angles, including with former Stanford president John
0:00:52 Hennessey on the transition of ideas from academia to startups, and with CEO of Novartis,
0:00:58 Vashrini Vasan, on R&D and innovation inside Big Pharma, as well as an episode on what disruption
0:01:03 theory is and isn’t by searching for those on our website. But in this interview, we cover
0:01:09 what is a loonshot versus a moonshot, and how do or don’t those connect to the concept of disruption.
0:01:13 And given that those are some buzzy buzzwords, after briefly defining them and considering
0:01:18 how companies can innovate regardless of what we call it, we then go into some deep scientific
0:01:24 concepts and analogies to share a very different framework than in typical management literature
0:01:30 with practical advice for teams and companies of all sizes to nurture new ideas and initiatives.
0:01:36 And we even dismantle some common dogma around culture, product, and market along the way.
0:01:42 But first, Safi begins briefly with an analogy to historical empires and why new ideas didn’t take
0:01:50 off there. For thousands and thousands of years, truth was established by the divine ruler or the
0:01:58 tribal leader. And in the 17th century, plus or minus a few years in Western Europe, this kind of
0:02:06 new idea spread. But instead of asking what was true and false, there was a universal truth that
0:02:13 anybody could figure out through measurement and experiment. And that idea, now called the
0:02:18 scientific method, changed the world more than any other idea. It essentially democratized truth.
0:02:22 If you got rid of all the previous books, you destroyed them, they should all be able to come
0:02:29 back to the extent that they were right. And in some sense, you asked me what a loonshot is,
0:02:34 this idea that there are natural laws in the universe and we can access them through
0:02:39 measurement and experiment. We take that for granted today.
0:02:40 It was really controversial at the time.
0:02:48 Not only was it controversial, it was subversive, it was threatening. It didn’t just suddenly appear,
0:02:53 it was a 2000 year story of it, that same idea appearing and being quashed and appearing and
0:02:59 being quashed and appearing, you know, for hundreds if not thousands of years. So why did that idea
0:03:05 appear there? And in fact, so much of the stuff that we take for granted and fed into that
0:03:11 revolution came from China, Islam and India, the astronomy, the medicine, the mathematics,
0:03:16 which came from algebra and India. Many of the technologies, paper and printing had been
0:03:23 used in China and developed in China and not for 50 years or 100 years, but for a thousand years.
0:03:28 China did a paper printing compass, currencies, scholar elites,
0:03:34 agricultural mining that was unheard of, that was not seen in Europe until many centuries later.
0:03:40 So why Western Europe? What was, and I just kept getting stuck on this question, especially
0:03:46 when you grow up in, you know, a Western culture, you’re told, well, basically there was the Greeks
0:03:50 and nothing. Oh, no, you’re so right. I had the same upbringing. I completely relate to that.
0:03:54 That was like how I’d grown up. It was basically 30 second history of science,
0:03:58 Greeks, nothing new and done. History started and were modern. And maybe Einstein.
0:04:04 But the more I started to think about it and the more I realized like there’s something in the
0:04:11 question of why did these big empires did so well for so long, but why do they have trouble
0:04:17 innovating? And I realized there was a close parallel to my industry. So I was in the drug
0:04:23 discovery industry. So the parallel, China, India and the Islamic empires were a lot like
0:04:28 the Merck, Pfizer, Novartis of today. Fascinating. They were global, everywhere, dominant,
0:04:33 they’re phenomenal at franchises. So in India, you’ve had the Taj Mahal. You know, in China,
0:04:38 they built the Great Wall and just incredible, the grand canals. They had a series, sequels.
0:04:44 But the little tiny ideas got quashed. And where did the incredibly innovative revolutionary new
0:04:49 drugs come from? From them teaming market of hundreds and hundreds of small biotechs like the
0:04:53 one that I was at. Essentially, what you’re pointing to is that if you think about how we
0:04:57 parallel that now, biotechs are just great examples of startups in general, you know,
0:05:00 and that startups are maybe the ultimate loon shot, right? Especially the reason why often you
0:05:05 have to do in a startup is that if you try to do it in the big established empire, you know,
0:05:09 that you’d be either going against what the empire does, or it’d be heretical, it’d be crazy.
0:05:15 Right. So a moon shot is a big, exciting goal, like you’re in cancer, eliminating poverty. But
0:05:22 it’s a destination. To go to the moon. Yeah. Like in 1961, when President Kennedy said, let’s put a
0:05:29 man on the moon, that was the original moon shot. But if you rewind the tape, 40 years earlier,
0:05:35 there was a guy named Robert Goddard who explained how he might get there with jet propulsion and
0:05:41 liquid fuel rockets. Kennedy, when he announced his moon shot, was widely applauded. Goddard,
0:05:46 when he explained that crazy idea, might get there was ridiculed. In fact, there was an editorial
0:05:51 in the Times when he was talking about his idea that said, well, obviously this guy who, you know,
0:05:55 has this quote unquote chair in physics at Clark University or wherever, you know, he doesn’t
0:06:01 understand the basic principles of physics that we teach in high school every day. About 49 years
0:06:08 after the Times printed that, the day after the successful Apollo 11 launch to the moon,
0:06:14 the Times printed a retraction and said, apparently rocket flight does not violate the laws of physics
0:06:18 and quote, the Times regrets the error. Such a victory lap. But that doesn’t happen all the time.
0:06:22 The fundamental question for anyone listening to this episode who kind of has a lot of
0:06:28 understanding of the history of innovations is how do you go from being a Goddard to a Kennedy?
0:06:32 There’s a fine balance somewhere in between them. You don’t want to be the person completely
0:06:36 ridiculed. You don’t also want to get there so late that it’s already an accepted idea to succeed.
0:06:41 So when do you know a moon shot becomes a moon shot? Kennedy’s was a moon shot because it worked
0:06:46 too, right? I mean, if NASA was not able to get to the moon, it would have been a moon shot,
0:06:50 presumably, right? On this idea of moon shots, I do want to talk briefly about Google X,
0:06:54 which is in some ways given both a good name and a bad rap to the concept of moon shots in
0:06:58 multiple levels. One, because they talk about what they do as moon shots. Two, because they
0:07:04 actually have a project called loon, which has been mixedly received. And so that’s another context.
0:07:08 But three, I think what is really interesting here is when you think about a moon shot as
0:07:13 this destination, as you’ve described it, a desire to go somewhere. I think a lot of people
0:07:19 have a hard time in organizations in terms of teasing apart. What is a shot that’s worth further
0:07:23 nurturing, regardless of the destination, whether it fails or not? Because I agree,
0:07:27 there’s no shortage of ideas. But it’s precisely because there’s no shortage of ideas that I want
0:07:33 to know when can we pick which ones to invest in, because there’s still limited resources.
0:07:37 Yeah, you’re right. A moon shot is a destination, and we get there by nurturing loon shots,
0:07:40 which are these crazy ideas. And your question is, how do you prioritize?
0:07:43 Exactly. How do you winnow those ideas as well?
0:07:49 I think you want to find the ones that challenge your belief system, because those are the ones
0:07:57 that are most dangerous to you and your business. So the reason you nurture loon shots is when
0:08:03 there’s some new crazy idea that appears that’s going to kill one of your business lines, or take
0:08:08 away your customers, or even open a completely new business line, where would you rather find
0:08:13 out about it? By reading about it in the press release from one of your large customers who’s
0:08:18 going to move all their business away, or would you rather find out about it from one of your
0:08:23 people who presents you one morning with this crazy idea that you were sure was not true.
0:08:29 And all of a sudden, he or she shows you this great proof of concept that there’s some traction
0:08:33 to it. I love that you said proof of concept and traction, actually. I’m so glad you said that,
0:08:38 because when Astro Teller, who heads up GoogleX, wrote an op-ed for me at Wired, he made the argument
0:08:43 that it’s identifying a big problem, articulating a solution that can actually solve the problem
0:08:47 if it existed, even if it does not already exist, which is to your point that it is dangerous and
0:08:52 not out there, but more precisely, when you mention the proof of concept, that there is an
0:08:56 indicator right now that this thing can become a reality. And that, I think, is like the key.
0:09:01 So I spent a while talking about this with Astro, who actually I knew well before GoogleX. And one
0:09:06 of the things that Astro pointed out, which we both agree on, is that using the word moonshot,
0:09:11 unless you really know what you’re doing, and they obviously really know what they’re doing at X,
0:09:18 can be very misleading. And it comes back to why there are a few phrases in the English language
0:09:23 that give me more gas pain than the phrase disruptive innovation. Here’s the problem.
0:09:29 When people talk about disruption, they’re talking about a market. And if you look across history,
0:09:35 if you look across the ideas that change a field of science or a field of business,
0:09:42 or even world history, none of the champions knew anything about where their idea would end up.
0:09:48 I’ll give you an example, the transistor. It was developed at Bell Labs, although it was based on
0:09:52 a lot of federal research, but it was developed at Bell Labs, why they were looking for better
0:09:58 switches. By every definition, it was an incremental innovation, not a quote disruptive.
0:10:02 Once they figured it out, they didn’t know what they could do. It wasn’t usable in the phone
0:10:07 system initially, because it was too expensive and too unreliable. So after a bunch of years of
0:10:10 trying to figure out what they could do with this damn thing, they came up with hearing aids.
0:10:17 So do you think the scientists working on semiconductor junctions in 1946 and 1947
0:10:23 told their bosses, we’re working on disrupting the hearing aid market? The word disruptive
0:10:28 innovation makes sense in hindsight, if you’re a history professor. Right. But not for predicting
0:10:32 and knowing the future. Right. So the reason using the word disruptive innovation or moonshot,
0:10:36 unless you know what you’re doing, can be very misleading, is that it causes you to overlook
0:10:42 the small things that end up having a tremendous difference, because they challenge accepted
0:10:46 belief, but you thought maybe it’ll just be for hearing aids. In fact, it’s not so much an accident,
0:10:53 but it’s this emergent kind of building block that’s created that has so much like second,
0:10:58 third order effects. We had Brian Arthur on this podcast, and he’s like the father of complexity
0:11:02 theory and economics. And one of it’s, he’s the one who’s sort of studied when network effects tip,
0:11:07 when these systems sort of, it’s a very sudden, and they’re always like small, tiny tips. There’s
0:11:13 never some big specific iPhone moment that is a, you know, discrete thing. It’s actually this
0:11:17 continuum of things. What you want to do is you want to fund a portfolio of moonshots,
0:11:23 irrespective of market. Nothing kills a great idea faster, and has been more of a disaster to
0:11:29 innovation, certainly in my industry and drug discovery, than the idea of market projections.
0:11:34 Forget the market protections. Does it challenge accepted beliefs? If so, let’s go try it.
0:11:38 That’s what it’s all about. Well, then let me push some more then, because
0:11:42 then I want to really understand then what distinguishes a good moonshot from a really
0:11:47 bad one. Because when you are a company, you have limited resources. When you’re an entrepreneur,
0:11:52 you have limited time and energy to devote into what product you’re going to build.
0:11:54 And you want to nurture them, but which ones do you nurture?
0:11:57 Exactly. How do you think about that here? I mean, this is your job.
0:12:00 Yeah. So first off, you know, it’s interesting from the point of view of venture capital that
0:12:04 we’re looking for things that have that huge disruptive potential. That’s very much our business.
0:12:10 And it’s usually the things that Ben has, I think it’s been, or Chris Dixon, talk about how often
0:12:13 really great companies look like bad companies. Yeah, they both talk about it and they’re actually
0:12:19 quoting Peter Thiel. Okay. So for example, like, you know, going on the web and meeting strangers
0:12:23 to ride in cars with you, these things at one time seem like crazy. And now it’s a good idea only,
0:12:29 because in part what made it seem like a bad idea was also what opened the door for a team to come
0:12:34 in and have that opportunity. So, you know, for what we’re looking for, especially on the biocide,
0:12:38 I’m looking for things where they’re right at that cusp between where the science has largely
0:12:43 been done, where we’re not investing in science projects, and that it’s about to scale into engineering.
