You Don't Need More Freedom. You Need Better Constraints | David Epstein - EP 818

27:18And it failed because I don't think we ever had the right framing. And because we didn't have the right framing, the CEO of the company and the CFO could never understand how to explain it. And it's so interesting to me now because we spent hundreds of millions trying to implement this. Home Depot is now spending 15 or 16 billion trying to replicate what we were doing in 2005 to 2010. And it kind of made me think about if General Magic had gotten it right with the smartphone, where would they be now compared to Apple? And are there any lessons from what I just briefly mentioned about Lowe's and what happened to General Magic that you might be able to tie together based on your own research?

28:10Yeah. I mean, without knowing more details about the problem of framing? Because I'd love to hear more about that. So if you're willing to tell me a little bit more about the CEO and the CFO, were they trying to frame it internally, externally, or what was the problem they were having explaining it externally, what this thing is, or internally so that the teams could cohere? Or can you tell me a little bit more about that? I think the issue was that the CEO used to be the CFO. And so he was so used to explaining things from a shareholder value perspective. And I don't think he really looked at it because he wasn't an operator who grew up working on a cash register, working in the stores, and really having to interface with the customer.

28:57I don't think he understood how difficult it was to interface with Lowe's at that time when you were a customer. And for whatever reason at that time, they did not want to have a loyalty program. They were 100 % against it. So we were trying to put in a loyalty program underneath the surface, but doing it in a completely different way that no one had approached because we didn't want you to have to present a card. We just wanted you to feel the loyalty because we understood you so much. In fact, we developed this tool called Single View of the Customer that without even having a loyalty program, we were able to use data matching and had a 92 % accuracy rate on understanding who our customer was, regardless if they were in their first home or their second home or third.

29:47But I don't think that Robert really understood how to explain this issue from looking at it from a customer standpoint, because he was always trying to explain it from a shareholder value standpoint. And they're two completely different things on how you value the problem we were trying to solve. Interesting. Okay. That problem you're trying to solve issue, I think tracks absolutely with General Magic. So to give people a little quick background, General Magic, this is a story about the danger of too much freedom in the absence of constraints. This was a company conceived in the 80s, started in the early 90s, that was so visionary that Goldman Sachs took them public in the first concept IPO in Silicon Valley history.

30:32They went public just with an idea, not a product. Three former Apple employees, in one of whom who'd become the CEO named Mark Peratt was, in the late 80s, was making schematics that if I showed you right now, you would think was the iPhone, right? Thin glass rectangle, no protruding buttons, just a touchscreen, rectangular apps. It's a phone, it's a computer, it's an entertainment center, et cetera. So he saw the future. The web did not exist when he was drawing that. He really saw the future. And so it drew this incredible collection of talent, money to the team. And they were going to build basically the iPhone starting in 1990, totally audacious.

31:14They were even doing patents that are coming back now because of autonomous AI agents. Like they were just, I read his PhD dissertation from Stanford in 1976 where he coined the term information economy. And like this guy was an incredible visionary. And they had so much talent they had this like literally i remember they made a trip to japan and ceos mike stern who was the general counsel and cfo at one point recounted to me being chased down a train platform by a ceo of one of the largest companies in the world please take my money and so they had this huge consortium of international companies and they were innovating like crazy i mean earliest form of usb precursors to emojis.

31:59They made a virtual gathering space where agents could interact for commerce. And what did they call it in 1990? The cloud. They were just so far ahead. But because they had so much money and so much talent, they could do anything. So they frequently did do anything. And they just added and added and the project grew. More people, more resources. When it finally shipped way late. It shipped with a 200 page user manual. And it wasn't clear to anyone what problem it was solving for them because it had so much stuff that it wasn't really coherent. And so it was an absolute flop. The company goes from stock price doubling on day one to basically worthless in two years.

32:44And one of the problems I interviewed dozens of former employees that they would tell me is we did not have a clear customer. There's this saying that they called their customer Joe Sixpack and nobody knew who the heck that was, right? There's this saying that people don't want a quarter inch drill. They want a quarter inch hole in the wall. Not what cool thing can I make for them, but what problem can I solve? And they did not have that. And so they were so unfocused because they were not defining a user problem that they were solving. And so it was so amorphous. Every time they added some feature, they just retrofit the story to say, well, here's why people need this, instead of actually understanding the customer problem and building to that.

