In this episode of the McKinsey Global Institute’s Forward Thinking podcast, co-host Michael Chui talks with Andy McAfee. McAfee is a principal research scientist at the MIT Sloan School of Management, co-founder and co-director of MIT’s initiative on the digital economy, and the inaugural visiting fellow at the Technology in Society organization at Google. His new book, The Geek Way: The Radical Mindset that Drives Extraordinary Results, will be published on November 16. In this podcast, he answers questions, including:
- What is a geek?
- How often does a technology alpha geek become a business alpha geek?
- What geek norms are associated with success?
- Does geek culture make diversity a challenge?
Michael Chui (co-host): Janet, what do you think are the most influential companies in the global economy?
Janet Bush (co-host): I’d have to say the big tech companies. Technology influences so many aspects of our lives as consumers, and also the ways that companies operate.
Michael Chui: Indeed. And what’s interesting about these companies is not only that they are developing and deploying new generations of disruptive technologies. Today’s guest has written a book that argues that many of these influential tech companies operate in different ways, which he has named “the Geek Way.”
Janet Bush: Well, that’s certainly an interesting title. I’m curious how geeks run companies.
Michael Chui: Andy, welcome to the podcast.
Andrew McAfee: Michael, it’s always a pleasure to talk with you.
Michael Chui: Terrific. Well, let’s start with how you ended up doing what you’re doing today. Where’d you grow up? What’d you study? What was your path to what you’re doing now?
Andrew McAfee: The short answer is I have no idea and it was largely random. But let’s try to tell a story that makes sense. I grew up in the middle of the country, in Indiana. And I am old enough that I was an impressionable young kid when these things called personal computers first started to appear in the world.
And I was the right kind of kid to be receptive to them, because I was kind of nerdy. I went to math camp. I always had my nose in a book. So when these things appeared and I heard about them, I thought they were just innately fascinating. And I don’t think I’ve ever let go of that feeling.
Michael Chui: That’s great. And so that’s an interesting childhood. It turns out I was once a CIO of Bloomington, Indiana.
Andrew McAfee: This story might not be totally foreign to you, Michael, or a lot of our listeners.
Michael Chui: What happened after that playing with computers as a kid?
Andrew McAfee: Playing with computers as a teenager. Mainly because I didn’t have any friends or dates. So why not just keep playing with computers? And then I got admitted to MIT, which was the gateway to this big, broad world, full of intimidating geeks. And I went there in the fall of 1984 to start my undergraduate degrees. And with the exception of a hiatus at Harvard, I’ve been at MIT for most of my adult life, and most of my career.
Michael Chui: What was the detour to Harvard? What happened there?
Andrew McAfee: I got done with my master’s degrees in 1990, and I worked for a couple years and realized that I wasn’t loving it. And I decided to go back to school. And I went to the business school at Harvard and did my doctorate there. And then I taught on the faculty there for about a decade. And then in 2009, I came back to the mother ship, two stops farther up on the Red Line [subway] in Cambridge, and went back to MIT.
Michael Chui: It’s a bit unusual to go to business school and get a doctorate, though, right? Most people go and get a master’s of business administration.
Andrew McAfee: Yes. Exactly. And to be clear, you used the right term. It’s not a PhD. It’s a doctorate. My degree is technically a doctorate in business administration. You can only imagine how much my PhD friends like to beat up on me for that. But I’m fascinated by the business world. And I’m fascinated by this intersection of technology and business.
And when I came back to MIT, Michael, like you know, I started working a ton with Erik Brynjolfsson, who’s a very, very, very good economist, and a good guy. Very, very, very, very good guy. And he helped put more of an economics spin on the work that I was doing and our thinking. And in the books we wrote together, economics is the base. And I think that has stayed with me.
Michael Chui: You are a best-selling author, some co-authored with Erik. And you have a book entitled The Geek Way: The Radical Mindset that Drives Extraordinary Results, available where good books are sold on November 14th of 2023.
Andrew McAfee: But you can preorder it now. Don’t be shy.
Michael Chui: But anyway, it is really interesting. Thank you for the privilege of being able to take a look at it. And I’d love to get into it a little bit. There’s a lot there, right? And so one of the things that you talk about, it’s entitled “the geek way,” and over 20 years ago, another author, Tim O’Reilly, talked about watching the alpha geeks. And I think he described them as the people who are inventing the future, which I think he mostly meant technologists. But tell me what you mean by geeks in this.
Andrew McAfee: Tim and I come from the same place, which is looking at a bunch of people who founded the computer industry in Northern California. And geek is a pretty appropriate term from them, because geek went from being the person who bites the head off a chicken in a circus sideshow to being somebody who spent a little bit too much time in front of a computer screen.
A geek is anybody who gets properly obsessed with a really hard, deep problem.
By the ’80s or ’90s, that was the primary definition of geek. But I think both Tim and I want to broaden that definition out. And I think the term has broadened out. And for me, a geek is anybody who gets properly obsessed with a really hard, deep problem. Can’t let it go, is very tenacious, and is willing to embrace unconventional solutions, is not tied to the conventional wisdom, or the mainstream, or the status quo.
So for me, geek is a lot broader than computer nerd. My working definition for it is an obsessive maverick. And the geeks that I became obsessed with—a lot of them were computer nerds, to be super clear—but they, in my eyes, were business geeks. And the problem they got obsessed with, properly obsessed with, is: how do we run and grow a company and keep alive its ability to innovate, to be agile, to be responsive, to execute at high levels, even as we grow, even as we scale, and even as we age?
And to sharpen that a little bit, I think the geeks got obsessed with avoiding the dysfunctions that seemed to plague companies throughout the industrial era. And they really wanted none of that. They wanted to do something different and better. And I’m not saying that they’ve perfected the formula, but I think they’ve succeeded. And my super shorthand message of the book is that the geeks have given the company an upgrade.
Michael Chui: Let’s get into some of those specific things that you’ve discovered about these business geeks. I want to pause for just a moment here, just because there are a set of VCs—historically, a lot of times you’d have these technologists found companies. And you’d hear this term, “adult supervision.” A businessperson would come in and provide the business expertise.