0:12:47 And the small crazy ideas that changed the course of science, business, or history,
0:12:54 they always stumble. I had this story about Jim Black, Mr. James Black, who was a Nobel laureate,
0:12:59 very famous, actually chemist, pharmacologist, originally was advising us. I was saying, God,
0:13:04 he asked me why I was looking kind of down. You know, we were working on this really promising
0:13:07 drug program in the lab, and it just didn’t, you know, we just had these bad results. And he’s
0:13:12 like leaned over, you know, padded me on the knee, and this is like 10 o’clock, a couple whiskies in,
0:13:19 and he’s like, Oh, Sophie, I can’t do a Scottish. Oh, Sophie, it’s not a good drug unless it’s
0:13:23 been killed three times. That’s fabulous. And so I’ve always thought of that as the three deaths
0:13:28 of the luncheon. And that’s actually a little different than some of the philosophy here,
0:13:34 especially in Silicon Valley of, you know, fail fast and pivot. Because in many cases,
0:13:39 the really good ideas that completely transform industries, they fail the first few times.
0:13:43 And it’s a false fail. And it’s a false fail. What do you mean by a false fail? I mean,
0:13:47 I think about in the context of false positives, false negatives. False fail is it’s a failure
0:13:53 that’s due to a flaw in the experiment, rather than a fundamental flaw of the idea. I mean,
0:13:57 just as a simple example, Peter Thiel and Ken Howrie, who I remember Ken Howrie was telling
0:14:02 me this story, you know, a couple years ago, that when they were first looking at Facebook,
0:14:08 everybody was passing on social networks because it was like the 15th or 16th or whatever social
0:14:13 network. And right around that time, Friendster was just sort of going down the toilet. And I
0:14:17 remember what Ken and Peter did was they got access to the, they were like, really, is it
0:14:24 the case? Maybe that’s the case. But why don’t we just probe and get a little more background.
0:14:29 And they asked where they knew the CTO and the team behind that they knew that the website was
0:14:33 having a lot of problems. So they got the user retention data. And they looked at it and they
0:14:38 were like, holy shit, you know, people are staying on this freaking website, even though
0:14:42 it’s crashing, they’re staying on it for hours. And they’re like, wait a minute,
0:14:46 maybe it’s not a bad business model. Maybe actually, it’s a really good business model
0:14:49 and people are leaving because the website sucks. That’s a false fail.
0:14:54 And when the lesson is to false fail within an org, if the false fail takes down your org,
0:14:58 then the other startups that come after you get to benefit from your failure.
0:15:02 But if you can sort of understand the false field internally, that’s your best bet to
0:15:06 keep on moving, right? And I think that these false failures, there really wasn’t a failure.
0:15:09 It was if failures of execution is not a failure at the loonshot.
0:15:11 And I’m curious to get your take on this, Hafi, because you’ve lived this,
0:15:15 is that for a lot of drug design projects, when they get canceled,
0:15:19 Pharma never has the time or the interest. Sometimes it seems to go back and understand why.
0:15:24 I remember the head of R&D of a very large, one of the top few Pharma one time, we were having
0:15:30 dinner, we were discussing doing a large partnership together. And we were talking about
0:15:34 killing projects and your people say, oh, you got to have balls to kill a project. And he said,
0:15:38 killing a project is the easiest thing in the world. It never comes back.
0:15:43 And that’s one of the reasons it’s so difficult to innovate inside, let’s say in my field,
0:15:46 in our field, inside a large drug discovery or research organization,
0:15:50 is because it’s so hard to keep a project going.
0:15:52 That’s fascinating, because I would have argued, I would have thought it’s actually
0:15:57 much harder to kill something than to actually keep something going that shouldn’t be killed.
0:15:59 Now, it’s the easiest thing in the world to kill a project.
0:16:04 It’s much harder when, because of the three deaths of the luncheon, because really good ideas
0:16:08 often seem really stupid in the beginning, as soon as they hit that first stumble,
0:16:12 it’s much harder to justify to your colleagues.
0:16:14 That’s true in a big company. Do you think it’s true in a small company? In a small company,
0:16:19 if this is your thing, killing something is actually the dreaded pivot, right?
0:16:22 And so maybe in a small company, is it less likely that people want to kill it?
0:16:27 Well, this comes down to the incentives. The underlying theme of this book is let’s look at
0:16:34 structure rather than culture. I became a first-time CEO. I think I was 32 or three.
0:16:37 And probably like a lot of first-time CEOs and entrepreneurs,
0:16:42 I read everything I could find about management, how to be a better leader,
0:16:47 how to build a great company, how to empower my employees and get great returns from my
0:16:52 stakeholders and achieve our big mission. And almost everything I read about was about
0:16:59 building a great culture. And as a physicist, that felt kind of squishy to me. And there’s also
0:17:04 that sort of after the fact, let’s say somebody wins a lottery and you say, “Well, that’s
0:17:09 fantastic. What color socks was he wearing? Let’s all go wear.” I did want to see if there was a
0:17:16 little more of a scientific underpinning. Let’s do the behavioral economics, but of groups.
0:17:19 And as soon as you arrange people into groups, what are these sort of hidden incentives? That’s
0:17:24 the kind of thing that I didn’t think about for 10 years and kind of none of the traditional
0:17:28 management literature has really looked at. So it’s interesting you bring up behavioral
0:17:33 economics because prospect theory itself was a very dangerous idea to the establishment at the
0:17:38 time. And yes, there’s a lot of debates right now in the literature about how robust those findings
0:17:43 are so many years later. But that aside, it was a very important contribution. I’m curious for
0:17:49 how that connects into your worldview. Sure. In fact, that’s kind of one of the most fun things
0:17:54 for me because what Kahneman and Firsky did was they found a new Venn diagram intersection of
0:18:00 psychology and economics. And this is a new Venn diagram intersection of physics and economics.
0:18:06 Oh, interesting. I love that. So you simply look at the incentives for individuals whenever you
0:18:12 organize into a group. It’s only relevant inside a group. And then you borrow an idea from physics,
0:18:17 which is a phase transition. How does this phase transition idea come into play? I mean,
0:18:21 you’re a physicist by training. So I get that there’s a physics aspect to this. But how does
0:18:26 this play in the innovation organizational context? Every phase transition in nature is a result of
0:18:32 two competing forces, a tug of war. And the phase transition is as you adjust the balance,
0:18:39 you change certain parameters of design or the environment, you adjust the balance between
0:18:44 those two forces and boom, at some point, they cross and the system snaps. I’ll give you an
0:18:48 example with water. There’s a glass of water right here, which your audience can’t quite see.
0:18:52 I’ll take a photo. But you can imagine. When I stick my finger in the glass of water,
0:18:59 I can swish it around, right? Except as I gradually lower the temperature, all of a sudden,
0:19:04 at a critical temperature, boom, the behavior completely changes.
0:19:12 So when it freezes. Exactly. Why? How did those molecules know to go from sloshing around and
0:19:17 being fluid to being completely rigid? It’s the same molecules. And it’s the exact same molecules.
0:19:23 There’s no CEO molecule with a bullhorn. And that’s sort of the other counterintuitive thing
0:19:28 about this is it doesn’t matter what the CEO does. Okay, it’s not about the culture. Yeah,
0:19:34 it’s about understanding those two forces and what are the balance. So here’s the two forces
0:19:40 in a glass of water. There’s one force, entropy, which wants to makes water molecules run around
0:19:46 and be free. Right. The creative force, one could argue. If one wants to do that, entropy is actually
0:19:53 a hard physical quantity. And it is the case that molecules have two competing forces, one of which
0:19:58 is entropy, which wants to make them run around and be free. And the other is binding energy.
0:20:04 It wants to lock every molecule of water, exactly 2.8 angstroms from its neighbor.
0:20:10 When the temperature is high, entropy wins. Now, as you gradually lower the temperature,
0:20:16 the entropy force gets weaker and weaker and weaker. And the binding energy gets more and more
0:20:22 important. And then what happens? Boom. At 32 Fahrenheit, those two things cross. Below 32,
0:20:28 it becomes totally rigid. So it has nothing to do with the CEO molecule yelling, be rigid, be fluid.
0:20:32 Let’s adopt a fluid culture and all slush around. The CEO can set the temperature.
0:20:37 That’s exactly right. Control parameter is the equivalent of temperature. It controls the
0:20:42 behavior of the system. It turns out when you work out the mathematics of this, of the incentives,
0:20:47 the two forces, there are four control parameters inside organizations, elements of organizational
0:20:53 design. And not the CEO, the executive team, the board has control over, just like you have control
0:20:58 over a glass of water. In classic innovation parlance, it is sort of disruptive versus non-disruptive
0:21:03 core versus non-core, these two forces. However you describe them, the point is that organizations
0:21:08 clearly have these two competing forces. So back to not being, you know, the cultural stuff that
0:21:14 management literature gives us, what structurally needs to happen to make organizations know how
0:21:21 to navigate these two forces? Okay. So when you bring a group of people together and you organize
0:21:27 them and you set a mission and you tie a reward system, incentives tied to that mission, all of
0:21:35 a sudden you create two forces. And here’s what they are. The first one is stake in outcome.
0:21:43 Think of that as equity. So let’s say to stick with our industry, you have a small biotech company
0:21:46 and you’re developing a new cancer drug. Your stake in outcome is your stake in that drug
0:21:52 success. Does it work or not? Now when you’re a small, your stake in outcome is enormous. If the
0:21:58 drug works, everybody’s a hero and a millionaire. If the drug fails, everybody’s unemployed.
0:22:02 Right. What’s the second force? It’s perks of rank.
0:22:08 Interesting. So you can think of this almost as equity in cash or equity in base salary.
0:22:14 When you’re a tiny small company, stakes in outcome are so much bigger than perks of rank.
0:22:21 Now let’s flip it. Let’s say you, Sonal, are a manager at Pfizer. So now let’s look at your
0:22:26 incentives. What’s the stake in outcome? Well, if you make a good drug, let’s say it sells
0:22:31 500 million a year, that’s a pretty good drug. Plus you’re going to be fighting for the next seven
0:22:35 or 10 years. Everybody wants your budget and there’s going to be the three deaths with the
0:22:38 loon shot. You’re going to have to try to defend it. So your stake in outcome is pretty tiny.
0:22:44 What’s the perks of rank? Well, if there’s a committee meeting and you sit around the table
0:22:48 and keep, you know, trashing these loon shots and doing it in a very thoughtful wise, and maybe
0:22:54 you’re very funny and kind of witty, and they’re like, well, this young person, she’s got a great
0:22:59 head. She’s aligned with our vision. Right. What might you get? Promoted. Exactly. And then what
0:23:04 happens when you’re promoted? 30% bump in salary. And not much change in stake in equity, actually.
0:23:11 Exactly. So we just went from small, a small company, stake in outcome was much bigger.
0:23:17 Perks of ranks were irrelevant. To large, where stakes in outcome is very small and perks of
0:23:21 rank. So that’s very much equivalent to the water. These are the dynamic two forces that you’re
0:23:26 describing. Right. And instead of temperature, it’s size. So somewhere between small and large,
0:23:31 as you grow and you grow and you grow, all of a sudden the balance of incentives suddenly shift.
0:23:38 And the same person, the same sonal, would go from pounding the table. This is the most awesome idea
0:23:42 to kind of making fun of it at committee meetings. You know, I think this makes a ton of sense. And
0:23:46 we see it in big companies. And I think the challenge then is, you know, as your company is
0:23:51 growing, what can you do about it? So let me ask you a question. When it snows overnight,
0:23:56 what do you sprinkle on your sidewalk? Salt. Why? Because it blocks the snow from
0:23:59 hardening and freezing over. Exactly. And the reason it does that is because adding salt to
0:24:05 water lowers the freezing temperature. This is a control parameter. It is another. So
0:24:09 the really interesting thing about complex systems, which is what we’re talking about,
0:24:13 whether it’s a glass of water or humans understanding incentives when you bring people
0:24:18 together in a group, is that they have more than one control parameter. There are four other control
0:24:24 parameters. And to what Vijay was saying earlier, that’s what a manager or a leader or an executive
0:24:29 team or a board can control. You may not control size. You may need 100 people. You may need 500
0:24:34 people. You may need 10,000 people. And that, you can’t control, but you can add a little salt.