33:25So people internal to General Magic who recognized this problem did some interesting things. One guy said, oh, the web happened. He built a website, a low-level service engineer built a website on his free time. He called Auction Web to facilitate commerce between strangers. And General Magic had a right to any IP developed by employees. So he brought it to the leaders, said, look, it's doing one of the things we want. This is way too small. This is small potatoes. We have this much larger vision. So he took it outside and changed the name. And that became eBay. Right. Pierre Omidyar was a low-level service engineer.

34:05Another guy, Jeff Hawkins, was creating an app that turned strokes with a stylus into writing. And when General Magic kept expanding, he said, you know what? There is a clear customer problem. Busy professionals want to sync their contacts and calendar and take them on the go. So he moved out and created his own device that was just contacts, calendar, memo pad, period. Three of the bajillion things that General Magic was doing, and that was the Palm Pilot, smash hit in the same era. So the people who really spent time defining a customer problem they were solving, that made so much difference because it channeled and focused their work.

34:47It sounds like similarly to the experience at Lowe's, there was, at least for the CEO maybe, a lack of really focusing tightly on what customer problem exists and what is the way that we're solving that. I think the mistake when I really think about it is that what we should have done is we should have gone back to like Amazon starting with books or NVIDIA with gaming or DeepMind with video games. I think if we had solved a smaller problem first and then another smaller problem and not tried to solve the huge thing, then it probably would have been a different story because there would have been success stories that could have been touted and we could have trickled the tiny pebbles.

35:38And so I think it was a sequencing issue. We tried to go for the Big Bang instead of sequencing a lot of small wins that would have been let up and allowed people to see the vision along the way. And I think it was that vision gap that probably cost us the long-term solution that we were so really 18 months from accomplishing. That's so interesting. And Afiz, I want to hear how the decision was made eventually to kill it, because that's often very hard once a lot of money has been spent. But that sequencing, it reminded me of, so inside the box, I write a bit about the work of a Danish professor named Bent Flubier, who spent his career studying projects.

36:18And he kept a database of 16 ,000 projects. And he led a bunch of projects himself and was involved in projects that were over a billion dollars budget and things like that. And he defined what he said was the typical path. What he found, first of all, was that only eight and a half percent of projects came in on time and budget and 0.5 % did that and delivered the promised benefits. And the typical pattern for not working out well is what he called think fast, act slow, where there'd be a big idea that was so enticing that it would quickly become a big project. And instead of starting small, doing some experimentation, some iteration, defining the boundaries, biting off that small chunk.

37:03You just go into big execution really quickly. And then it's act slow because once there's so much momentum, it's actually hard to adjust to these lessons that you're learning while you're going. The reverse, what he called think slow, act fast, was spending time being small and defining the boundaries and picking up, defining those smaller problems you're going to start with. And then when you move into execution, you can act much more quickly. And or as he colloquially, he said Pixar planning. So one of the reasons I wanted to write a bunch about Pixar is because in his work, he said that Pixar planning is like the best way of planning.

37:37And he was talking about all these sorts of rules and constraints they have, but things like, to your point, where they will keep a director in story development with a tiny team for years, forcing them to refine the core of that story. What is it about? Stripping away characters. That's why there was no ennui and schadenfreude in the first Inside Out as they were stripping it away, making it simpler. And it might seem really inefficient to spend years in that mode, but the costs only explode once they move into production and complexity explodes. And so if they can stay in that small group, then when they really have the boundaries defined, they move into production, they can act really quickly.

38:15And they were always finding ways to be small, to your point. I don't know if you remember, but in theaters, Pixar used to always start their films, and maybe still do. I'm not sure. I haven't seen one in theater lately, with a short. It would be like a five to ten minute short. And those were always them experimenting with some new kind of animation, where it was, all right, we're going to just try it for a short and see how this works. And sometimes it worked out and sometimes it didn't. But they were always looking for these ways to be as small as possible while they could. It's a very good point.

