Transcription complète de Legibility

materially affect the outcome. 00:05:22.016 --> 00:05:30.817 Just as the commercial forester found it convenient to overlook minor forest products, so the cadastral official tended to ignore all but the main commercial use of a field. 00:05:31.896 --> 00:05:48.074 The fact that a field designated as growing wheat or hay might also be a significant source of bedding straw, gleanings, rabbits, birds, frogs, and mushrooms was not so much unknown as ignored, lest it needlessly complicate a straightforward administrative formula. 00:05:49.194 --> 00:06:01.334 The most significant instance of myopia, of course, was that the cadastral map and assessment system considered only the dimensions of the land and its value as a productive asset or as a commodity for sale. 00:06:02.274 --> 00:06:12.314 Any value that the land might have for subsistence purposes or for the local ecology was bracketed as aesthetic, ritual, or sentimental values." 00:06:15.161 --> 00:06:15.442 Wow. 00:06:17.422 --> 00:06:21.281 I am gonna place a bet that I know where that came from. 00:06:21.713 --> 00:06:23.033 All right, let's see. 00:06:23.074 --> 00:06:23.514 Let's see. 00:06:23.574 --> 00:06:24.454 I'm curious. 00:06:25.442 --> 00:06:38.901 So there are a set of books that I've run across in my life that before I read them, I thought about the world one way, and after I read them, I saw the world a different way. 00:06:38.982 --> 00:06:42.682 And the book Seeing Like a State was one of those books. 00:06:43.372 --> 00:06:47.021 And I'm gonna put my money on Seeing Like a State. 00:06:47.423 --> 00:06:48.341 Is that where this came from? 00:06:49.603 --> 00:06:51.425 We have a winner! How many times... 00:06:52.863 --> 00:06:53.925 i'm so glad 00:06:54.783 --> 00:06:56.134 We have a winner. 00:06:56.514 --> 00:07:07.303 I'm so thrilled because this is, gosh, we've been doing this now for a while and finally one of us has guessed the book correctly. 00:07:07.425 --> 00:07:07.903 Yes. 00:07:08.264 --> 00:07:13.884 And yes, that book is on my list of books that have changed how I see everything. 00:07:14.687 --> 00:07:26.708 The book is called Seeing Like a State, and the subtitle is significant, so I'm gonna read it: How Certain Schemes to Improve the Human Condition Have Failed. 00:07:27.728 --> 00:07:29.067 And it's by James C. 00:07:29.127 --> 00:07:29.588 Scott. 00:07:30.588 --> 00:07:31.947 So, you've read this? 00:07:33.019 --> 00:07:34.459 I have read that cover to cover. 00:07:34.699 --> 00:07:39.160 I really wanna read it again because I found it a very difficult book to read. 00:07:39.819 --> 00:07:42.040 And it wasn't... 00:07:42.259 --> 00:07:43.819 it was just dense, right? 00:07:43.879 --> 00:08:46.019 It was full of new ways of thinking that were so obviously right and well-researched, but were so counter to how I saw the world that it seemed like a portal that I went through that I thought the world existed one way, which is trees are just trees and plants are just plants and maps are just maps, and I came out the other side thinking,"Wow, there is an entirely different schema and viewpoint for looking at all of this," not just the difference between something being static versus dynamic, or not just a difference between something being complex or dynamically complex, but something being fundamentally measurable in a different way, and that there would be an entirely different construct for how to think about what you're looking at. 00:08:46.600 --> 00:08:59.922 Because seeing the forest for the trees is like the old way of looking at it, and seeing the forest through the eyes of the state, if you will, like the government state, is an entirely different way. 00:09:00.572 --> 00:09:17.841 And recognizing that the economics and land use and monetary value and tax base and people's intentions and hist- like, all of that goes into looking at how land functions. 00:09:19.101 --> 00:09:20.282 But it's not just land. 00:09:20.481 --> 00:09:21.522 Everything's like that. 00:09:22.101 --> 00:09:24.241 And it's the, it's this gen... 00:09:24.501 --> 00:09:25.392 I don't know what you call it. 00:09:25.392 --> 00:09:33.302 It's like this generalizable principle for looking at any map now, and looking... 00:09:33.381 --> 00:09:42.501 And not just a physical, not just like a road map, but any conceptual map of the world can be viewed now through multiple different points of view. 00:09:43.062 --> 00:09:46.782 I am so curious what brought this to our conversation. 00:09:47.081 --> 00:09:47.312 Yes. 00:09:47.481 --> 00:09:57.082 So that's what I wanted to bring out here because in hearing us talk about this book, it might start sounding a little nerdy and abstract. 00:09:57.092 --> 00:10:13.072 It's like,"What are you talking about maps and cadastral surveys and all this stuff?" I've been having conversations with line of business leaders in organizations, and they are being tasked by their managers to adopt AI. 00:10:15.052 --> 00:10:18.722 And there is a lot of that happening right now. 00:10:19.552 --> 00:10:22.113 Teams are adopting the technology. 00:10:23.773 --> 00:10:45.107 And I think it's fair to say that it's being done in a kind of hectic, experimental, ad hoc, somewhat bottom-up manner, where within certain boundaries, people are being given the tools and encouraged to experiment and explore with how they might be used in a business context. 00:10:46.927 --> 00:10:57.207 And I'm increasingly hearing that there is a lot of experimentation, and the experimentation is happening in a kind of very undirected fashion. 00:10:58.628 --> 00:11:06.927 the question becomes, What is the right balance between bottom-up and top-down transformation," right? 00:11:07.028 --> 00:11:17.508 Because management wants to make sure that initiatives are in service of the business's goals. 00:11:18.587 --> 00:11:26.447 We want to sell more, we want to produce more, we want to produce more effectively, more efficiently, more