0:24:37 It’s interesting to think about how far the analogy goes and where it breaks down. So I think one of
0:24:42 the most powerful things about it is the fact that you can take the same person from one face to the
0:24:46 other. I agree. Just like with the molecule. Yeah. So that part is really interesting. And also,
0:24:51 just what you’re talking about, you can imagine doping and putting in certain types into given
0:24:55 phases to change their behavior. And I think we’ve seen that in various groups, right? You can put a
0:24:59 certain type of person to a team. And so some people will change with phases. Some people are
0:25:04 just intrinsically different. I thought where you were going with this was doping in semiconductors,
0:25:07 which actually gave us the transistor. Well, you know, but that doping either one,
0:25:10 you’re talking about alloys or semiconductors is the same concept, right? Actually, it’s exactly
0:25:15 right. If you take an insulator like silicon and you dope it a little bit and then you get a metal
0:25:20 and it’s a phase transition. That’s what Phil Anderson won his Nobel Prize for, is the localization
0:25:26 transition, the metal insulator transition. But to get to your question, yes, this is a phase
0:25:32 transition. That’s kind of the theme of the book. CEOs at the end of the day, they’re sort of lamenting,
0:25:36 oh, no one cares. Well, that’s because they don’t own half the equity of the company like you do.
0:25:41 But when you were a small startup, of course, everybody, their incentives were aligned.
0:25:45 And of course they rushed out because they were going to be unemployed if the loonshot failed.
0:25:49 I find it’s a very hopeful idea, quite honestly, because you’re saying that you can,
0:25:53 you essentially can be a big company that can figure out how to innovate. You can be a small
0:25:57 company that doesn’t innovate, that actually sometimes goes into the frozen zone without
0:26:02 figuring out how to keep the entropy moving and the creativity and the excitement moving,
0:26:08 and that you can more precisely be both at the same time. And you can file up or down various
0:26:13 parameters to make that happen. So I find that’s a very hopeful message. A question I have, and
0:26:18 when you think about the history of innovation, kind of going back full circle to what makes
0:26:23 something a loonshot versus a moonshot. And yes, I heard you, a moonshot’s a destination,
0:26:30 a loonshot could be a false fail along the way. I still want to understand how you know when a
0:26:37 field, an area becomes more ready to be optimized a certain way.
0:26:42 This reminds me a lot of just the general topic of how a biology right now is shifting from
0:26:47 science to engineering. And there’s lots of different trends and reasons why. We talk about
0:26:51 these two different groups, the sort of creative types, the inventors, the artists versus the
0:26:56 the soldiers, the people who get things done. We’re seeing that shift in biology as one can
0:27:01 finally engineer biology in many ways. So in that field then, when you move from say artistry to
0:27:06 soldierry, for lack of a better phrase, the way I’ve heard you describe this is like the bespoke
0:27:11 to the machine automator. How do you in this particular example know? Yeah, I mean, that’s the
0:27:16 billion dollar or trillion dollar question, right? Yeah, literally. And so, you know, what it comes
0:27:21 down to when, at least for me, what I’m looking for is both in Safi’s language that there is a
0:27:27 destination that we see where you would land, and that at each step of the way that there is a way
0:27:31 to do it through engineering like execution, there’s all these different places where bio plays
0:27:38 a role, either with consumer or with healthcare or other industries. And in each one, it’s just not
0:27:42 enough that there’s a technology in our product that there really is probably market fit. And
0:27:46 that’s always unavoidable. And probably market fit is one of the hardest things for anyone to
0:27:52 sort of just theorize, you know, because you can very easily be wrong, or you can be wrong,
0:27:56 could be three years too early. It’s not wrong, it’s when. It’s when. And so, some of that also,
0:28:01 you have to have the intuition that it’s now, you can have theses for why it’s now. In the end,
0:28:06 the market will tell you whether you’re right. Like the classic example, Xerox was a great example
0:28:10 of a disruptor in earlier days with just the copying machine. People are like, “Well, I don’t
0:28:14 need copies. What do I want copies for?” You know, it just wasn’t part of the workflow. You wouldn’t
0:28:18 think to use a machine that doesn’t exist. And that only after people start to understand, “Oh,
0:28:24 actually, it was really useful to have this,” did you create a market. And so, in a sense that
0:28:30 device had product market fit before the market was there. And so, I think, I agree with Sophie in
0:28:34 the sense that just because the market isn’t there now, isn’t a reason to kill it. But obviously,
0:28:38 there has to become a market. Otherwise, there’s no one to buy it.
0:28:42 I would then argue just to reconcile these different views. It basically is how you,
0:28:46 the word market, just might be over-indexed. Because what we’re really talking about is a
0:28:51 pull, a draw, a need. But some might even say a want. Because sometimes you don’t need something.
0:28:55 You might just want something but not know that you want it in the classic Jobsian sense.
0:29:01 Right. In the sense of some, I mean, I can bring it to very nuts and bolts. Someone’s got to pay
0:29:04 for it. And this is one of the mistakes I see over and over again with startup founders is that
0:29:10 when they’re small, it feels like they’re all artists. They do a great job of building an org.
0:29:14 And then suddenly, the org is not what they’re comfortable with or familiar with because they
0:29:18 have built an army. And they actually never want to join the army. Right. They don’t want to be navy,
0:29:22 actually, in that context, right? So, when you think of artists versus soldiers, this idea that
0:29:26 there’s the creatives and the builders and the people have to fight the war on the ground,
0:29:30 another way to think about it is Jobs’ famous quote about pirates versus navy. I mean,
0:29:33 there’s so many different variations of this over time in innovation history.
0:29:37 I think our companies have both too. But hopefully, there’s something for the soldiers to start
0:29:41 working on, you know, to go to war on, even though, you know, you constantly have to continue to
0:29:46 innovate. So you need both. But so when I think about how I’m going to find them, I want to find
0:29:50 things that are just at that transition. If it’s only artists, it’s not going to work. But once
0:29:54 they’re there, we have to think about which ones have the opportunity to become the moonshots
0:29:58 eventually. And which ones also, frankly, that can continue to innovate. And so one of the things
0:30:02 I thought was interesting about the book was that phase transition interplay between the two and
0:30:07 how it’s not a bad thing to have that phase transition. And that was maybe one of the more
0:30:10 surprising things. There are two things that I find very hopeful. I know a lot of people have
0:30:17 found very helpful. One is, and they’re both a little bit against the grain of what you often hear.
0:30:20 I love things that are against the grain. More against the grain, the better.
0:30:25 Especially what I hear from my tech friends in Silicon Valley. So one is don’t hire people from
0:30:31 big companies or big people from big companies are a terrible fit. And I thought that as well.
0:30:34 We’re doing the great stuff because we’re the risk takers. We’re the innovators and all those big
0:30:40 corporate guys, they’re risk averse and that’s why you shouldn’t hire. And then as you grow up,
0:30:43 you start doing these partnerships with large farmers. So you’re spending a lot of time with
0:30:48 these people and you get to know them well. And then these people are just like us. If they’re
0:30:53 advisor and they have no stake in outcome and a ton of stake in getting promoted, what do you think
0:30:57 they’re going to do? They’re going to work on being promoted. And if they’re at a 50-person
0:31:03 startup and being promoted or titled is totally irrelevant, but whether the project works or not,
0:31:06 means whether they’re unemployed or not, what are they going to do? They’re going to pound
0:31:11 the table and save the thing if it stumbles. So it’s about the subtle influence of incentives.
0:31:16 And when you really peel back the layers on that, you tease out all these things that you can do
0:31:21 to balance the incentives better for what you’re trying to achieve. So that’s one hopeful thing.
0:31:25 It’s not really about there, these two kinds of people. You should never hire a big corporation.
0:31:28 It’s the same person in different environments. Exactly. It’s the molecule that lands on the
0:31:34 block of ice versus the molecule that drops in a glass of water. So okay, now you say these
0:31:40 organizations have these phase transitions when they grow, they go from being totally fluid and
0:31:46 embracing wild new ideas to being quite rigid. So what do we do if we are a 500,
0:31:52 a thousand, or 10,000-person company? And so the point is that innovating well is a phase of
0:31:58 organization, just like being solid or liquid is a phase of matter. It’s a property of the whole
0:32:02 that doesn’t depend so much on the details of the parts. That’s what it means in
0:32:06 science or physics of an emergent behavior. And actually, it’s just sort of straightforward
0:32:12 mathematics. And the book that innovating well and a focus on politics or not is a phase of human
0:32:15 organization. I love that you actually have an equation that you outlined in the appendix,
0:32:20 which is awesome. We’ve never seen that in a management book. That took a lot of convincing
0:32:25 with the publisher. That’s another story. It got relegated. To the appendix. We’re way back there.
0:32:33 But anyways, if it’s a phase of organization, there are operationally excellent, high execution,
0:32:39 or you are innovating well. Larger companies today or startups that have grown need to do both.
0:32:43 Yeah. Mark talks about how startups are on the sort of five-year cycle theory. And I think of
0:32:48 this a little bit like growing versus sustaining organizations that even startups, by definition,
0:32:53 after a certain amount of time, have this inevitable gravity that begins to hit because
0:32:56 there are other startups coming out there sort of doing things faster.
0:33:01 Part of it is that also these environments keep on changing as we were talking about. And so
0:33:05 as the environment changes, the question is how we can sort of push back on the environment. And
0:33:09 maybe to abuse the analogy in physics with phase transitions, you can have hysteresis.
0:33:15 Which is… You can push the environment past where something… So there’s a really fun thing
0:33:20 to do. And so kids, you could do this at home. For the beer, put in the freezer. You take it out.
0:33:25 And it looks like water. But it’s already below the freezing temperature. And you just knock it.
0:33:29 And it’ll suddenly, it’ll freeze right in front of you. I’ve never done that. I’ve never got kids out of it.
0:33:34 The knocking kind of catalyzes this. And so this is a historic effect that it really should have been
0:33:39 a solid. But that it was lingering in the old phase for a little bit. What’s the point of that?
0:33:43 So the point is, you may think you have innovation in your org. And you may think that you have the
0:33:49 environment. But it is just a message of the past. And in an instant, you can lose it.
0:33:54 Just like the beer can shift. And so you need to really understand what your environment is.
0:33:59 And as a CEO, as a leader of something, you have to understand your org. You have to look to see,
0:34:04 do you have this phase segregation? Do you have two healthy phases there? The pirates in the navy,
0:34:08 the artists and the soldiers. And then what do you need to do to sort of affect the balance?
0:34:12 And it can be the hysteretic argument is that it can be misleading sometimes, too.
0:34:18 I love that beer example. So it is absolutely true that matter can’t be in two phases at the same time.
0:34:25 You can’t be solid and liquid at the same time. So if complex systems can never be in two phases
0:34:32 at the same time, you can’t be solid and liquid. You can’t be water and ice. How can you do both?
0:34:36 There’s one exception to the rule that you can’t be in two phases at the same time.
0:34:44 And that’s right on the edge of a phase transition. Life at 32 Fahrenheit. What happens is,
0:34:51 matter separates. You get blocks of ice and pools of liquid. So you get phase separation
0:34:56 and they coexist at the same time. The second thing that happens is you get something called
0:35:00 dynamic equilibrium, which the molecules go back and forth and back and forth. Little molecules
0:35:05 swimming in the pool of liquid hits on the block of ice and freezes. A molecule on the surface of
0:35:10 the ice starts shaking loose and then goes into the pool. And it’s back. It’s a continuous cycle.
0:35:14 I mean, if you think about the word dynamic equilibrium itself as a seeming contradiction,
0:35:17 but it’s not in this context because it’s dynamic and in equilibrium at the same time.
0:35:24 Exactly. It’s a balance cycle of exchanging back and forth with neither side dominating the other.
0:35:30 And that’s what happened during World War II. Vannevar Bush came in, told FDR, he said,
0:35:35 “We’re going to lose this war.” At a 10 minute meeting in the summer of 1940, he said, “We are
0:35:40 far behind the technologies that are going to be crucial to winning this war.”
0:35:46 The culture of the military, not only is it resistant to new technologies like any
0:35:50 military culture, but it should be. It is the right culture for their job.
0:35:53 In this case, they’re literally soldiers.
0:35:58 They’re literally soldiers. And he said, “I want you to give me a group of people,
0:36:03 give me the authority to mobilize the nation’s scientists for war and we’ll develop the radical
0:36:08 new technologies and I’ll report just to you.” So he created phase separation within the federal
0:36:13 government. And to this day, that really is the origin of the national research infrastructure
0:36:20 of the United States. The National Science Foundation, the NIH, DARPA, came out of Vannevar
0:36:27 Bush’s idea during the Second World War. But he made sure where most, where this fails so often,
0:36:32 so many years, so many times across companies. I mean, Xerox Park is a great example of that
0:36:37 because it was overly siloed. They missed the second part, which is the dynamic equilibrium.