38:48When I took over the application development organization, everything we were doing was a typical waterfall type of development. And one of the things I was happy to bring in, and hopefully I think they use it now as the main thing, is I brought in agile development. Because what was happening, I remember I met with the SVP of strategy one time and he says to me, his name was Scott. He says, John, you guys are absolutely incredible at delivering projects that have absolutely no value by the time they're done. And it was hard to hear it, but he was right because you'd get so far into these things that you had so much sunk cost fallacy going on that you just wanted to bring it to an end, even if it was irrelevant by the time you were done with it.

39:36And what the agile development really allowed us to do was to really use trial and error and to be much closer with the customer to experiment with what we were actually, whether that customer was an internal person or external person, is what we're building actually solving the problem. And if not, it's much easier to throw that away after a couple of months of work than it is after two to three years and tens of millions of dollars. Speaking of which, with that big effort, because that sunk cost fallacy is so difficult, such a strong bias for humans, how was the decision made to end that sort of holistic customer experience program that you were talking about?

40:18Painfully. I'm sure. At the time, this has got to be a very rare occurrence. Our president and chief operating officer was the brother of my boss, the chief information officer. So you don't ever see that in Fortune 50 companies. And it was the brother, the older brother, who was the COO who eventually made the decision. And what it came down to is we're working with a lot of larger vendors on the outskirts of this project. But the innermost, the hardest thing to work on was this very complicated data engine that had to keep track of all the requests that the customer was making in real time. to run it across all the other data layers that were in the company.

41:02And we partnered with a startup, which we might as well have just bought instead of working with them the way we were. And it was really their inability to deliver on that technology that kept being this huge hurdle. It kept the, we have to wait, we have to wait, we have to wait and keep giving them more and more money. And after you're giving this company tens of millions of dollars to survive, and you're the one who's basically bankrolling them, eventually a decision had to be made. Got it. Those are tough decisions. I always admire it because in some ways it would be easier, if more expensive, to bring it to the end anyway.

41:42Although, like you said, then creating something. What was the phrase that you, the hard feedback that you got great at creating solutions that are no longer needed or something like that? They're basically irrelevant by the time they come into production. So I kind of admire those. Anyone who can defeat the sunk cost fallacy at any time. I have some admiration for it, even though it's painful. And I admired Scott for at least being honest with me, giving me an opportunity to try to fix it. Well, I want to talk to you about another project. So as I was reading the book, one of the stories that really stayed with me came out of chapter six, and it was a story about UPS.

42:17And why this was relevant for me is UPS was trying to answer a question that when I was a consultant at Arthur Anderson, I was trying to answer for a different company, waste management. And that is, how do you find the most efficient route for thousands of drivers making thousands of garbage stops or deliveries every single day? And what was interesting to me is when we went into this project, we went in believing the challenge was building a better route. But what we discovered, and we discovered this because we actually sat in the garbage trucks and did the routes with the drivers, was that the software could optimize miles, but what it couldn't optimize was the drivers were carrying years of local knowledge that the system couldn't possibly know, neighborhoods, alleys, traffic patterns, customers, all the little things that never showed up in the data.

43:19And so as I was reading about what UPS went through, I found myself thinking, man, I've lived through this before. What was one of the biggest surprises for you that the technology only became powerful when it started learning from the people? Or was there another insight that you gained? I think the first big surprise for me was the extent to which the drivers, once they made this, had to remap all of America because Google Maps wasn't accurate enough for UPS, and made this routing algorithm. The first big surprise to me and to them was that the people who actually had to use it basically rejected it, and their managers rejected it.

44:01And part of that was because it did weird things. it would tell them if they were driving down one road and there was a delivery to make on a side street a human will stop off and make it and it would tell them often to skip it and come back later or the one that that several drivers told me about was if you go in a cul-de-sac and it would tell you to zigzag instead of walking around the cul-de-sac maybe you're going to save two meters or something, but it's weird to a human. Then on top of that, it didn't necessarily take into account there things like, oh, someone has to sign for this package.