There are some VCs, some of whom you quote in the book, who have taken the view [that] the best business leaders are the technologists. Do you have a reflection on that? How often does the technology alpha geek become a business alpha geek as well?
Andrew McAfee: Much more than never. Think about Reed Hastings, who was a computer science master’s graduate at Stanford. His first company built debugging tools. This guy is a good old-fashioned technologist founder. And when he decided to start Netflix, to co-found Netflix—and keep in mind, the original goal of Netflix was not dominating Hollywood or completely changing the economics of the filmed entertainment industry, it was mailing DVDs to us, because they could fit in a normal first-class envelope. That was it.
What’s astonishing to me is how successful Netflix, and a bunch of other companies that I talk about in the book, have been. And my whole message is that a big part of their success is that their founders kept on working the problem of, again, how do I keep the growth going? How do I keep this spirit of innovation, entrepreneurship? To use Jeff Bezos’s term, how do I keep this a day-one organization?
Because in a sense, what the geeks are terrified of are some of the consequences of a phrase that you used, which is “adult supervision.” And it seems like over and over again, when we try to layer in, air quotes, “adult supervision” into companies, we wind up instead with bureaucracy, and sclerosis, and cultures that get jammed up and they can’t get anything done. And they become kind of stultifying places to work.
When we try to layer in ‘adult supervision’ into companies, we wind up instead with bureaucracy, and sclerosis, and cultures that get jammed up and they can’t get anything done.
And the geeks are like, “Whatever flavor of adult supervision that is, we do not want that. We want to find something very, very different.” And, again, I’m not saying that they’ve nailed it. But I think they have made a big, big step forward in how to have the right kind of oversight, adult supervision, and avoid some of those dysfunctions.
Michael Chui: And I think that’s what’s interesting about what you’ve written, is there is an interesting nuance to it. You can’t run a giant organization without some amount of structure and organization, but figuring out how some of these geek norms shape that and create the right balance, I think, is one of the things that comes through.
I mentioned geek norms. What are the geek norms that you discovered that underlie the extraordinary performance that you’ve identified?
Andrew McAfee: The way I boil down and describe the geek way is to describe it in a set of norms. And norms, as anybody who studies human cultures will tell you, norms are everywhere. They’re critically important. And they’re the community policing of any culture. They are the behaviors that are expected of you.
And if you don’t follow those behaviors, the community will let you know. It’s not that the boss will let you know. It’s not that you’re violating corporate policy. It’s that there are expectations for how you’re going to behave. And if you are not in line with those, you are going to hear about it. And the environment might become an uncomfortable one for you.
In a lot of organizations, and a lot of these kind of bureaucratic, sclerotic, old-fashioned companies, the norms are, “Don’t speak up. Don’t discuss the undiscussables. Stay in line. Just kind of follow the corporate rule book.”
I’m not saying the geeks don’t have rule books. But they believe in some very, very different norms. I describe four of them.
The first one is science. And science usually means beakers, and test tubes, and people with very, very advanced degrees. That’s not what we’re talking about. That’s not what the geeks believe science is. Science is an argument, it’s a debate, it’s back and forth. We’ve got to get to the bottom of this and we have to do it in a group, governed by a ground rule for how you’re going to settle those debates: evidence, analysis, experimentation—not seniority, charisma, beauty of PowerPoint, and things like that. So the first thing the geeks believe very deeply in is having evidence-based arguments.
Science is an argument, it’s a debate, it’s back and forth. The first thing the geeks believe very deeply in is having evidence-based arguments.
The second one is ownership, is actually trying to make real these concepts of autonomy and empowerment we’ve been hearing about forever. And they’re a lot easier to find in business books than they are in actual businesses. When you look at the evidence about the kinds of corporate cultures that come out of Northern California, or that were incubated there—and I want to say this one more time, they are far from perfect, but the people who work at these companies report higher levels of autonomy, and responsibility, and empowerment than we see almost any place else in the economy.
The third great geek norm is speed. And speed is not just velocity. It’s primarily iteration, or cadence. How often are we putting something out there in the world, getting feedback on it from reality, from a customer, and incorporating that feedback in getting to the next cycle? It’s a complete departure from the up-front, planning-heavy, analysis-heavy approaches to managing big projects that we built up during the industrial era. The geeks believe in an MVP, a minimum viable plan. And after that, start iterating, because that’s actually how you learn.
And then the last great geek norm is openness. It’s a pretty close synonym to psychological safety, which Amy Edmondson talks with such passion about. Do you have an organization where people will speak truth to power? And on the flip side, does the hierarchy of the organization, do the bosses show openness to being corrected, to pivoting, to letting go of their ideas, to realizing that what they’re doing isn’t working, and that they need to move on? That openness is a two-way street here.
So I put those four norms together. And I think that science, ownership, speed, and openness are the things that I believe deeply differentiate geek cultures from what they’re replacing.
Michael Chui: Well, let’s talk about each of these. Why don’t we start with science? A lot of our listeners might be familiar with A/B testing, this idea, create a controlled experiment and try to develop some finding from it. And that’s the empirical evidence that you talked about.
But you also talked about creating an argument, and that just collecting data is not actually the thing which distinguishes, because, again, as you’ve said before, I think lots of organizations over decades would say, “We have evidence-based management discussions.” And so what really distinguishes these companies?
There’s a piece where you talk about, can we have too much science? And so designers will say, “Yeah. If I just designed based on doing a bunch of experiments, I wouldn’t have creativity,” or what have you. So how do you think about bringing this empirically based argument in a way that actually distinguishes a company?
Andrew McAfee: Michael, like you point out, a lot of companies have been talking about how evidence-driven and how data-driven they are. I think a key difference is that what a lot of companies do is have a HIPPO, who’s going to make a decision.
HIPPO stands for ‘highest-paid person’s opinion.’ And it’s how most companies make most of their decisions.