0:36:39 Constantly moving back and forth.
0:36:44 So here’s the message for entrepreneurs or leaders or managers, which is also a little
0:36:51 different than what you get typically in Silicon Valley and is different than the model of a great
0:36:57 manager, a great leader. And that is, there is this myth that the great manager, the great
0:37:03 leader of a technology company is this product, product, product person and they build their…
0:37:07 Oh, you’re really fighting some major dogma right now. This is going to be blasphemous.
0:37:08 I want to hear it.
0:37:13 They build the company, this great company on the back of their inventions.
0:37:21 So the companies across history that have done that have failed spectacularly. They may survive
0:37:27 for a while when there’s a Moses on the mountain with a staff anointing the next Holy Loonshot.
0:37:32 They may survive for a while, that’s the hysteresis effect,
0:37:36 but eventually they’ll turn. That’s exactly what happened with Polaroid when Edwin Land,
0:37:42 who certainly deserved Nobel Prizes for some of his breakthroughs, stood on top of that mountain
0:37:46 and anointed the next Holy Loonshot and that was the end of Polaroid.
0:37:51 So the difference, the ones who have really succeeded, let’s say Vannevar Bush,
0:37:55 a theater veil at AT&T, the ones who really succeeded didn’t have that mindset.
0:38:01 Vannevar Bush was a brilliant inventor. He invented the first analog computer. He anticipated
0:38:04 much of the personal computer industry and certainly the internet and…
0:38:06 He wrote that memex memo, right? I remember reading it.
0:38:11 The famous essay published in the Atlantic, which inspired Doug Engelbart and a lot of
0:38:16 that traces back to Vannevar Bush, but as he likes to say, “I made no original contribution
0:38:21 to the war. None of my inventions ever amounted to anything.” What he did, rather than lead as a
0:38:28 Moses, he led as a gardener and he saw his job as maintaining life at 32 Fahrenheit.
0:38:32 Setting up the structure, the scaffolding that allowed that dynamic equilibrium,
0:38:38 that living at the edge. He maintained the balance between his group and the military.
0:38:44 And so if there’s a lesson there, it’s not about your ideas. Your job is to maintain life at 32
0:38:51 Fahrenheit. Your job is to maintain the exchange of… Because where innovation fails is not in the
0:38:56 supply of new idea. It’s in the transfer of those new ideas to the field. It’s actually like Xerox,
0:39:01 exact same thing. They were giving these computers to people selling typewriters and what were they
0:39:04 being commissioned on? Typewriters. Exactly. It was a very sales-driven organization.
0:39:10 It was a subtle influence of incentives. Bush recognized that the dynamic equilibrium,
0:39:16 it’s not just transfer one way from the scientist or the artist or the creative engineers to the
0:39:21 salespeople or the soldiers in the field, but the other way. Because most ideas suck when they first
0:39:27 come out of the lab. There’s a huge jump from a technology to a product, like a really beautiful
0:39:30 product. And then even if you have a really beautiful product, you still have to create a
0:39:35 market and sell to sell it. I was outside of my Stanford office after I’ve been at ACSNZ for a
0:39:40 while. And a founder at Stanford came by and he’s like, “You present this whole new tech.” And he’s
0:39:44 like, “Oh my god.” And I was like, “Well, this is really interesting, but I’m not sure you could
0:39:48 build a company.” And he’s like, “You know, the physics was so hard, building a company will
0:39:54 be easy.” But actually, the tech is not enough. And this goes back to Steve Jobs’ story is learn
0:39:59 to love your artists and soldiers equally. So when Steve Jobs started his first time at Apple,
0:40:06 he was exactly like that. I only love the artist. And people who know the history know this quite
0:40:12 well. It was a disaster. So the Lisa that he tried to run was a flop, the Apple 3 that when it was
0:40:16 a flop. And when he developed the Macintosh, although it was a phenomenal ad that goes down in the
0:40:22 history of advertising, it was a disaster as a product. The sales completely plummeted. He
0:40:28 loved his artist and he ridiculed the soldiers. And it caused enormous dysfunction and it almost
0:40:34 made Apple go bankrupt. And then he did this same thing at Next. When his next computer company,
0:40:38 it was product, product, product, it’s all about the product, da, da, da, da, da. And that was,
0:40:42 that product flopped. And then he bought this little company, this Lucas Film Group, because
0:40:46 they had a computer. And he thought this is an even bigger, faster computer, the Pixar Image
0:40:52 Computer called PIC. And that product flopped. But eventually over 10 years, when he came back
0:40:58 to Apple the second time, he had Johnny Eve, that was his artist, and he had Tim Cook. Tim Cook was
0:41:02 known as the Attila the Hun of Inventory before. And the myth around him, oh, is all his ideas,
0:41:07 and he championed the product. That’s not really what happened. In fact, when he did that, it was
0:41:13 a disaster. And only by evolving and learning to love his artists and shoulders equally,
0:41:19 that’s when things really took off. And that’s why Vannevar Bush succeeded. And so many scientists
0:41:24 before him failed, because soldiers love soldiers. Artists love artists. He not only worked well
0:41:29 with the military, he loved them and respected them. And that’s why he was able to bridge the
0:41:33 two. Yeah. I would argue just to Vijay’s point too, though, that this is why there’s such a
0:41:39 resurgence in modern tech management literature and startups in particular, around the important
0:41:43 role of product managers and product management. Because traditionally, there was so much discussion
0:41:49 around engineering and sales and design, all being sort of separated and siloed. But now,
0:41:52 through this, and it’s not a new role, it’s been around for ages. But in a sense, that product
0:41:58 manager role is to connect the engineers to set the temperature. Exactly. There’s sort of like
0:42:02 this person who maintains a dynamic equilibrium between orgs, they move and travel. I actually
0:42:07 think it’s actually one of the most important schools of thought for the future of the firm
0:42:11 in the startup to company environment, which is the evolving role of product management in this.
0:42:16 Because that is the concrete instantiation of exactly what you’re describing, Safi.
0:42:20 There’s all these different threads going on here. Bush and his famous memo right after World War
0:42:26 II define what basic research is. And in Safi’s concept of dynamic equilibrium,
0:42:32 by, I think, inadvertently talking about the separation between science and engineering,
0:42:35 he kind of pulled engineering away from scientists. Oh, that’s fascinating.
0:42:40 Yeah. And that now there was these two different disciplines and science is done by scientists
0:42:45 and engineering is done by engineers. And that actually took a little bit about that
0:42:50 equilibrium that was there our way. And so what we’re seeing now is something different where
0:42:54 scientists are doing translational work, which is another name for engineering.
0:42:59 And in biology in particular, the shift from science to engineering is becoming very powerful.
0:43:03 And one of the things I think is always a topic very germane, I think, to our listeners and to me
0:43:09 is one of the more things that we can do that to practically learn from this. Because I think the
0:43:14 analysis is very compelling. But if you’re a CEO right now, what can you do? I mean, what are the
0:43:19 knobs? And so in the book, you talk about one of the classic things being the magic number.
0:43:24 Basically, a number of people at which beyond that, it’s hard to, you have to start to create
0:43:29 an org and a hierarchical structure too often. And the magic number is canonically 150, but
0:43:35 often I see the magic number being 30 when it’s not built right. And so how can you push it
0:43:38 such that you can try to maintain what we loved about the start from the beginning?
0:43:43 Right. So this idea of a magic number, like the Dunbar number comes from the idea that the number
0:43:48 of neurons in your brain is fixed. So sometimes, as you know, in science or in physics, you might
0:43:53 have a good observation, but the wrong theory. So if you just look at the hidden influence of
0:43:59 incentives out pops a size, but it’s not a fixed number. I have a little fun with it because I
0:44:05 show how if you plug in certain things, it can be 150, but it can be as low as 30 or as high as
0:44:11 5000, because it depends on these four parameters. So when you adjust these four parameters,
0:44:14 you make those numbers bigger. What are the four parameters again? I don’t actually think we listed
0:44:21 them. There’s equity fraction. For example, to what extent is a person being incentivized by
0:44:26 base salary, in which case there’ll be a ton more focus on politics versus to what extent are they
0:44:32 being rewarded on their project? Right. Another parameter is you might call it return on politics.
0:44:41 To what extent is the system’s design so that individuals can lobby their managers for promotion
0:44:46 and that has an impact? Obviously, the more impact lobbying has, the more political and the more
0:44:52 careers matter, but some companies, actually Google does this, McKinsey does this and others,
0:44:56 they take the promotion decisions completely away from the manager. So that’s
0:45:02 another example. Right. And then the third? The third one is project skill fit. So if you’re very
0:45:07 if you’re very good and working more on your project, moves the project, if you’re a coffee
0:45:12 machine designer, you’re good at coffee machine design, then you’re likely to spend more time on
0:45:16 your coffee machine. I love that idea. It’s like founder market fit. It’s like writer topic fit.
0:45:19 It’s this idea that you really match the skills and passion of the person
0:45:23 so well with the thing they’re working on. This is project skill fit. Now, if somebody puts you
0:45:28 on coffee machine design and you’re like me, not a very good designer with very little aesthetic
0:45:33 sense, then it doesn’t matter if you spend an extra hour or 10 hours or 100 hours on your project,
0:45:37 it’s going to be still the same lousy coffee machine. You might as well spend your time on
0:45:43 politics and lobbying. So project skill fit is another parameter. And the fourth parameter?
0:45:49 Management span. So there’s a lot of literature on management span, but they kind of miss the
0:45:55 point. There isn’t a right management span. There’s the right span for your goal. So for example,
0:46:00 if you are building planes, you don’t want to assemble 10 planes and see which eight fall
0:46:06 out of the sky. You don’t want a very innovative risk taking. Right. You don’t want to be doing
0:46:09 that with like autonomous cars, for instance. Right. So there you actually want to narrow
0:46:14 a span because that encourages quality control and redundancy checking because you don’t want
0:46:19 those eight planes. On the other hand, if you want to do a ton of experiments and have groups
0:46:24 work together well, you actually want a very wide span. So there’s no right answer for a company.
0:46:28 And that’s what like all these things that average across the company make no sense at all. Right.
0:46:33 If the group that’s assembling planes, you want one kind of system, one kind of metric,
0:46:38 you want narrow spans, you want high quality controls, the group that’s creating crazy new
0:46:42 technologies for those planes, you want those spans as wide as possible with very different
0:46:50 metrics. So there’s a handful of these parameters and those are what you adjust to tune for the
0:46:55 goals that you want to achieve. Yeah. So many of the world’s problems that can be addressed through
0:46:59 bioengineering, healthcare comes off naturally into many different sub parts about healthcare,
0:47:02 but just the world around us is created by biology, whether we’re talking about
0:47:08 the food we eat, the air we breathe, just about everything. And so that is a huge opportunity
0:47:12 for companies to come in and become huge companies because of the potential of the problems they
0:47:17 can address. And it’s just fascinating to think about how the dynamics of human nature comes into
0:47:22 play here and that there really are these knobs that a CEO can turn. And now I think the question is,
0:47:26 can you really read your organization right and understand what the temperature is?
0:47:30 Finding the thermometer or the equivalent starts to become, I think, really critical.
0:47:33 And if you can find that and know where these knobs are, I think you could
0:47:37 have a huge impact on your work. Yeah, that’s great. Thank you for joining the A6NC podcast.
0:47:40 Thank you so much. Thanks. It was a lot of fun to be here.
0:00:07 Today, we have one of our book launch episodes with Safi Bakal, author of “Loone Shots.”
0:00:14 The subtitle is, “Had a Nurture the Crazy Ideas that Win Wars, Cure Diseases, and Transform Industries.”
0:00:20 So joining me to co-host this episode is A6NZ General Partner for Bio, V.J. Punday,
0:00:23 who also happened to be a Miller Fellow with the author back in their grad school days.
0:00:27 Speaking of, Safi has an academic background in theoretical physics,
0:00:33 co-founded a biotech startup developing new drugs for cancer, and then led its IPO and served as its
0:00:37 CEO for over a decade. So he has a really interesting vantage point that unites perspectives from
0:00:43 academia, startups, and big company innovation. By the way, you can find other episodes touching
0:00:47 on some of these themes from different angles, including with former Stanford president John
0:00:52 Hennessey on the transition of ideas from academia to startups, and with CEO of Novartis,
0:00:58 Vashrini Vasan, on R&D and innovation inside Big Pharma, as well as an episode on what disruption
0:01:03 theory is and isn’t by searching for those on our website. But in this interview, we cover
0:01:09 what is a loonshot versus a moonshot, and how do or don’t those connect to the concept of disruption.