44:38They're not home, but I know they work right around the corner or something like that. Or I know I don't want to be here when camp drop-off is happening or whatever it is, all these little things. And because of that, I think the lack of feeling like they could still use their own local knowledge and because of some of the oddities of the way the routing algorithm told them to behave, they just plain rejected it. And so you had this incredible, expensive, time-intensive project that does nothing for you if you don't have the change management right. And so it actually did not take them, it was intensive, but it took them a lot longer to integrate this into the workflow than it did to actually create the routing algorithm.

45:20And so they had to go back to square one and ended up meeting with every driver eventually. showing them how these things worse they started making these video game consoles where you could play with the routing algorithm with orion and see how it worked and if you change things and they were trying to make it go from a black box to a glass box as they said and they started making games like beat the computer if you can do better job do it i remember one of the drivers who'd been working for 20 years he came home the first day he tried it and said see i told you your algorithm stinks. I beat it.

45:53And when they looked at his data, he'd gone 135 miles to what the suggested, Orion suggested a route that was like 140. But he'd been going 150 for 15 years. And so this competition with the machine got him to do better. And so I think it was just surprising, less surprising now that I've done that research for chapter six, but that visions of change are much more compelling when they include visions of continuity. I think that's I'm paraphrasing something that Adam Grant has said, but it's so true that the change in many cases with technological disruption, the change management is the hard part.

46:33And I think that's totally true with AI too, right? I think one of the reasons I've spent a lot of time with one company over the last year and a half that helps other companies implement AI. And what's become apparent to me is that one of the reasons why we're seeing, despite this incredibly unprecedented rate of adoption, that some of the productivity numbers are not keeping up, is that we're often doing a poor job of actually integrating it into the known work streams and caring about that and thinking about that. And so I think the surprise to me was the degree to which you have to, the bigger a change is, the more important it is to include a vision of continuity and to ground it in something familiar for people and that that goes throughout design right that's why your electric car has a plug-in where the non-existent gas tank is and why the folders on your desktop look like old hanging folders because this the what are big changes we're trying to embed them in things that communicate to you what it is that this thing does and how it connects to the things you're used to david one thing i have learned in two decades of doing many technology projects is technology projects rarely fail because the technology doesn't work.

47:44They fail because people don't change. And you're right, because there isn't a change management strategy. Well, as you were alluding to, I think one of the smartest decisions that the UPS team made was realizing they couldn't ask drivers to trust a system that the drivers couldn't understand. So I really liked how they brought the drivers in to see how the routes were built, experiment with them, understand the reasoning. And I love this analogy of turning a black box into a glass box. It made me wonder if explainability isn't just a design choice. Is it what actually creates trust? To some degree, the algorithm, right, of the code.

48:28but they did want some hands-on experience with this thing just to start to tease out its behavior and understand it and see that there was some logic to it and some sense to it and so I absolutely I think so for sure that it's connecting with things that are understandable to us that make sense that really build that trust one thing I really loved about the UPS story is that in the end, they thought they were solving a routing problem, and it turned out they were solving a learning problem. And why I like this story is it's a great example of what happens over and over again in your book. We think we're solving one problem, but the real breakthrough comes when we discover we've been asking the wrong question all along, which is what we started with.

49:18Is that one of the main threads that you wanted to connect all these different stories? Yes. Yes. I think you put it very well. I'm not sure I have anything to add to the way that you just summarized that. Yes. I wanted to bridge from there to, I always think our books are influenced by someone greatly. For me, my book was really influenced by the lifetime of work by psychologist Gordon Flett. In your book, as I read it, Herbert Simon keeps coming back again and again throughout the book. And he wrote in 1970 that a wealth of information creates a poverty of attention. Today, that feels really prophetic.

50:03Why do you think Simon understood something about our future in 1970 that almost all other people missed? Man, if you read his writing about what was coming with AI starting the 50s he was seeing a lot of what was coming he was a founding father of AI but I think Simon first of all he was a massive polymath right he won the Nobel in economics he won the highest award in psychology won the Turing award in computer science and so one I think he saw across disciplines but at the same time in his own words he was really just studying one problem which is how humans make decisions and he was so enmeshed in that and the workings of the brain and modeling the workings of the brain.