HIPPO is my new favorite business acronym. It stands for “highest-paid person’s opinion.” And it’s how most companies make most of their decisions. There could be some analysis that the data nerds did, but then they tee it up to the HIPPO. And the HIPPO essentially says, “Look. If the evidence aligns with my prior beliefs, with my intuition, with my big HIPPO gut, great, we’ll follow the evidence. If not, we’re going to go my way. After all, the reason I’m up high on the org chart is because of my experience and my intuition and my judgment. If the evidence doesn’t agree with that, we’re not going to follow the evidence.”
Whether or not it’s that explicit, I think that’s what happens at a lot of companies. And the geeks try very hard to say, “No, actually, that’s not what we’re doing.” I’ve got a quote that I love from Richard Feynman at the start of the chapter about science. And Feynman says, “Look. If your guess is wrong, it doesn’t matter who you are. It doesn’t matter how many Nobel Prizes you have around your neck. It doesn’t matter how august you are, how successful you’ve been. If your guess doesn’t line up with the evidence, if it doesn’t line up with reality, it’s wrong, and you need to walk away from it.”
That’s really what I mean. For me, that’s at the heart of science. And it’s important to spend a little bit of time on why science works so well, why it is so powerful. I think the key part of the answer is it is the best corrective we have ever come up with, the best corrective we humans have ever come up with to our own overconfidence.
Michael, I love my ideas. I’m pretty sure you like your ideas a lot, too. And if that’s the end of the line, we will stand or fall, and defend those ideas to our death. Science is a way to correct that overconfidence. And what’s critical is that we cannot be left to the job of refining or stress-testing our own ideas. That’s completely inappropriate. All we will do is reinforce how amazing our ideas are.
The group, however, is really good at assessing our ideas. So there’s this wonderful interplay between the individual level. Michael, you have a great idea. I have a great idea. We’re so proud of it. Get it to the group and let the group evaluate it.
It turns out that I think we humans have been wired by evolution, wired to be deeply lousy at evaluating our own ideas, and really, really good at evaluating the ideas of others. And science taps into that. It’s a group-level process. Norms are group-level things. And it’s a group-level process that defines the ground rule for making progress: What does the evidence say?
Michael Chui: Why is overconfidence evolutionarily advantageous?
Andrew McAfee: I put this great quote in The Geek Way from House of Games, the David Mamet film from the ’80s. There was a con man. And you know “con man” is short for “confidence man.” And he says to his mark, and he’s trying to get her deep into the game, he says, “It’s called a con game. Why? Because you give me your confidence? No. Because I give you mine.”
Confidence is a great thing for a human being to have. We follow more confident people. We’re more likely to trust them, to listen to them, to ally with them. The data on this are overwhelming. And so what appears to have happened—and this is a theory that I believe; I think it holds up really well—is that we have been wired to be chronically overconfident, because confidence is so valuable for us.
In evolutionary terms, it increases our fitness like crazy. Gives us higher status, gives us all the stuff that we really need. But our brains are not wired for reality. Our brains are wired to tell ourselves the best story they can get away with about ourselves.
I put a bunch of really cool studies and evidence in the book. One of them that really knocked me on my heels was this study where they gave people pictures that had been doctored. And they were doctored to either make the people look more attractive or less attractive, according to conventional standards. And they said, “Which one of these is the most accurate picture?” And on average, people picked the 20 percent more attractive picture. Like, “Yeah. That’s me. That’s what I look like.”
Now, the 50 percent more attractive might be more than our brains could plausibly tell ourselves. We’re not that handsome. But 20 percent, yeah, we can get away with that.
And over and over again, we see this overconfidence. It’s been called the hardest bias to get rid of, the most chronic human bias, the easiest one to elicit. The genius of science is that it doesn’t try to train you and me to be less overconfident. Maybe that training works, but it’s really, really hard. I think it’s about as hard as training us not to like calorie-rich foods. Our overconfidence is that deep.
What science does instead is say, “Great. Go be overconfident. Be overconfident about producing your evidence and presenting your arguments. But then the group gets to figure out where reality is, where the truth lies.” That process is a great reality discovery mechanism.
Michael Chui: I think a lot of our listeners will think: I can think of some alpha geeks, and they seem not to be lacking in confidence. So that would probably resonate. But then how does that work so that the group can change that? Because don’t those alpha geeks just become the HIPPOs?
Andrew McAfee: The incredibly successful, unbelievably insightful tech founders are not immune from this. And I tell the story in The Geek Way of Reed Hastings almost tanking the company, way back in 2011. Do you remember Qwikster? Remember his genius idea to split Netflix into two companies? There was one that was going to be Netflix, that was going to stream, because Hastings saw the streaming revolution coming. There was another one that was called Qwikster, where we were going to manage our DVD queues, if we still wanted that.
Now, that meant we had to manage two accounts. We were going to pay about 60 percent more. It was a deeply bad idea. It was an incredibly bad idea. But Hastings believed in it. And his company went along with him, because he was the visionary, passionate CEO of the place.
And when Qwikster almost sank the company, and it tanked the stock price on the order of 75 percent, Hastings, who had already published the culture deck, he had already been trying very, very hard with his head of people operations, Patty McCord, and a bunch of his colleagues, he’d been trying very hard to build a culture where people would speak truth to power and speak back to the boss.
And he realized in the wake of Qwikster that he had failed at that. And so he asked around. He said, “Why?” And they said, “Reed, you were so confident. And you’re a charismatic, prestigious guy at the top of the org chart. We thought it was a dumb idea. But we thought you’re usually right. We’re going to go along with it.”
And then what Hastings did not do is yell at his employees or say, “Wow. I had better train myself not to be overconfident.” Instead, he did something brilliant. He instituted a formal policy—and Netflix is not big on formal policies—where he said: If you’re going to launch something big, Qwikster level, you have to farm for dissent. You have to write up a memo, give it to a bunch of your colleagues, and say, “What do you like and not like about this?”
In other words, you’ve got to go out and do some science-y kind of thing. And I tell a couple of other stories in The Geek Way about how Netflix eventually got better at becoming an organization where people were comfortable talking back to the boss. It is not an overnight process. And one speech by the CEO won’t do it.