0:01:13 And given that those are some buzzy buzzwords, after briefly defining them and considering
0:01:18 how companies can innovate regardless of what we call it, we then go into some deep scientific
0:01:24 concepts and analogies to share a very different framework than in typical management literature
0:01:30 with practical advice for teams and companies of all sizes to nurture new ideas and initiatives.
0:01:36 And we even dismantle some common dogma around culture, product, and market along the way.
0:01:42 But first, Safi begins briefly with an analogy to historical empires and why new ideas didn’t take
0:01:50 off there. For thousands and thousands of years, truth was established by the divine ruler or the
0:01:58 tribal leader. And in the 17th century, plus or minus a few years in Western Europe, this kind of
0:02:06 new idea spread. But instead of asking what was true and false, there was a universal truth that
0:02:13 anybody could figure out through measurement and experiment. And that idea, now called the
0:02:18 scientific method, changed the world more than any other idea. It essentially democratized truth.
0:02:22 If you got rid of all the previous books, you destroyed them, they should all be able to come
0:02:29 back to the extent that they were right. And in some sense, you asked me what a loonshot is,
0:02:34 this idea that there are natural laws in the universe and we can access them through
0:02:39 measurement and experiment. We take that for granted today.
0:02:40 It was really controversial at the time.
0:02:48 Not only was it controversial, it was subversive, it was threatening. It didn’t just suddenly appear,
0:02:53 it was a 2000 year story of it, that same idea appearing and being quashed and appearing and
0:02:59 being quashed and appearing, you know, for hundreds if not thousands of years. So why did that idea
0:03:05 appear there? And in fact, so much of the stuff that we take for granted and fed into that
0:03:11 revolution came from China, Islam and India, the astronomy, the medicine, the mathematics,
0:03:16 which came from algebra and India. Many of the technologies, paper and printing had been
0:03:23 used in China and developed in China and not for 50 years or 100 years, but for a thousand years.
0:03:28 China did a paper printing compass, currencies, scholar elites,
0:03:34 agricultural mining that was unheard of, that was not seen in Europe until many centuries later.
0:03:40 So why Western Europe? What was, and I just kept getting stuck on this question, especially
0:03:46 when you grow up in, you know, a Western culture, you’re told, well, basically there was the Greeks
0:03:50 and nothing. Oh, no, you’re so right. I had the same upbringing. I completely relate to that.
0:03:54 That was like how I’d grown up. It was basically 30 second history of science,
0:03:58 Greeks, nothing new and done. History started and were modern. And maybe Einstein.
0:04:04 But the more I started to think about it and the more I realized like there’s something in the
0:04:11 question of why did these big empires did so well for so long, but why do they have trouble
0:04:17 innovating? And I realized there was a close parallel to my industry. So I was in the drug
0:04:23 discovery industry. So the parallel, China, India and the Islamic empires were a lot like
0:04:28 the Merck, Pfizer, Novartis of today. Fascinating. They were global, everywhere, dominant,
0:04:33 they’re phenomenal at franchises. So in India, you’ve had the Taj Mahal. You know, in China,
0:04:38 they built the Great Wall and just incredible, the grand canals. They had a series, sequels.
0:04:44 But the little tiny ideas got quashed. And where did the incredibly innovative revolutionary new
0:04:49 drugs come from? From them teaming market of hundreds and hundreds of small biotechs like the
0:04:53 one that I was at. Essentially, what you’re pointing to is that if you think about how we
0:04:57 parallel that now, biotechs are just great examples of startups in general, you know,
0:05:00 and that startups are maybe the ultimate loon shot, right? Especially the reason why often you
0:05:05 have to do in a startup is that if you try to do it in the big established empire, you know,
0:05:09 that you’d be either going against what the empire does, or it’d be heretical, it’d be crazy.
0:05:15 Right. So a moon shot is a big, exciting goal, like you’re in cancer, eliminating poverty. But
0:05:22 it’s a destination. To go to the moon. Yeah. Like in 1961, when President Kennedy said, let’s put a
0:05:29 man on the moon, that was the original moon shot. But if you rewind the tape, 40 years earlier,
0:05:35 there was a guy named Robert Goddard who explained how he might get there with jet propulsion and
0:05:41 liquid fuel rockets. Kennedy, when he announced his moon shot, was widely applauded. Goddard,
0:05:46 when he explained that crazy idea, might get there was ridiculed. In fact, there was an editorial
0:05:51 in the Times when he was talking about his idea that said, well, obviously this guy who, you know,
0:05:55 has this quote unquote chair in physics at Clark University or wherever, you know, he doesn’t
0:06:01 understand the basic principles of physics that we teach in high school every day. About 49 years
0:06:08 after the Times printed that, the day after the successful Apollo 11 launch to the moon,
0:06:14 the Times printed a retraction and said, apparently rocket flight does not violate the laws of physics
0:06:18 and quote, the Times regrets the error. Such a victory lap. But that doesn’t happen all the time.
0:06:22 The fundamental question for anyone listening to this episode who kind of has a lot of
0:06:28 understanding of the history of innovations is how do you go from being a Goddard to a Kennedy?
0:06:32 There’s a fine balance somewhere in between them. You don’t want to be the person completely
0:06:36 ridiculed. You don’t also want to get there so late that it’s already an accepted idea to succeed.
0:06:41 So when do you know a moon shot becomes a moon shot? Kennedy’s was a moon shot because it worked
0:06:46 too, right? I mean, if NASA was not able to get to the moon, it would have been a moon shot,
0:06:50 presumably, right? On this idea of moon shots, I do want to talk briefly about Google X,
0:06:54 which is in some ways given both a good name and a bad rap to the concept of moon shots in
0:06:58 multiple levels. One, because they talk about what they do as moon shots. Two, because they
0:07:04 actually have a project called loon, which has been mixedly received. And so that’s another context.
0:07:08 But three, I think what is really interesting here is when you think about a moon shot as
0:07:13 this destination, as you’ve described it, a desire to go somewhere. I think a lot of people
0:07:19 have a hard time in organizations in terms of teasing apart. What is a shot that’s worth further
0:07:23 nurturing, regardless of the destination, whether it fails or not? Because I agree,
0:07:27 there’s no shortage of ideas. But it’s precisely because there’s no shortage of ideas that I want
0:07:33 to know when can we pick which ones to invest in, because there’s still limited resources.
0:07:37 Yeah, you’re right. A moon shot is a destination, and we get there by nurturing loon shots,
0:07:40 which are these crazy ideas. And your question is, how do you prioritize?
0:07:43 Exactly. How do you winnow those ideas as well?
0:07:49 I think you want to find the ones that challenge your belief system, because those are the ones
0:07:57 that are most dangerous to you and your business. So the reason you nurture loon shots is when
0:08:03 there’s some new crazy idea that appears that’s going to kill one of your business lines, or take
0:08:08 away your customers, or even open a completely new business line, where would you rather find
0:08:13 out about it? By reading about it in the press release from one of your large customers who’s
0:08:18 going to move all their business away, or would you rather find out about it from one of your
0:08:23 people who presents you one morning with this crazy idea that you were sure was not true.
0:08:29 And all of a sudden, he or she shows you this great proof of concept that there’s some traction
0:08:33 to it. I love that you said proof of concept and traction, actually. I’m so glad you said that,
0:08:38 because when Astro Teller, who heads up GoogleX, wrote an op-ed for me at Wired, he made the argument
0:08:43 that it’s identifying a big problem, articulating a solution that can actually solve the problem
0:08:47 if it existed, even if it does not already exist, which is to your point that it is dangerous and
0:08:52 not out there, but more precisely, when you mention the proof of concept, that there is an
0:08:56 indicator right now that this thing can become a reality. And that, I think, is like the key.
0:09:01 So I spent a while talking about this with Astro, who actually I knew well before GoogleX. And one
0:09:06 of the things that Astro pointed out, which we both agree on, is that using the word moonshot,
0:09:11 unless you really know what you’re doing, and they obviously really know what they’re doing at X,
0:09:18 can be very misleading. And it comes back to why there are a few phrases in the English language
0:09:23 that give me more gas pain than the phrase disruptive innovation. Here’s the problem.
0:09:29 When people talk about disruption, they’re talking about a market. And if you look across history,
0:09:35 if you look across the ideas that change a field of science or a field of business,
0:09:42 or even world history, none of the champions knew anything about where their idea would end up.
0:09:48 I’ll give you an example, the transistor. It was developed at Bell Labs, although it was based on
0:09:52 a lot of federal research, but it was developed at Bell Labs, why they were looking for better
0:09:58 switches. By every definition, it was an incremental innovation, not a quote disruptive.
0:10:02 Once they figured it out, they didn’t know what they could do. It wasn’t usable in the phone
0:10:07 system initially, because it was too expensive and too unreliable. So after a bunch of years of
0:10:10 trying to figure out what they could do with this damn thing, they came up with hearing aids.
0:10:17 So do you think the scientists working on semiconductor junctions in 1946 and 1947
0:10:23 told their bosses, we’re working on disrupting the hearing aid market? The word disruptive
0:10:28 innovation makes sense in hindsight, if you’re a history professor. Right. But not for predicting
0:10:32 and knowing the future. Right. So the reason using the word disruptive innovation or moonshot,
0:10:36 unless you know what you’re doing, can be very misleading, is that it causes you to overlook
0:10:42 the small things that end up having a tremendous difference, because they challenge accepted
0:10:46 belief, but you thought maybe it’ll just be for hearing aids. In fact, it’s not so much an accident,
0:10:53 but it’s this emergent kind of building block that’s created that has so much like second,
0:10:58 third order effects. We had Brian Arthur on this podcast, and he’s like the father of complexity
0:11:02 theory and economics. And one of it’s, he’s the one who’s sort of studied when network effects tip,
0:11:07 when these systems sort of, it’s a very sudden, and they’re always like small, tiny tips. There’s
0:11:13 never some big specific iPhone moment that is a, you know, discrete thing. It’s actually this
0:11:17 continuum of things. What you want to do is you want to fund a portfolio of moonshots,
0:11:23 irrespective of market. Nothing kills a great idea faster, and has been more of a disaster to
0:11:29 innovation, certainly in my industry and drug discovery, than the idea of market projections.
0:11:34 Forget the market protections. Does it challenge accepted beliefs? If so, let’s go try it.
0:11:38 That’s what it’s all about. Well, then let me push some more then, because
0:11:42 then I want to really understand then what distinguishes a good moonshot from a really
0:11:47 bad one. Because when you are a company, you have limited resources. When you’re an entrepreneur,
0:11:52 you have limited time and energy to devote into what product you’re going to build.
0:11:54 And you want to nurture them, but which ones do you nurture?
0:11:57 Exactly. How do you think about that here? I mean, this is your job.
0:12:00 Yeah. So first off, you know, it’s interesting from the point of view of venture capital that
0:12:04 we’re looking for things that have that huge disruptive potential. That’s very much our business.
0:12:10 And it’s usually the things that Ben has, I think it’s been, or Chris Dixon, talk about how often
0:12:13 really great companies look like bad companies. Yeah, they both talk about it and they’re actually
0:12:19 quoting Peter Thiel. Okay. So for example, like, you know, going on the web and meeting strangers
0:12:23 to ride in cars with you, these things at one time seem like crazy. And now it’s a good idea only,
0:12:29 because in part what made it seem like a bad idea was also what opened the door for a team to come
0:12:34 in and have that opportunity. So, you know, for what we’re looking for, especially on the biocide,
0:12:38 I’m looking for things where they’re right at that cusp between where the science has largely
0:12:43 been done, where we’re not investing in science projects, and that it’s about to scale into engineering.
0:12:47 And the small crazy ideas that changed the course of science, business, or history,
0:12:54 they always stumble. I had this story about Jim Black, Mr. James Black, who was a Nobel laureate,
0:12:59 very famous, actually chemist, pharmacologist, originally was advising us. I was saying, God,
0:13:04 he asked me why I was looking kind of down. You know, we were working on this really promising
0:13:07 drug program in the lab, and it just didn’t, you know, we just had these bad results. And he’s
0:13:12 like leaned over, you know, padded me on the knee, and this is like 10 o’clock, a couple whiskies in,
0:13:19 and he’s like, Oh, Sophie, I can’t do a Scottish. Oh, Sophie, it’s not a good drug unless it’s
0:13:23 been killed three times. That’s fabulous. And so I’ve always thought of that as the three deaths
0:13:28 of the luncheon. And that’s actually a little different than some of the philosophy here,
0:13:34 especially in Silicon Valley of, you know, fail fast and pivot. Because in many cases,
0:13:39 the really good ideas that completely transform industries, they fail the first few times.