50:47In some ways, he was fighting neoclassical economic theory, which modeled humans as rational actors. The more options will always make a better decision. And he was saying, theoretically, that's true. In practice, it's completely false because of the way our brains work. Sometimes more options will paralyze us. Sometimes they will lead to better decisions that we'll regret more. And sometimes that will just read to worse decisions. And so I think he was just so like acutely aware of how the things that seem like they should only be good, more information, more tools, because of the way our brains actually work can often have an ill fit for the way our minds actually work.

51:36And I think he was attuned to that because he had studied some economics and then realized that what he was learning did not, in many cases, represent real human behavior. So I think he was attuned to this kind of disconnect in a way earlier than most people were. I do have to say, I love it when I read a book and it introduces a person to me who I knew who Simon was, but I didn't understand the depth of going all the way from political science to winning a Nobel Prize in economics. And it was wild. It is very wild. He's like under every third rock in social sciences that I turn over. There's something in Simon.

52:17He was incredibly prolific also and just a brilliant person. And one other thing I wanted to quickly talk about is in episode 301 a few years ago, I interviewed Gloria Mark. And it was one of the most fascinating conversations I've had on this podcast because we spent a lot of time talking about attention residue and task switching. And reading your own thoughts on Gloria, I realized that as we were talking about Simon, he anticipated that entire conversation that we both had with Gloria decades before these things ever existed. How did learning this and talking to Gloria change how you think about protecting your own attention.

53:01Some of Gloria's work is the, to me, the sort of the, let's say, scariest research that went into the book. Because when she started studying people in the workplace, basically 25 years ago, she found people were tasks switching about every three minutes on average. And then I think 2012, it was every 75 seconds. And then by 2022, it was every 45 seconds. And the trouble is that multitasking is not really possible. You can talk and walk at the same time. So if one of the things is an automatic function, then you can, but not between real cognitive tasks. And two of the most interesting things, I think, from her work were, one, she was monitoring physiological measures, and she found that the number of tasks, task switches someone made during the day predicted their stress as measured by heart rate variability.

53:51And so this whole area of how attempting to multitask impacts stress was new to me and a little frightening. and also her finding that we become accustomed to a certain cadence of interruption, where if we're being interrupted by anything all day, whether it's people, whether it's notifications, whatever, you train your attention so that even if you say, today I have to focus and I put the phone away, you will self-interrupt with intrusive thoughts at the rate to which you become accustomed, as if we have this internal distraction barometer that wants to maintain a certain pace. And I'm like, I found that frightening.

54:28And I see that in myself when I'm, I'll go into these periods where I'm writing a book where I'm like really locked in and I won't even turn on my phone until my workday is done. And then I may shift into these modes where actually I'm task switching like crazy. And I can feel how hard it is for me to then sit down and read at the depth I want to and things like that. And with Gloria's advice, I started batching my work where she found people in offices and is a little old now. So maybe it's more, we'll check email on average, 77 different times a day. And so what you really want to do is if you have to answer all those, if you really have to answer all those emails, can you put that into one or three or five blocks a day where you're just answering email?

55:12And then you're not, what you don't want is 77 different tiny blocks throughout the day. And so I've been much more intentional about structuring my work where I'm monotasking in a certain block, taking a little break then moving to the next block and for the last thing of every day i do what i call the hemingway principle where hemingway would stop writing in the middle of a sentence so that he knew exactly where to pick up the next day and so my last act of every work day is to write down the important thing i'm going to start with the next day because i think it protects me from what's called the mere urgency effect in psychology where people will prioritize tasks that feel urgent like answering a message over things that are important even if those things are objectively much better use of time.

55:52So her work really just made me much more proactive in thinking about the way that I structure everything during my workday. Because I think at this point, as her work shows, if you're not structuring your attention, it is being structured for you. And I want to be the one driving it. I love it. And it's such an important message. I wanted to make sure we got it into this interview. Last question for you, David, is if someone listening wants to conduct their own theory of constraints audit this weekend or this week, what are the first questions they should start asking themselves to potentially uncover the bottleneck that's limiting everything else?