To get back to your initial point, the alpha geeks of the high-tech industries, the alpha geeks that I’m trying to learn from, that I think have built these brilliant companies, they are not immune from overconfidence. What the really good ones do is try to build companies that can compensate for it, and not let it tank the place.
Michael Chui: This might be a little bit out of order, but we’ve segued a little bit into this norm around openness, in terms of being able to challenge the boss, or challenge the person who would otherwise be a HIPPO. And, you mentioned earlier, psychological safety. How do you create that? Because, again, I think a lot of people have in their minds some iconic founders. I think you have some vignettes as well, where people can be, let’s say, extraordinarily frank when dealing with others. How do those two things—how can they both be true?
Andrew McAfee: You can be extraordinarily frank when dealing with others. But the instant you come across as abusive, or domineering, or not willing to listen, then you’ve shut the door on psychological safety.
You can be extraordinarily frank when dealing with others. But the instant you come across as abusive, or domineering, or not willing to listen, then you’ve shut the door on psychological safety.
I tell this story in The Geek Way that has stuck with me for a long time. I was at MIT by the time, when Brian Halligan, who was then the CEO of HubSpot, came to me. He co-founded it with Dharmesh Shah. He was the CEO of the place. It was not yet an extremely successful public company, but it was growing quickly. And he wanted to do some internal employee education.
So he came to me, and we brainstormed about it for a bit. And normally, that’s the end of the line. The educator and the CEO think about what they want to do, and then they go build the curriculum. Halligan took a different approach. He said, “OK, come on into the office. We’re going to present this to the HubSpotters.” And I thought, “Oh. That’s interesting.”
So he set up a meeting. There were about 20 people in the conference room. And Halligan and I talked. And then he was last, and he sat down. And that’s the cue for everybody to go, “What a great idea, Brian. Gosh. Thanks so much, Boss. We’re so happy with this.”
And instead, there was a ‘child’ in the room. He had to be a brand-new hire. And he stood up and he said, “There are a couple things I don’t like here.” And then he went on from there. And I thought, “Oh, wow. I get to watch a career-limiting moment unfold right in front of me.”
And I looked around the room to kind of watch the temperature rising. And it didn’t. And I realized that I was the only person in the room who found that at all exceptional or even interesting. And I looked at Halligan. And his body language didn’t change at all. He was looking at this kid with this very open expression on his face. And when the kid got done, Halligan said, “Yeah, that’s a good point. I hadn’t thought of that.”
And the reason I’m bringing this up and telling this anecdote at some length is that that is an unbelievably strong signal to the organization that, again, the dominant, the prestigious person—how do they react to these kinds of situations?
In addition to being overconfident, one of the other incredibly fundamental aspects of being a human being is that we learn from others. We do it consciously, we do it subconsciously. We do it much, much more than we’re aware of. And in particular, we have learned to imitate and emulate prestigious people.
That’s actually a decent strategy for learning how to survive in your environment. So the prestigious people in an organization have even more clout, even more influence than we usually think that they do. So when people in a room all look over at the CEO, who’s just been challenged by a junior employee, and they see him react that way, man, that is an extremely strong signal.
That’s not the only thing you need to do for openness and psychological safety. But the things that your prestigious people do matter a huge amount. And so one of the things I tell leaders, or even wannabe leaders, is: the way you react to that kind of challenge, as long as it’s not a jerk one, or it’s not just mindlessly confrontational, the way you react to that kind of stuff will go a long way to determining whether you have openness in your organization.
Michael Chui: There is a dialogue about folks who would say psychological safety is antithetical to performance, that organizations get caught up with all these internal messages, challenging the strategy, challenging this and challenging that. They lose the edge on executing. They coddle their employees. And therefore, they’re slow.
You’ve talked about some of the dysfunctions of quote-unquote “industrial era” companies. And yet, they say, we need people to be a lot less comfortable. We don’t want them to have safety. We want them to feel a little bit stressed, so they’ll perform at high levels. How do you think about that?
Andrew McAfee: I think that’s a misunderstanding of what psychological safety is. And this word “safety” has been expanded to cover all kinds of things. And I think that’s part of the confusion. Psychological safety is actually the opposite of a safe space, of freedom from being challenged.
Psychological safety is actually the opposite of a safe space, of freedom from being challenged.
That’s really not what it is at all. And it’s also not, every opinion you have is fantastic and needs to be voiced. If you’re showing up and you’re just shooting off your mouth to a discussion that’s supposed to be evidence based, you will hear about it in a science-based organization. That’s not a valid contribution here.
And I’ve watched geek organizations be very clear to people that the way you’re talking is not helpful. And it’s not part of what we do here. The last thing that psychological safety is not is, “Whatever you want to do throughout the working day is fine by us.”
In other words, let’s say you and I are trying to build a Netflix competitor. But we hire a bunch of people who believe deeply in climate change and in fixing global warming. That’s great. I believe in fixing global warming, too. But if they want to sit around five hours out of the working day and debate global warming with their colleagues, I think you and I need to have a way to say, that’s actually not what this company does for a living. And you can either get on board with the goals of the company, and contribute to them, or you can try to work on something that is more in line with the climate goals that seem to matter so much to you.
So it’s absolutely not a blandly cuddly environment, or everything you say is fantastic. It’s a place where challenges are accepted and acceptable, without people immediately getting defensive, and domineering, and shutting it down.
Michael Chui: Got it. Well, you mentioned their alignment with overall goals. One of the other geek norms you mentioned was ownership. What do you mean by that? Again, just the word itself is something I think people have talked about for decades. So what do you mean particularly here?
Andrew McAfee: Ownership is very close to these buzzwords that we’ve been tossing around for a while. It’s about autonomy and empowerment, responsibility. But like I said before, those buzzwords are usually a lot easier to find on the page than they are out there in organizations.
Bezos was super articulate about this when he got asked about day-one versus day-two organizations. And I think he had a deep insight. Because he talks about Amazon as a day-one organization. He’s been doing it for a long time. And one of his employees asked him in an all-hands, “Jeff, what’s day two?”