0:13:43 And it’s a false fail. And it’s a false fail. What do you mean by a false fail? I mean,
0:13:47 I think about in the context of false positives, false negatives. False fail is it’s a failure
0:13:53 that’s due to a flaw in the experiment, rather than a fundamental flaw of the idea. I mean,
0:13:57 just as a simple example, Peter Thiel and Ken Howrie, who I remember Ken Howrie was telling
0:14:02 me this story, you know, a couple years ago, that when they were first looking at Facebook,
0:14:08 everybody was passing on social networks because it was like the 15th or 16th or whatever social
0:14:13 network. And right around that time, Friendster was just sort of going down the toilet. And I
0:14:17 remember what Ken and Peter did was they got access to the, they were like, really, is it
0:14:24 the case? Maybe that’s the case. But why don’t we just probe and get a little more background.
0:14:29 And they asked where they knew the CTO and the team behind that they knew that the website was
0:14:33 having a lot of problems. So they got the user retention data. And they looked at it and they
0:14:38 were like, holy shit, you know, people are staying on this freaking website, even though
0:14:42 it’s crashing, they’re staying on it for hours. And they’re like, wait a minute,
0:14:46 maybe it’s not a bad business model. Maybe actually, it’s a really good business model
0:14:49 and people are leaving because the website sucks. That’s a false fail.
0:14:54 And when the lesson is to false fail within an org, if the false fail takes down your org,
0:14:58 then the other startups that come after you get to benefit from your failure.
0:15:02 But if you can sort of understand the false field internally, that’s your best bet to
0:15:06 keep on moving, right? And I think that these false failures, there really wasn’t a failure.
0:15:09 It was if failures of execution is not a failure at the loonshot.
0:15:11 And I’m curious to get your take on this, Hafi, because you’ve lived this,
0:15:15 is that for a lot of drug design projects, when they get canceled,
0:15:19 Pharma never has the time or the interest. Sometimes it seems to go back and understand why.
0:15:24 I remember the head of R&D of a very large, one of the top few Pharma one time, we were having
0:15:30 dinner, we were discussing doing a large partnership together. And we were talking about
0:15:34 killing projects and your people say, oh, you got to have balls to kill a project. And he said,
0:15:38 killing a project is the easiest thing in the world. It never comes back.
0:15:43 And that’s one of the reasons it’s so difficult to innovate inside, let’s say in my field,
0:15:46 in our field, inside a large drug discovery or research organization,
0:15:50 is because it’s so hard to keep a project going.
0:15:52 That’s fascinating, because I would have argued, I would have thought it’s actually
0:15:57 much harder to kill something than to actually keep something going that shouldn’t be killed.
0:15:59 Now, it’s the easiest thing in the world to kill a project.
0:16:04 It’s much harder when, because of the three deaths of the luncheon, because really good ideas
0:16:08 often seem really stupid in the beginning, as soon as they hit that first stumble,
0:16:12 it’s much harder to justify to your colleagues.
0:16:14 That’s true in a big company. Do you think it’s true in a small company? In a small company,
0:16:19 if this is your thing, killing something is actually the dreaded pivot, right?
0:16:22 And so maybe in a small company, is it less likely that people want to kill it?
0:16:27 Well, this comes down to the incentives. The underlying theme of this book is let’s look at
0:16:34 structure rather than culture. I became a first-time CEO. I think I was 32 or three.
0:16:37 And probably like a lot of first-time CEOs and entrepreneurs,
0:16:42 I read everything I could find about management, how to be a better leader,
0:16:47 how to build a great company, how to empower my employees and get great returns from my
0:16:52 stakeholders and achieve our big mission. And almost everything I read about was about
0:16:59 building a great culture. And as a physicist, that felt kind of squishy to me. And there’s also
0:17:04 that sort of after the fact, let’s say somebody wins a lottery and you say, “Well, that’s
0:17:09 fantastic. What color socks was he wearing? Let’s all go wear.” I did want to see if there was a
0:17:16 little more of a scientific underpinning. Let’s do the behavioral economics, but of groups.
0:17:19 And as soon as you arrange people into groups, what are these sort of hidden incentives? That’s
0:17:24 the kind of thing that I didn’t think about for 10 years and kind of none of the traditional
0:17:28 management literature has really looked at. So it’s interesting you bring up behavioral
0:17:33 economics because prospect theory itself was a very dangerous idea to the establishment at the
0:17:38 time. And yes, there’s a lot of debates right now in the literature about how robust those findings
0:17:43 are so many years later. But that aside, it was a very important contribution. I’m curious for
0:17:49 how that connects into your worldview. Sure. In fact, that’s kind of one of the most fun things
0:17:54 for me because what Kahneman and Firsky did was they found a new Venn diagram intersection of
0:18:00 psychology and economics. And this is a new Venn diagram intersection of physics and economics.
0:18:06 Oh, interesting. I love that. So you simply look at the incentives for individuals whenever you
0:18:12 organize into a group. It’s only relevant inside a group. And then you borrow an idea from physics,
0:18:17 which is a phase transition. How does this phase transition idea come into play? I mean,
0:18:21 you’re a physicist by training. So I get that there’s a physics aspect to this. But how does
0:18:26 this play in the innovation organizational context? Every phase transition in nature is a result of
0:18:32 two competing forces, a tug of war. And the phase transition is as you adjust the balance,
0:18:39 you change certain parameters of design or the environment, you adjust the balance between
0:18:44 those two forces and boom, at some point, they cross and the system snaps. I’ll give you an
0:18:48 example with water. There’s a glass of water right here, which your audience can’t quite see.
0:18:52 I’ll take a photo. But you can imagine. When I stick my finger in the glass of water,
0:18:59 I can swish it around, right? Except as I gradually lower the temperature, all of a sudden,
0:19:04 at a critical temperature, boom, the behavior completely changes.
0:19:12 So when it freezes. Exactly. Why? How did those molecules know to go from sloshing around and
0:19:17 being fluid to being completely rigid? It’s the same molecules. And it’s the exact same molecules.
0:19:23 There’s no CEO molecule with a bullhorn. And that’s sort of the other counterintuitive thing
0:19:28 about this is it doesn’t matter what the CEO does. Okay, it’s not about the culture. Yeah,
0:19:34 it’s about understanding those two forces and what are the balance. So here’s the two forces
0:19:40 in a glass of water. There’s one force, entropy, which wants to makes water molecules run around
0:19:46 and be free. Right. The creative force, one could argue. If one wants to do that, entropy is actually
0:19:53 a hard physical quantity. And it is the case that molecules have two competing forces, one of which
0:19:58 is entropy, which wants to make them run around and be free. And the other is binding energy.
0:20:04 It wants to lock every molecule of water, exactly 2.8 angstroms from its neighbor.
0:20:10 When the temperature is high, entropy wins. Now, as you gradually lower the temperature,
0:20:16 the entropy force gets weaker and weaker and weaker. And the binding energy gets more and more
0:20:22 important. And then what happens? Boom. At 32 Fahrenheit, those two things cross. Below 32,
0:20:28 it becomes totally rigid. So it has nothing to do with the CEO molecule yelling, be rigid, be fluid.
0:20:32 Let’s adopt a fluid culture and all slush around. The CEO can set the temperature.
0:20:37 That’s exactly right. Control parameter is the equivalent of temperature. It controls the
0:20:42 behavior of the system. It turns out when you work out the mathematics of this, of the incentives,
0:20:47 the two forces, there are four control parameters inside organizations, elements of organizational
0:20:53 design. And not the CEO, the executive team, the board has control over, just like you have control
0:20:58 over a glass of water. In classic innovation parlance, it is sort of disruptive versus non-disruptive
0:21:03 core versus non-core, these two forces. However you describe them, the point is that organizations
0:21:08 clearly have these two competing forces. So back to not being, you know, the cultural stuff that
0:21:14 management literature gives us, what structurally needs to happen to make organizations know how
0:21:21 to navigate these two forces? Okay. So when you bring a group of people together and you organize
0:21:27 them and you set a mission and you tie a reward system, incentives tied to that mission, all of
0:21:35 a sudden you create two forces. And here’s what they are. The first one is stake in outcome.
0:21:43 Think of that as equity. So let’s say to stick with our industry, you have a small biotech company
0:21:46 and you’re developing a new cancer drug. Your stake in outcome is your stake in that drug
0:21:52 success. Does it work or not? Now when you’re a small, your stake in outcome is enormous. If the
0:21:58 drug works, everybody’s a hero and a millionaire. If the drug fails, everybody’s unemployed.
0:22:02 Right. What’s the second force? It’s perks of rank.
0:22:08 Interesting. So you can think of this almost as equity in cash or equity in base salary.
0:22:14 When you’re a tiny small company, stakes in outcome are so much bigger than perks of rank.
0:22:21 Now let’s flip it. Let’s say you, Sonal, are a manager at Pfizer. So now let’s look at your
0:22:26 incentives. What’s the stake in outcome? Well, if you make a good drug, let’s say it sells
0:22:31 500 million a year, that’s a pretty good drug. Plus you’re going to be fighting for the next seven
0:22:35 or 10 years. Everybody wants your budget and there’s going to be the three deaths with the
0:22:38 loon shot. You’re going to have to try to defend it. So your stake in outcome is pretty tiny.
0:22:44 What’s the perks of rank? Well, if there’s a committee meeting and you sit around the table
0:22:48 and keep, you know, trashing these loon shots and doing it in a very thoughtful wise, and maybe
0:22:54 you’re very funny and kind of witty, and they’re like, well, this young person, she’s got a great
0:22:59 head. She’s aligned with our vision. Right. What might you get? Promoted. Exactly. And then what
0:23:04 happens when you’re promoted? 30% bump in salary. And not much change in stake in equity, actually.
0:23:11 Exactly. So we just went from small, a small company, stake in outcome was much bigger.
0:23:17 Perks of ranks were irrelevant. To large, where stakes in outcome is very small and perks of
0:23:21 rank. So that’s very much equivalent to the water. These are the dynamic two forces that you’re
0:23:26 describing. Right. And instead of temperature, it’s size. So somewhere between small and large,
0:23:31 as you grow and you grow and you grow, all of a sudden the balance of incentives suddenly shift.
0:23:38 And the same person, the same sonal, would go from pounding the table. This is the most awesome idea
0:23:42 to kind of making fun of it at committee meetings. You know, I think this makes a ton of sense. And
0:23:46 we see it in big companies. And I think the challenge then is, you know, as your company is
0:23:51 growing, what can you do about it? So let me ask you a question. When it snows overnight,
0:23:56 what do you sprinkle on your sidewalk? Salt. Why? Because it blocks the snow from
0:23:59 hardening and freezing over. Exactly. And the reason it does that is because adding salt to
0:24:05 water lowers the freezing temperature. This is a control parameter. It is another. So
0:24:09 the really interesting thing about complex systems, which is what we’re talking about,
0:24:13 whether it’s a glass of water or humans understanding incentives when you bring people
0:24:18 together in a group, is that they have more than one control parameter. There are four other control
0:24:24 parameters. And to what Vijay was saying earlier, that’s what a manager or a leader or an executive
0:24:29 team or a board can control. You may not control size. You may need 100 people. You may need 500
0:24:34 people. You may need 10,000 people. And that, you can’t control, but you can add a little salt.
0:24:37 It’s interesting to think about how far the analogy goes and where it breaks down. So I think one of
0:24:42 the most powerful things about it is the fact that you can take the same person from one face to the
0:24:46 other. I agree. Just like with the molecule. Yeah. So that part is really interesting. And also,
0:24:51 just what you’re talking about, you can imagine doping and putting in certain types into given
0:24:55 phases to change their behavior. And I think we’ve seen that in various groups, right? You can put a
0:24:59 certain type of person to a team. And so some people will change with phases. Some people are
0:25:04 just intrinsically different. I thought where you were going with this was doping in semiconductors,
0:25:07 which actually gave us the transistor. Well, you know, but that doping either one,
0:25:10 you’re talking about alloys or semiconductors is the same concept, right? Actually, it’s exactly
0:25:15 right. If you take an insulator like silicon and you dope it a little bit and then you get a metal
0:25:20 and it’s a phase transition. That’s what Phil Anderson won his Nobel Prize for, is the localization
0:25:26 transition, the metal insulator transition. But to get to your question, yes, this is a phase
0:25:32 transition. That’s kind of the theme of the book. CEOs at the end of the day, they’re sort of lamenting,
0:25:36 oh, no one cares. Well, that’s because they don’t own half the equity of the company like you do.