56:29I think the first thing I would probably do, I'll say the first thing that I did for myself, because I launched some months ago, the video operations, I made a YouTube channel, it's a total workflow that I'm not used to and all these kinds of things. And the problem early on was, I just could not move enough scripts through to make it viable basically and get ready for filming and so I stole something from a genetics lab that I visited where I made all of the steps of the process visible by with post-its on a wall and I said which of these steps where is the work piling up because it's not all these steps and it was it actually became pretty clear where some of the work was once I laid it out that way I said it's definitely not here it's definitely not here It might be here.

57:13Where's the work piling up? And once we identified that spot where work is piling up, it doesn't matter how efficient all the other steps are because the work just piles up at that one step. And so once we had that identified, I could talk to the team and say, all right, this is the bottleneck. How do we widen this step? Stop spending effort on the other steps because it's just making those steps faster and piling even more up at the bottleneck. And we need to focus our energy on how do we widen this one step? So I would make every step in the process visible. Just you can look at it on the wall, say, where's the work piling up?

57:44And then don't spend energy on the places where work isn't piling up because that's just running the machines faster so they can pile up at the slowest step. Awesome. Well, David, for a listener who wants to learn more about you, where's the best place for them to go? DavidEpstein.com. I have a free newsletter that you can find there and a free YouTube channel where I turn a lot of these things into sort of video lessons. That's new. So yeah, but it's all, you can find it all at davidepstein.com. And I would add an excellent sub stack that I subscribe to personally. Oh, thank you. Thank you. I try to pretend that nobody whose brain I admire reads that sub stack because for me, it's a place where I want to be a little free to be very imperfect.

58:23So I appreciate that. I will attempt to forget it. Awesome. Well, David, such an honor to have you today. Thank you so much for joining us. Total pleasure. Thank you for having me. I hope today's conversation with David Epstein gave you a new way to think about how you approach change, uncertainty, and the problems right in front of you. One of the ideas I keep coming back to is David's story about Mendeleev. We love the mythology of the breakthrough. The genius who suddenly sees something no one else can see. But the reality was much more ordinary. Mendeleev wasn't sitting around trying to change the world.

59:00He had a textbook to finish. And that constraint forced him to experiment, look for patterns, and ultimately see something much bigger. I think there's a powerful lesson in that. Sometimes we don't need more freedom, more options, or a bigger vision. Sometimes we need to pay closer attention to what's already in front of us. Because the problem you're trying to solve right now may actually be the doorway to what's next. If something from this conversation resonated with you, I'd love to hear about it. Send me a message on Substack, LinkedIn, or Instagram, or share this episode with someone who could use a different perspective on how to navigate an unpredictable world.

59:40And make sure you're following PassionStruck so you don't miss what's coming next. Because next week, we're taking everything we've been talking about—adaptability, uncertainty, constraints, and the ability to keep moving forward—and putting it to the ultimate test. I'm joined by Cole Brower. However, Cole became the first American woman to race solo non-stop and unassisted around the world. 27 ,000 miles, alone at sea, facing storms, exhaustion, isolation, and moments when everything could have gone wrong. But what's most fascinating isn't simply that she finished, it's what she discovered about herself along the way.

1:00:21At one point after being told she was too short to compete in the Southern Ocean, Cole found herself at what she describes at one of the lowest points in her life. And then something changed. She decided to go anyway. She went from being told she couldn't do it to sailing around the world alone. And there's one moment in our conversation that I think captures the spirit of this entire series. When you hit rock bottom is exactly what I need. to go back up. While I'm at the bottom, all I can do is go up. I can't really go any lower. I wish I didn't have to hit rock bottom to find exactly where I am going in life, but maybe that's kind of the beauty of it too.

1:01:03That's next week on Passion Struck, because David reminded us that we don't always need to escape our constraints. Sometimes we need to learn how to work within them. And Cole shows us what happens when you take that idea to the edge of what you believe is possible. Until then, remember, you don't have to know exactly where you're going to take the next step. Sometimes you just have to be willing to take it. I'm John Miles, and as always, stay curious, stay intentional, and stay passion-struck.