And Bezos’s answer was great. He said, “Day two is when you start trying to manage by anything other than the goals of the organization.” I’m paraphrasing. He said, “For example, process. Did we follow the process? And as long as we followed the process, and as long as our processes are great, we’re spending all of our time thinking about the process, and training the process, and designing the process. That’s a goal of an organization.” Bezos said, “You’ve got to watch that. Because the process can take over.”
I think that is a brilliant, very concise description of how bureaucracy happens. “We can’t have that happen again. We need a process. And we have to have a lot of coordination in this organization. So we’re going to build this very cross-functional thing.” And you’re expected to be part of that. And you really can’t get anything done without that.
That sounds really good on paper. And, Michael, I think your career is long enough that like me, you remember the business process reengineering craze that started in the mid-’90s and took corporate America by storm. And a lot of people who built their managerial careers in that time are huge fans of process, because it sounds so great, right? Define how the organization is going to work and monitor adherence to that process.
Man, the danger there is that process accumulates like plaque on teeth or something, or like barnacles on the hull of a ship. They can just get thicker, and thicker, and denser. And it can kind of take over. So I love Bezos’s insight that the process is not the thing. The goals of the organization are the thing.
I learned a story about Amazon that I actually didn’t know before, as I was researching the book. It turned out that Amazon, in the late ’90s, when it was growing like crazy, was on its way to becoming a sclerotic, bureaucratic, process-heavy mess, because they had an innovation process.
You had to submit your innovation idea and tell what resources you needed from the rest of the organization. And you got one of three emails back. The good one was, “Hey, your innovation idea got approved. And the other people who you need to contribute to it are on board. And they’re going to start helping you.”
That was the good one. The medium one was, “Your idea did not get approved. But you also don’t have to provide resources to anybody else.” Again, this is process working as designed. This was how they designed the company.
And then the third email was the worst of all. It was like, “Your ideas did not get approved. And you have to provide help and resources to other teams, while still being responsible for all of your performance objectives.”
Now, that middle email tells me something very deep about bureaucracies. It is great news when they leave you alone. And that third email tells you that they very often don’t. And it was just clear that this was jamming up Amazon, and everybody hated it, and it was slowing down the company.
And Bezos and his team said, “We have to do something different. We have to walk away from this.” That was actually the dawn of Amazon Web Services. Because they said, “First of all, we have to make it so you don’t need to ask the IT, the technology department for resources if you want to try something. We need to build a modular, and robust, and scalable tech infrastructure, so that you don’t need to ask permission. You can just start banging on the finance module, and the warehousing module, and all that stuff.”
The other thing they had to do was modularize the organization itself and work really hard to reduce interdependencies, and—here’s the weird part—reduce communication, reduce coordination. All those things became kind of like a tax at Amazon, or a thing that they wanted to minimize. And that is so antithetical to the way that I was originally taught to think about business, and about improving business.
Man, stop it. Stop all the coordination, the communicating, the process. All that becomes overhead. Try to build an organization where you delegate responsibility as far down as possible, you give people the autonomy and the tools to do what they need. Now, as part of that, you have to define their goals, and make sure that they know what they’re tracking to. And that’s where OKRs and the V2MOM process at Salesforce come from. Those are alignment processes.
And that’s a process in a bureaucracy that the geeks believe heavily in. But then their mantra is, “Get out of the way.” And if you’ve hired the right people and pointed them in the right direction, they will go execute. They will go innovate. They’ll go get things done. But this day-two process-heavy approach, a lot of the geeks that I learned a bunch from, man, they are so terrified of that. They think that’s the way to a long, slow, sad decline.
Michael Chui: Let’s talk about this alignment thing. Because there’s a passage in here, when I read it, which I was a little shocked by. But I think there’s some subtlety here, where it said, “The business geeks take a radical step. They stop coordination, collaboration, and communication.”
Overall, those seem like good things. So I suspect you don’t mean there’s none of that going on in organizations.
Andrew McAfee: Yeah, maybe “stop” was not the right verb. Maybe “minimize,” or “trim,” or “walk back.”
Michael Chui: My sense is that what you were saying there, when you’re talking about modularizing the organization, all of that stuff happens—coordination, communication, collaboration happens—within that module, who is attempting to accomplish something.
But as you were saying, there are organizational and technical means to make sure everyone else whom you need in order to get some things done in a big company can do it. Famously, Amazon has their application programming interfaces, where you can just access these capabilities. But you don’t have to ask for permission to do so.
There’s an online technical door where you can get whatever accounting you need, or whatever website you need. How do you hold those two things in your head at once? You actually need communication, collaboration, coordination, and you need to have a modular organization.
Andrew McAfee: Amazon’s answer to that excellent question is two-pizza teams, and then later, single-threaded leaders, because sometimes the teams got too big for two pizzas.
Michael Chui: What’s a two-pizza team?
Andrew McAfee: A two-pizza team was Amazon’s organizational solution to the bureaucracy that they were building up. And they said, what we want is to divide our groups.
Let’s say you and I are responsible for introducing shoes into Italy, or whatever. The team that’s going to do that is no bigger than could be fed by two pizzas, two large pizzas. That’s a relatively small team. So the fundamental unit of organization is a pretty small team.
And then what we’re going to do is work to reduce the required interdependencies between our team and the rest of Amazon. Now, of course, you’ve got to pick up the phone once in a while. You have to integrate. You have to coordinate and whatnot. But they want that to be minimized. And they want it to be as ad hoc as possible. And, again, this is a big departure from what we were doing in the internet era.
There’s a great book called Working Backward that describes this process at Amazon. And they said the teams that worked hardest up-front to reduce their interdependencies and to stop the coordination and communication were the ones with the best results down the road. And bureaucracies can be either hard—“You cannot get reimbursed unless you submit your expenses through exactly this process”—or soft.
There are gatekeepers, and there are informal custodians of things in the organization. And if you don’t touch base with them and go through them, you’re unlikely to be successful. A lot of the geek companies that I’ve studied are working really hard to reduce both the hard and the soft bureaucracies.