0:25:41 But when you were a small startup, of course, everybody, their incentives were aligned.
0:25:45 And of course they rushed out because they were going to be unemployed if the loonshot failed.
0:25:49 I find it’s a very hopeful idea, quite honestly, because you’re saying that you can,
0:25:53 you essentially can be a big company that can figure out how to innovate. You can be a small
0:25:57 company that doesn’t innovate, that actually sometimes goes into the frozen zone without
0:26:02 figuring out how to keep the entropy moving and the creativity and the excitement moving,
0:26:08 and that you can more precisely be both at the same time. And you can file up or down various
0:26:13 parameters to make that happen. So I find that’s a very hopeful message. A question I have, and
0:26:18 when you think about the history of innovation, kind of going back full circle to what makes
0:26:23 something a loonshot versus a moonshot. And yes, I heard you, a moonshot’s a destination,
0:26:30 a loonshot could be a false fail along the way. I still want to understand how you know when a
0:26:37 field, an area becomes more ready to be optimized a certain way.
0:26:42 This reminds me a lot of just the general topic of how a biology right now is shifting from
0:26:47 science to engineering. And there’s lots of different trends and reasons why. We talk about
0:26:51 these two different groups, the sort of creative types, the inventors, the artists versus the
0:26:56 the soldiers, the people who get things done. We’re seeing that shift in biology as one can
0:27:01 finally engineer biology in many ways. So in that field then, when you move from say artistry to
0:27:06 soldierry, for lack of a better phrase, the way I’ve heard you describe this is like the bespoke
0:27:11 to the machine automator. How do you in this particular example know? Yeah, I mean, that’s the
0:27:16 billion dollar or trillion dollar question, right? Yeah, literally. And so, you know, what it comes
0:27:21 down to when, at least for me, what I’m looking for is both in Safi’s language that there is a
0:27:27 destination that we see where you would land, and that at each step of the way that there is a way
0:27:31 to do it through engineering like execution, there’s all these different places where bio plays
0:27:38 a role, either with consumer or with healthcare or other industries. And in each one, it’s just not
0:27:42 enough that there’s a technology in our product that there really is probably market fit. And
0:27:46 that’s always unavoidable. And probably market fit is one of the hardest things for anyone to
0:27:52 sort of just theorize, you know, because you can very easily be wrong, or you can be wrong,
0:27:56 could be three years too early. It’s not wrong, it’s when. It’s when. And so, some of that also,
0:28:01 you have to have the intuition that it’s now, you can have theses for why it’s now. In the end,
0:28:06 the market will tell you whether you’re right. Like the classic example, Xerox was a great example
0:28:10 of a disruptor in earlier days with just the copying machine. People are like, “Well, I don’t
0:28:14 need copies. What do I want copies for?” You know, it just wasn’t part of the workflow. You wouldn’t
0:28:18 think to use a machine that doesn’t exist. And that only after people start to understand, “Oh,
0:28:24 actually, it was really useful to have this,” did you create a market. And so, in a sense that
0:28:30 device had product market fit before the market was there. And so, I think, I agree with Sophie in
0:28:34 the sense that just because the market isn’t there now, isn’t a reason to kill it. But obviously,
0:28:38 there has to become a market. Otherwise, there’s no one to buy it.
0:28:42 I would then argue just to reconcile these different views. It basically is how you,
0:28:46 the word market, just might be over-indexed. Because what we’re really talking about is a
0:28:51 pull, a draw, a need. But some might even say a want. Because sometimes you don’t need something.
0:28:55 You might just want something but not know that you want it in the classic Jobsian sense.
0:29:01 Right. In the sense of some, I mean, I can bring it to very nuts and bolts. Someone’s got to pay
0:29:04 for it. And this is one of the mistakes I see over and over again with startup founders is that
0:29:10 when they’re small, it feels like they’re all artists. They do a great job of building an org.
0:29:14 And then suddenly, the org is not what they’re comfortable with or familiar with because they
0:29:18 have built an army. And they actually never want to join the army. Right. They don’t want to be navy,
0:29:22 actually, in that context, right? So, when you think of artists versus soldiers, this idea that
0:29:26 there’s the creatives and the builders and the people have to fight the war on the ground,
0:29:30 another way to think about it is Jobs’ famous quote about pirates versus navy. I mean,
0:29:33 there’s so many different variations of this over time in innovation history.
0:29:37 I think our companies have both too. But hopefully, there’s something for the soldiers to start
0:29:41 working on, you know, to go to war on, even though, you know, you constantly have to continue to
0:29:46 innovate. So you need both. But so when I think about how I’m going to find them, I want to find
0:29:50 things that are just at that transition. If it’s only artists, it’s not going to work. But once
0:29:54 they’re there, we have to think about which ones have the opportunity to become the moonshots
0:29:58 eventually. And which ones also, frankly, that can continue to innovate. And so one of the things
0:30:02 I thought was interesting about the book was that phase transition interplay between the two and
0:30:07 how it’s not a bad thing to have that phase transition. And that was maybe one of the more
0:30:10 surprising things. There are two things that I find very hopeful. I know a lot of people have
0:30:17 found very helpful. One is, and they’re both a little bit against the grain of what you often hear.
0:30:20 I love things that are against the grain. More against the grain, the better.
0:30:25 Especially what I hear from my tech friends in Silicon Valley. So one is don’t hire people from
0:30:31 big companies or big people from big companies are a terrible fit. And I thought that as well.
0:30:34 We’re doing the great stuff because we’re the risk takers. We’re the innovators and all those big
0:30:40 corporate guys, they’re risk averse and that’s why you shouldn’t hire. And then as you grow up,
0:30:43 you start doing these partnerships with large farmers. So you’re spending a lot of time with
0:30:48 these people and you get to know them well. And then these people are just like us. If they’re
0:30:53 advisor and they have no stake in outcome and a ton of stake in getting promoted, what do you think
0:30:57 they’re going to do? They’re going to work on being promoted. And if they’re at a 50-person
0:31:03 startup and being promoted or titled is totally irrelevant, but whether the project works or not,
0:31:06 means whether they’re unemployed or not, what are they going to do? They’re going to pound
0:31:11 the table and save the thing if it stumbles. So it’s about the subtle influence of incentives.
0:31:16 And when you really peel back the layers on that, you tease out all these things that you can do
0:31:21 to balance the incentives better for what you’re trying to achieve. So that’s one hopeful thing.
0:31:25 It’s not really about there, these two kinds of people. You should never hire a big corporation.
0:31:28 It’s the same person in different environments. Exactly. It’s the molecule that lands on the
0:31:34 block of ice versus the molecule that drops in a glass of water. So okay, now you say these
0:31:40 organizations have these phase transitions when they grow, they go from being totally fluid and
0:31:46 embracing wild new ideas to being quite rigid. So what do we do if we are a 500,
0:31:52 a thousand, or 10,000-person company? And so the point is that innovating well is a phase of
0:31:58 organization, just like being solid or liquid is a phase of matter. It’s a property of the whole
0:32:02 that doesn’t depend so much on the details of the parts. That’s what it means in
0:32:06 science or physics of an emergent behavior. And actually, it’s just sort of straightforward
0:32:12 mathematics. And the book that innovating well and a focus on politics or not is a phase of human
0:32:15 organization. I love that you actually have an equation that you outlined in the appendix,
0:32:20 which is awesome. We’ve never seen that in a management book. That took a lot of convincing
0:32:25 with the publisher. That’s another story. It got relegated. To the appendix. We’re way back there.
0:32:33 But anyways, if it’s a phase of organization, there are operationally excellent, high execution,
0:32:39 or you are innovating well. Larger companies today or startups that have grown need to do both.
0:32:43 Yeah. Mark talks about how startups are on the sort of five-year cycle theory. And I think of
0:32:48 this a little bit like growing versus sustaining organizations that even startups, by definition,
0:32:53 after a certain amount of time, have this inevitable gravity that begins to hit because
0:32:56 there are other startups coming out there sort of doing things faster.
0:33:01 Part of it is that also these environments keep on changing as we were talking about. And so
0:33:05 as the environment changes, the question is how we can sort of push back on the environment. And
0:33:09 maybe to abuse the analogy in physics with phase transitions, you can have hysteresis.
0:33:15 Which is… You can push the environment past where something… So there’s a really fun thing
0:33:20 to do. And so kids, you could do this at home. For the beer, put in the freezer. You take it out.
0:33:25 And it looks like water. But it’s already below the freezing temperature. And you just knock it.
0:33:29 And it’ll suddenly, it’ll freeze right in front of you. I’ve never done that. I’ve never got kids out of it.
0:33:34 The knocking kind of catalyzes this. And so this is a historic effect that it really should have been
0:33:39 a solid. But that it was lingering in the old phase for a little bit. What’s the point of that?
0:33:43 So the point is, you may think you have innovation in your org. And you may think that you have the
0:33:49 environment. But it is just a message of the past. And in an instant, you can lose it.
0:33:54 Just like the beer can shift. And so you need to really understand what your environment is.
0:33:59 And as a CEO, as a leader of something, you have to understand your org. You have to look to see,
0:34:04 do you have this phase segregation? Do you have two healthy phases there? The pirates in the navy,
0:34:08 the artists and the soldiers. And then what do you need to do to sort of affect the balance?
0:34:12 And it can be the hysteretic argument is that it can be misleading sometimes, too.
0:34:18 I love that beer example. So it is absolutely true that matter can’t be in two phases at the same time.
0:34:25 You can’t be solid and liquid at the same time. So if complex systems can never be in two phases
0:34:32 at the same time, you can’t be solid and liquid. You can’t be water and ice. How can you do both?
0:34:36 There’s one exception to the rule that you can’t be in two phases at the same time.
0:34:44 And that’s right on the edge of a phase transition. Life at 32 Fahrenheit. What happens is,
0:34:51 matter separates. You get blocks of ice and pools of liquid. So you get phase separation
0:34:56 and they coexist at the same time. The second thing that happens is you get something called
0:35:00 dynamic equilibrium, which the molecules go back and forth and back and forth. Little molecules
0:35:05 swimming in the pool of liquid hits on the block of ice and freezes. A molecule on the surface of
0:35:10 the ice starts shaking loose and then goes into the pool. And it’s back. It’s a continuous cycle.
0:35:14 I mean, if you think about the word dynamic equilibrium itself as a seeming contradiction,
0:35:17 but it’s not in this context because it’s dynamic and in equilibrium at the same time.
0:35:24 Exactly. It’s a balance cycle of exchanging back and forth with neither side dominating the other.
0:35:30 And that’s what happened during World War II. Vannevar Bush came in, told FDR, he said,
0:35:35 “We’re going to lose this war.” At a 10 minute meeting in the summer of 1940, he said, “We are
0:35:40 far behind the technologies that are going to be crucial to winning this war.”
0:35:46 The culture of the military, not only is it resistant to new technologies like any
0:35:50 military culture, but it should be. It is the right culture for their job.
0:35:53 In this case, they’re literally soldiers.
0:35:58 They’re literally soldiers. And he said, “I want you to give me a group of people,
0:36:03 give me the authority to mobilize the nation’s scientists for war and we’ll develop the radical
0:36:08 new technologies and I’ll report just to you.” So he created phase separation within the federal
0:36:13 government. And to this day, that really is the origin of the national research infrastructure
0:36:20 of the United States. The National Science Foundation, the NIH, DARPA, came out of Vannevar
0:36:27 Bush’s idea during the Second World War. But he made sure where most, where this fails so often,
0:36:32 so many years, so many times across companies. I mean, Xerox Park is a great example of that
0:36:37 because it was overly siloed. They missed the second part, which is the dynamic equilibrium.
0:36:39 Constantly moving back and forth.
0:36:44 So here’s the message for entrepreneurs or leaders or managers, which is also a little
0:36:51 different than what you get typically in Silicon Valley and is different than the model of a great
0:36:57 manager, a great leader. And that is, there is this myth that the great manager, the great
0:37:03 leader of a technology company is this product, product, product person and they build their…
0:37:07 Oh, you’re really fighting some major dogma right now. This is going to be blasphemous.