And I quote Benedict Evans. Michael, I’m sure you read his stuff. He’s a fantastic technology analyst. And he summarized the result. If you can pull this off, he summarized the result really beautifully when he was talking about Amazon. He said, “Amazon is a machine for producing more little Amazons.” In other words, all these little teams. It just adds on to the Amazon gestalt.
If the teams are outside the company, great. You have all the APIs you need to sell. If you want Amazon to fulfill, we’ll do that. And if you don’t, you can do that. Amazon is kind of a self-replicating, self-propagating organization, because they’ve modularized so successfully.
And Evans says, “Look, you don’t need to fly to Germany and have a few meetings before you can start selling shoes in Italy. You don’t need to go to Seattle and the mother ship. You can just spin up what you’re doing and go.”
And for me, that’s when you’ve got the technological and the organizational environment right to actually have ownership. Otherwise, how do you take ownership when you’re in the middle of a gigantic, process-heavy bureaucracy? What does that even look like? I tell all these vivid anecdotes in the book about people caught in the middle of these things. And they just want to get their work done. And all they do is sit in the weekly meeting, where they go through the latest checklist of whatever. And it feels like nothing ever happens.
Michael Chui: We’ve actually been fortunate to have Ben Evans as a guest on the podcast previously. But I’m curious. Because if you have all of these modules, all these small companies that you’ve created within your own company, at some point you do have to make choices. What is going to show up first on the mobile app? What’s going to be on the web page when you open it up? How do you do that if you have all of these entrepreneurial units working to get things done?
Andrew McAfee: I think you set up a team that determines the ranking of stuff, or the appearance of stuff on the website. You have another team that’s responsible for that. And they don’t have to go get approval from everybody else. And we judge them based on their results.
These swarms, these teams, can have some hierarchy built into them, of course. But the fundamental idea is that we are going to devolve down. And we’re going to make sure that people understand what the overall goals of the company are, the organization are, and how they fit into that, and then go. And then just let them go.
Michael Chui: This book is primarily about culture. You do spend some time talking about just how important status and prestige is, and the way that that causes both maladaptive behaviors within organizations, as well as the way you can leverage those things in order to create higher performance.
But again, in this case that you’re talking about, you have all of these small fighting units, you might say, who are fighting for status or prestige. And don’t you end up with these same types of dynamics that you do within, as you said, sclerotic bureaucracies?
Andrew McAfee: I don’t think so, or I hope not. Michael, like you say, we human beings want status. We want it a lot. All social vertebrates want status, because status is the way to have high fitness. It’s really just that simple.
So think about everything from chickens to elephant seals to chimpanzees. These are all social animals. And wow, do they have very, very, very clear status hierarchies, very hard fought. We, humans, we have richer forms of status. But the idea that we’ve risen past that, we’ve evolved out of desiring status, it makes no sense whatsoever.
We are ultra-social creatures. We crave status. There’s a great book that I quote from a lot by Will Storr called The Status Game, where he just drives this point home. Our desire for money seems to level off. Our desire for power seems to level off. Our desire for status appears to be absolutely insatiable.
Our desire for money seems to level off. Our desire for power seems to level off. Our desire for status appears to be absolutely insatiable.
You bring up this point: how do we square that with the fact that the organization has its own goals, and people might not automatically go along with them? What the geeks try to do is take away opportunities for status that are not aligned with the goals of the organization.
So, for example, what that means is that they work very hard so that I do not have the right, or the formal or the informal ability, to block what you’re doing. And that’s all over the place in coordination-heavy, process-heavy organizations. “You’ve got to go through me to get your budgetary request approved, or to get the resources that you need.”
Great. That is a very, very clear kind of status to me. And there’s no guarantee that I’m going to use that power in the best interests of the company. I’ll probably use it to acquire more status. I’m not going to let go of my fiefdom voluntarily.
So they worked very hard to take away opportunities to gain status that aren’t aligned with the goals of the organization. I had the chance to interview Satya Nadella as part of the book. And he did this brilliant work. Michael, can you think of a better corporate comeback story in your career than Microsoft under Nadella?
Michael Chui: It’s been remarkable, for sure.
Andrew McAfee: Remarkable, right? I had the chance to interview him. And this was one of the questions that I asked. And he did a couple brilliant things. He didn’t say, “Now, stop all the infighting.” He might as well not waste his breath.
What he did instead was say, for example, “You can no longer own a resource inside Microsoft.” Now, subject to a couple important considerations. But if you want to go use data or code, you can now do that. And you don’t have to ask that formal or informal permission. So when the AI team wants to go try to use all the GitHub code to train up a code-writing assistant, they didn’t have to go ask formal permission. That is a right that you have inside Microsoft. So you can take away the opportunity to block and to gain status by being a blocker.
Now, I want to be clear. The people who are blocking other people are not doing it in a very deliberate, Machiavellian way. Our minds don’t work like that. But they do know that this is how they have status in the organization. And expecting them to voluntarily walk away from that is a fool’s errand. The geeks are much smarter about taking away the status opportunities that aren’t aligned with the goals of the organization.
Michael Chui: Got it. And the fourth geek norm you talked about is speed. Again, everyone says their organization is fast, or should be faster. What’s different here? What do you see?
Andrew McAfee: Here’s the test. Because everybody talks about their agile teams, and their tiger teams, and how they do an MVP, and they’ve got a Skunk Works. Here’s the test. How long can you kid yourself or anybody else about the actual progress that you’re making on a big project?
In most big projects, man, there is so much room to kid yourself, or to kid somebody else, or to try to kid your peers. I turned in my manuscript for The Geek Way late. I spend a lot of time in the book talking about overconfidence and how chronically late things are. Did I get out of that trap? I absolutely did not.
But what the geeks do, and the whole heart of agile, is to say to the teams that are working, “Look. Every week, we’re going to have a meeting, or we’re going to have an event, where your customer gets to say whether or not you delivered the thing that you said you were going to deliver. We’re not going to take your word for it. You can’t work on your own for a long, long time. We’re going to get very regular, very clear objective check-ins about whether or not you are keeping up with the goals, and with the pack here.”
The beauty of speed, the beauty of this iterative, agile approach is that it strips away those opportunities to fall behind.
And in most big projects, there are plenty of places to hide how late you are. And, again, you hide it from yourself, first and foremost. And then you hide it from the rest of the world. The beauty of speed, the beauty of this iterative, agile approach is that it strips away those opportunities to fall behind, or at least to fall behind in a way that’s not noticeable and obvious to the rest of the organization.
For me, that’s the heart of what’s going on. That’s how to tell if it’s actually agile or not. I tell a long case study in the book about VW’s attempts to do over-the-air updates for its first electric cars. And it’s just this classic, really disheartening story about chronic delays and unpleasant surprises. My guess is that there were people inside VW during that time talking about their agile development approaches. But they weren’t doing what I would call the great geek norm of speed.
Michael Chui: What’s the 90 percent syndrome?
Andrew McAfee: This is a phenomenon that comes up in a lot of big projects, so much so that there was actual research on the 90 percent syndrome. And it’s a pattern where things seem to be going well for the first 90 percent of the timeline of a big project. And then right when it feels like the finish line is in sight, all these amazing problems come up, all these deep problems.
They really put the entire project back on its heels, delay things by a factor of two sometimes, and it’s so pervasive that a couple of my colleagues at MIT decided to go study it. And they wrote this fantastic paper. And it’s got one of my favorite titles of any academic paper, because it’s called “The Liar’s Club.”
They said that they were going around trying to understand the 90 percent syndrome. And they were at, I believe, an automaker, and they said to a manager, “OK. You’ve got the big weekly status meeting. Tell us about it.” And he said, “Oh, you mean the Liar’s Club?”
And of course, my colleagues are like, “Wait. Hold on. What are you talking about?” And the guy said, “Look, me and the other team leads going to that Monday morning meeting, we’re all behind to some degree, especially as the project goes on and on. And we all know that we’re behind. We all know that the other guy is probably behind as well. But we also know that we just have to not be the first person whose lateness is exposed to everybody, whose lateness can no longer be denied. Because when that happens, we have to expand the timeline of the project. That means we all get that extra time. And as long as we’re not that one poor person, our reputation remains intact.”
So it’s cynical, but it is the absolute correct game theory play, strategy here, is to walk into that meeting and join the Liar’s Club.
Now I’ll say this one more time. I’m not just talking about overt, strategic behavior and Machiavellianism. The main person you lie to about how well you’re doing is yourself. And so the Liar’s Club can happen in between your own two ears. What the geek norm of speed really does is disband the Liar’s Club. And if you can do that, wow. Wow, are you going to run rings around the competition.
Michael Chui: I think that the 90 percent syndrome will resonate with a lot of our listeners.
Andrew McAfee: When I talk about it, you see a lot of heads nodding. And you can do relatively simple classroom exercises to demonstrate, even when you’re not lying to other people, even when you’re not being an overt part of the Liar’s Club, how easy it is to kid yourself about the progress that you’re making. It’s the default. And the underappreciated beauty of agile methods, and these iterative approaches, and the geek norm of speed is that it reduces your ability to deceive yourself and other people—but I think primarily yourself.
Michael Chui: This book is primarily around culture. And we’ve talked about geeks. And the observation has been made that sometimes the culture around tech companies isn’t as welcoming, or appears not to be as welcoming, in terms of just the demographics of people who are in tech.
Diversity has been a challenge, as you’ve actually mentioned in the book, as well. What are your reflections on that? Are the aspects of geek culture, including overconfidence, arguments, are those things that keep people outside of the ability to be in these extraordinary companies?
Andrew McAfee: Something is keeping diversity away from big tech. And the evidence on this is fairly clear. Businessweek did a really nice comparison of big tech companies versus other large organizations in other industries. And the tech leadership and professional ranks are much more pale, stale, and male than we find elsewhere in the economy.
I believe it’s a real problem. I don’t think we understand it very well. And it’s a problem, because in addition to whatever other reasons you have for wanting more diversity, you will get better solutions, you’ll be able to serve different kinds of customers, if the diversity of your own people goes up.
It’s absolutely true that tech is a little more demographically a monoculture in some ways that we should care a great deal about. By the same token, however, tech has been extraordinarily welcoming to neurodivergent people, and to people who have alternative identities or alternative sexual orientations. I think those things are also very true.
That’s not to excuse. And I want to draw also a distinction between geek—which is, are you following these four norms, and is that the DNA, is that the culture of your company?—versus tech. Not all tech is geek. Theranos certainly was not. And not all geek is tech. We see that in some very different industries as well.
So I want to be clear. My job is not to cheerlead for big tech. My job is to identify this new flavor of running a company, which I think works better. And I mean “better” both in the sense of turbocharging performance and providing healthy environments for people to work in.
I think the geek way is better, and therefore it’s going to spread. But I distinguish that from what we see with today’s giant tech companies. They’re not exactly the same.
Michael Chui: That makes sense. Let’s come back to that in a moment. But one of the seismic events of tech over the past 12 months has been the real burgeoning of generative AI as a topic. I’d love to get your reflections on how generative AI, which, by our own research, differentially affects geeks, if you think about people who have higher levels of educational attainment, if you think about people who are knowledge workers. And so what are your reflections, as you’ve observed generative AI?
Andrew McAfee: I agree with you. The sweet spot for generative AI appears to be higher in the education and the skill ladder than the sweet spot of previous, very powerful technologies. It’s going to be really interesting to see how that plays out.
But I think the geek organizations are going to be the most successful ones at incorporating generative AI, if they’re open to it. And my definition of geek, remember, one of the norms is openness, which is pretty close to the opposite of clinging to the status quo. And being open, and being welcoming to something new that comes in, and trying to harness it, as opposed to find reasons not to do it.
Well, here comes gen AI and it is going to be super disruptive, Michael, like you and your colleagues well know. We are in the early innings of how big a deal generative AI is going to be.