0:37:08 I want to hear it.
0:37:13 They build the company, this great company on the back of their inventions.
0:37:21 So the companies across history that have done that have failed spectacularly. They may survive
0:37:27 for a while when there’s a Moses on the mountain with a staff anointing the next Holy Loonshot.
0:37:32 They may survive for a while, that’s the hysteresis effect,
0:37:36 but eventually they’ll turn. That’s exactly what happened with Polaroid when Edwin Land,
0:37:42 who certainly deserved Nobel Prizes for some of his breakthroughs, stood on top of that mountain
0:37:46 and anointed the next Holy Loonshot and that was the end of Polaroid.
0:37:51 So the difference, the ones who have really succeeded, let’s say Vannevar Bush,
0:37:55 a theater veil at AT&T, the ones who really succeeded didn’t have that mindset.
0:38:01 Vannevar Bush was a brilliant inventor. He invented the first analog computer. He anticipated
0:38:04 much of the personal computer industry and certainly the internet and…
0:38:06 He wrote that memex memo, right? I remember reading it.
0:38:11 The famous essay published in the Atlantic, which inspired Doug Engelbart and a lot of
0:38:16 that traces back to Vannevar Bush, but as he likes to say, “I made no original contribution
0:38:21 to the war. None of my inventions ever amounted to anything.” What he did, rather than lead as a
0:38:28 Moses, he led as a gardener and he saw his job as maintaining life at 32 Fahrenheit.
0:38:32 Setting up the structure, the scaffolding that allowed that dynamic equilibrium,
0:38:38 that living at the edge. He maintained the balance between his group and the military.
0:38:44 And so if there’s a lesson there, it’s not about your ideas. Your job is to maintain life at 32
0:38:51 Fahrenheit. Your job is to maintain the exchange of… Because where innovation fails is not in the
0:38:56 supply of new idea. It’s in the transfer of those new ideas to the field. It’s actually like Xerox,
0:39:01 exact same thing. They were giving these computers to people selling typewriters and what were they
0:39:04 being commissioned on? Typewriters. Exactly. It was a very sales-driven organization.
0:39:10 It was a subtle influence of incentives. Bush recognized that the dynamic equilibrium,
0:39:16 it’s not just transfer one way from the scientist or the artist or the creative engineers to the
0:39:21 salespeople or the soldiers in the field, but the other way. Because most ideas suck when they first
0:39:27 come out of the lab. There’s a huge jump from a technology to a product, like a really beautiful
0:39:30 product. And then even if you have a really beautiful product, you still have to create a
0:39:35 market and sell to sell it. I was outside of my Stanford office after I’ve been at ACSNZ for a
0:39:40 while. And a founder at Stanford came by and he’s like, “You present this whole new tech.” And he’s
0:39:44 like, “Oh my god.” And I was like, “Well, this is really interesting, but I’m not sure you could
0:39:48 build a company.” And he’s like, “You know, the physics was so hard, building a company will
0:39:54 be easy.” But actually, the tech is not enough. And this goes back to Steve Jobs’ story is learn
0:39:59 to love your artists and soldiers equally. So when Steve Jobs started his first time at Apple,
0:40:06 he was exactly like that. I only love the artist. And people who know the history know this quite
0:40:12 well. It was a disaster. So the Lisa that he tried to run was a flop, the Apple 3 that when it was
0:40:16 a flop. And when he developed the Macintosh, although it was a phenomenal ad that goes down in the
0:40:22 history of advertising, it was a disaster as a product. The sales completely plummeted. He
0:40:28 loved his artist and he ridiculed the soldiers. And it caused enormous dysfunction and it almost
0:40:34 made Apple go bankrupt. And then he did this same thing at Next. When his next computer company,
0:40:38 it was product, product, product, it’s all about the product, da, da, da, da, da. And that was,
0:40:42 that product flopped. And then he bought this little company, this Lucas Film Group, because
0:40:46 they had a computer. And he thought this is an even bigger, faster computer, the Pixar Image
0:40:52 Computer called PIC. And that product flopped. But eventually over 10 years, when he came back
0:40:58 to Apple the second time, he had Johnny Eve, that was his artist, and he had Tim Cook. Tim Cook was
0:41:02 known as the Attila the Hun of Inventory before. And the myth around him, oh, is all his ideas,
0:41:07 and he championed the product. That’s not really what happened. In fact, when he did that, it was
0:41:13 a disaster. And only by evolving and learning to love his artists and shoulders equally,
0:41:19 that’s when things really took off. And that’s why Vannevar Bush succeeded. And so many scientists
0:41:24 before him failed, because soldiers love soldiers. Artists love artists. He not only worked well
0:41:29 with the military, he loved them and respected them. And that’s why he was able to bridge the
0:41:33 two. Yeah. I would argue just to Vijay’s point too, though, that this is why there’s such a
0:41:39 resurgence in modern tech management literature and startups in particular, around the important
0:41:43 role of product managers and product management. Because traditionally, there was so much discussion
0:41:49 around engineering and sales and design, all being sort of separated and siloed. But now,
0:41:52 through this, and it’s not a new role, it’s been around for ages. But in a sense, that product
0:41:58 manager role is to connect the engineers to set the temperature. Exactly. There’s sort of like
0:42:02 this person who maintains a dynamic equilibrium between orgs, they move and travel. I actually
0:42:07 think it’s actually one of the most important schools of thought for the future of the firm
0:42:11 in the startup to company environment, which is the evolving role of product management in this.
0:42:16 Because that is the concrete instantiation of exactly what you’re describing, Safi.
0:42:20 There’s all these different threads going on here. Bush and his famous memo right after World War
0:42:26 II define what basic research is. And in Safi’s concept of dynamic equilibrium,
0:42:32 by, I think, inadvertently talking about the separation between science and engineering,
0:42:35 he kind of pulled engineering away from scientists. Oh, that’s fascinating.
0:42:40 Yeah. And that now there was these two different disciplines and science is done by scientists
0:42:45 and engineering is done by engineers. And that actually took a little bit about that
0:42:50 equilibrium that was there our way. And so what we’re seeing now is something different where
0:42:54 scientists are doing translational work, which is another name for engineering.
0:42:59 And in biology in particular, the shift from science to engineering is becoming very powerful.
0:43:03 And one of the things I think is always a topic very germane, I think, to our listeners and to me
0:43:09 is one of the more things that we can do that to practically learn from this. Because I think the
0:43:14 analysis is very compelling. But if you’re a CEO right now, what can you do? I mean, what are the
0:43:19 knobs? And so in the book, you talk about one of the classic things being the magic number.
0:43:24 Basically, a number of people at which beyond that, it’s hard to, you have to start to create
0:43:29 an org and a hierarchical structure too often. And the magic number is canonically 150, but
0:43:35 often I see the magic number being 30 when it’s not built right. And so how can you push it
0:43:38 such that you can try to maintain what we loved about the start from the beginning?
0:43:43 Right. So this idea of a magic number, like the Dunbar number comes from the idea that the number
0:43:48 of neurons in your brain is fixed. So sometimes, as you know, in science or in physics, you might
0:43:53 have a good observation, but the wrong theory. So if you just look at the hidden influence of
0:43:59 incentives out pops a size, but it’s not a fixed number. I have a little fun with it because I
0:44:05 show how if you plug in certain things, it can be 150, but it can be as low as 30 or as high as
0:44:11 5000, because it depends on these four parameters. So when you adjust these four parameters,
0:44:14 you make those numbers bigger. What are the four parameters again? I don’t actually think we listed
0:44:21 them. There’s equity fraction. For example, to what extent is a person being incentivized by
0:44:26 base salary, in which case there’ll be a ton more focus on politics versus to what extent are they
0:44:32 being rewarded on their project? Right. Another parameter is you might call it return on politics.
0:44:41 To what extent is the system’s design so that individuals can lobby their managers for promotion
0:44:46 and that has an impact? Obviously, the more impact lobbying has, the more political and the more
0:44:52 careers matter, but some companies, actually Google does this, McKinsey does this and others,
0:44:56 they take the promotion decisions completely away from the manager. So that’s
0:45:02 another example. Right. And then the third? The third one is project skill fit. So if you’re very
0:45:07 if you’re very good and working more on your project, moves the project, if you’re a coffee
0:45:12 machine designer, you’re good at coffee machine design, then you’re likely to spend more time on
0:45:16 your coffee machine. I love that idea. It’s like founder market fit. It’s like writer topic fit.
0:45:19 It’s this idea that you really match the skills and passion of the person
0:45:23 so well with the thing they’re working on. This is project skill fit. Now, if somebody puts you
0:45:28 on coffee machine design and you’re like me, not a very good designer with very little aesthetic
0:45:33 sense, then it doesn’t matter if you spend an extra hour or 10 hours or 100 hours on your project,
0:45:37 it’s going to be still the same lousy coffee machine. You might as well spend your time on
0:45:43 politics and lobbying. So project skill fit is another parameter. And the fourth parameter?
0:45:49 Management span. So there’s a lot of literature on management span, but they kind of miss the
0:45:55 point. There isn’t a right management span. There’s the right span for your goal. So for example,
0:46:00 if you are building planes, you don’t want to assemble 10 planes and see which eight fall
0:46:06 out of the sky. You don’t want a very innovative risk taking. Right. You don’t want to be doing
0:46:09 that with like autonomous cars, for instance. Right. So there you actually want to narrow
0:46:14 a span because that encourages quality control and redundancy checking because you don’t want
0:46:19 those eight planes. On the other hand, if you want to do a ton of experiments and have groups
0:46:24 work together well, you actually want a very wide span. So there’s no right answer for a company.
0:46:28 And that’s what like all these things that average across the company make no sense at all. Right.
0:46:33 If the group that’s assembling planes, you want one kind of system, one kind of metric,
0:46:38 you want narrow spans, you want high quality controls, the group that’s creating crazy new
0:46:42 technologies for those planes, you want those spans as wide as possible with very different
0:46:50 metrics. So there’s a handful of these parameters and those are what you adjust to tune for the
0:46:55 goals that you want to achieve. Yeah. So many of the world’s problems that can be addressed through
0:46:59 bioengineering, healthcare comes off naturally into many different sub parts about healthcare,
0:47:02 but just the world around us is created by biology, whether we’re talking about
0:47:08 the food we eat, the air we breathe, just about everything. And so that is a huge opportunity
0:47:12 for companies to come in and become huge companies because of the potential of the problems they
0:47:17 can address. And it’s just fascinating to think about how the dynamics of human nature comes into
0:47:22 play here and that there really are these knobs that a CEO can turn. And now I think the question is,
0:47:26 can you really read your organization right and understand what the temperature is?
0:47:30 Finding the thermometer or the equivalent starts to become, I think, really critical.
0:47:33 And if you can find that and know where these knobs are, I think you could
0:47:37 have a huge impact on your work. Yeah, that’s great. Thank you for joining the A6NC podcast.
0:47:40 Thank you so much. Thanks. It was a lot of fun to be here.
with Safi Bahcall (@safibahcall), Vijay Pande (@vijaypande), and Sonal Chokshi (@smc90)
A ”moonshot” is a destination (like going to the moon, quite literally) — but nurturing ”loonshots” (which often involves a number of stumbles along the way) is how we get there. This goes beyond the trite mantra of failing fast! It is about not having ”false fails” or not killing the seemingly small ideas that could lead to outsized yet unexpected outcomes, observes Safi Bahcall (physicist, ex-startup founder, and CEO of a public biotech company), author of the new book, Loonshots: How to Nurture the Crazy Ideas That Win Wars, Cure Diseases, and Transform Industries.So in this episode of the a16z Podcast — in conversation with a16z bio general partner Vijay Pande and Sonal Chokshi — Bahcall shares why concepts like ”disruptive innovation” cause him gas; why doing market projections can sometimes be crap; and why most management books that focus on culture are b.s.Because CEOs and culture, argues Bahcall, do not control organizational behavior… but hidden incentives, ”phase transitions”, and specific control parameters do. So how can organizations — of any size, big or small — be in two states at the same time: both fluid and stable, soft and solid, with high entropy yet bound energy, and both artists and soldiers? The answer may be in a more scientific, less ”squishy” framework for management at the intersection of physics and economics. Big empires always miss the small but important new ideas… can this be why?