AI is very clearly going to distinguish defensive organizations and defensive cultures from open ones.
I think it’s very clearly going to distinguish defensive organizations and defensive cultures from open ones. I expect that a lot of the companies that you and I have both spent a lot of time studying are going to be comparatively quite successful with gen AI because of their norm of openness.
Michael Chui: Well, if you don’t mind, let’s wrap up with just a lightning round of quick questions, quick answers.
Andrew McAfee: Let’s do it.
Michael Chui: You can feel free to pass if you like. What’s your favorite source of information about developments in technology?
Andrew McAfee: Twitter/X.
Michael Chui: What’s your favorite source of information about developments in management?
Andrew McAfee: Twitter/X.
Michael Chui: Which current non-tech company best embodies the geek way?
Andrew McAfee: Bridgewater [Associates] is in some ways a very geeky organization, because for decades, [Ray] Dalio has been trying to build a completely open organization. He’s taken it way, way out there, to the point that there are few, if any, meetings that are off the record or that are not available to the entire company.
And he’s conducting this wild experiment, where instead of reputation being spread by gossip, which is the way it happens in every other human community that I’m aware of, he’s making his people assess each other’s reputations all the time with this app called the Dot Collector, on iPads, so that everybody’s reputation, the things that we informally whisper about—“Is this a good guy to work with? Is he smart? Is he ethical?”—these things are on the intranet at Bridgewater. And everybody can see everybody else’s Dot scores about the things the company cares about.
Now, I’m not sure if that’s the right idea for every place. But holy cow, is that geeky. It’s obsessive. It’s maverick. And Bridgewater has been leaning so deeply into this norm of openness for decades. And they’re well outside the high-tech industry.
By no traditional definition is rocket launching a digital tech industry. That’s its own industry. It’s called aerospace, right? SpaceX in some ways is an incredibly geeky organization. They’ve embraced the norm of speed like crazy. Iteration. Failure. They’re willing to blow up rockets. And I do not think it’s a coincidence that they’re putting most of the Earth’s payload into space, and that they are absolutely dominating satellite communications these days.
Michael Chui: What company from before the 1990s would best embody the geek way?
Andrew McAfee: Oh, right now, Microsoft would. Microsoft was founded in ’79.
Michael Chui: Who is the most surprising leader that you discovered that leads using geek norms?
Andrew McAfee: Surprising? I’ll talk about Microsoft again. I didn’t know Nadella before I interviewed him for the book. And I absolutely was not expecting Microsoft to come charging back as well as it has since he took over as CEO.
I write in the book how their share price was just as flat as a corpse’s EKG for about a decade. They were an also-ran. They were profitable, large, but they were an also-ran in the tech industry. And then in comes Nadella. And he has just breathed life back into that company in a way that I just haven’t seen before. And when I got to talk to him, he just struck me as, personally, a fairly geeky guy. But, wow, has he worked hard on putting the four great geek norms in place.
Michael Chui: What’s your favorite experiment of all time?
Andrew McAfee: OK. I’ll try to tell this story quickly. And this might not be my favorite of all time, because there are so many that qualify. But there was this amazing study where they put people in an fMRI [functional magnetic resonance imaging] machine and had them play a game where they think they’re tossing a ball, a virtual ball, tossing it back and forth on a computer screen with somebody else.
But it turns out that somebody else is a bot. There’s always a “but” here in these experiments. And at some point there are two bots, and they just start tossing the ball back and forth to each other. And they leave the human being in the fMRI out.
Now the whole point of this experiment is to induce social pain in that person—not physical pain, social pain, the pain of exclusion—so they can watch what happens inside the person’s brain. And when they took them out and they looked at their scans, they were indistinguishable from somebody experiencing physical pain.
This is wild to me. It’s such a beautiful experiment. And it drives home the point that for us, ultra-social human beings, all you need to do to maintain a culture is put in place the pain of social exclusion. Follow the norms, and if you don’t, we’re not going to beat you to a pulp. We’re just going to exclude you.
That is legitimately and truly painful to human beings. That’s the glue that holds human cultures together, is that threat and that pain of social exclusion. So that experiment and a bunch of others drive that home.
Michael Chui: That’s dark.
Andrew McAfee: It’s dark, right? Most social psychology experiments are kind of dark. I worry about those people.
Michael Chui: For what societal problem do you think the geek norms could most accelerate progress?
Andrew McAfee: This is really hopeful, but if you’re open, then maybe the level of polarization goes down. Open means being ready and willing to pivot, taking an argument onboard and changing what you believe. And holy cow, do we need more of that these days.
Michael Chui: What would you advise a graduating secondary school student to study?
Andrew McAfee: So heading into college?
Michael Chui: If they go to college.
Andrew McAfee: Yeah, if they go to college, that’s a big if, right? Study economics. Study statistics. And study this new discipline called cultural evolution, which asks and answers the question: Why are we human beings the only species that launches spaceships? I think they’ve gone farther than any other discipline toward answering that question, which is a pretty basic question.
So I’m assuming the person is kind of a geekily oriented person. Go do a lot of math. You need it. Go do economics. It’s the science of how people interact with each other. And go learn about this discipline called cultural evolution.
Michael Chui: What three people would you most like to have dinner with?
Andrew McAfee: I would love to have dinner with Maria Montessori, a writer named M. F. K. Fisher, who wrote about food and wrote about gastronomy. But that’s a little bit like saying that Hunter S. Thompson wrote about driving in a car. She was just a gloriously insightful writer. And I would love to have dinner with her.
And then probably Feynman. Just one of the most incandescent intelligences of the 20th century. One of my really nerdy habits is watching Feynman YouTube videos. And I’m just kind of slack-jawed at how good he was.
Michael Chui: And what would be your one piece of advice for listeners of this podcast?
Andrew McAfee: Seek the company of those who search for truth. Run from those who have found it.
Michael Chui: Andy McAfee, thanks for joining us.
Andrew McAfee: Awesome, Michael. Thanks very much.