cheaply. 00:11:28.008 --> 00:11:32.888 We want to reduce the number of defects, the number of returned items. 00:11:33.067 --> 00:11:33.488 I don't know. 00:11:33.827 --> 00:11:37.898 Different businesses have different numbers that they want to improve. 00:11:39.067 --> 00:11:56.100 And I feel like we might be starting to move into the phase in the deployment of AI where people are starting-- are asking the question:"So how do we use this? 00:11:56.479 --> 00:12:02.701 And how we avoid people developing dozens of agents that do the same thing, basically. 00:12:03.596 --> 00:12:16.062 And Seeing Like a State, like you, the first time I read it, and I have read it twice, and I'm actually considering reading it a third time now because it's one of these books that you revisit and get something new every time. 00:12:16.315 --> 00:12:16.634 Totally. 00:12:16.956 --> 00:12:17.174 Yeah 00:12:17.803 --> 00:12:26.702 But on reflection, I would say that this book is an argument against utopian top-down initiatives. 00:12:28.235 --> 00:12:29.035 Okay, fair enough 00:12:30.442 --> 00:12:44.842 And the way that Scott does that is by explaining how such systems came about at the highest level, at the state level, right? 00:12:45.403 --> 00:12:49.082 How they came about and why they came about. 00:12:51.158 --> 00:13:12.755 So the passage that I read, which was obviously dealing with how we manage land use and the taxation of land-based resources, one of the points he's making is that parceling land so that it can be accounted for is not something that exists in nature. 00:13:13.525 --> 00:13:14.145 That's layered. 00:13:14.346 --> 00:13:17.306 It's an abstraction layered by people. 00:13:18.096 --> 00:13:20.596 This is something else that stuck with me in the, book. 00:13:21.155 --> 00:13:30.096 Just the fact that we have surnames is something that did not emerge from a natural human society. 00:13:30.216 --> 00:13:35.806 Surnames were imposed by the state in order to tax people so that we could run a census. 00:13:36.456 --> 00:13:48.785 And it wasn't until human societies had a certain scale and they had the necessary information systems that they could aspire to manage a large population. 00:13:48.826 --> 00:14:04.225 And when you do that, you have to make population or the land or what have you legible, and that requires collapsing it to comparable measures. 00:14:04.926 --> 00:14:07.885 And that's what that little passage I read was about, right? 00:14:07.926 --> 00:14:23.011 Like, the land might be a source of multiple things, but if you're going to make it manageable, particularly in a world before computers, you're just gonna care about like how many trees can be grown there or how much copper can be extracted or what have you. 00:14:23.461 --> 00:14:34.292 And you collapse this very rich environment into this abstraction that you can measure and therefore you can manage. 00:14:34.961 --> 00:14:36.892 And in so doing, you lose a lot. 00:14:37.751 -->

00:14:47.831 And I wanted to discuss this in the light of these hugely disruptive bottom-up experiments that are happening in organizations. 00:14:48.522 --> 00:15:04.711 And what I expect is the next phase, which is how do we balance the potential innovation that's coming through these experiments with something that is recognizably manageable to the people trying to steer the ship. 00:15:07.447 --> 00:15:07.908 Yes. 00:15:09.586 --> 00:15:51.346 The two things that jump to mind upon hearing you reflect on this are, something that I know you're very familiar with, which is Stewart Brand's pace layers and another thing is the April 2011 Harvard Business Review article called The F Word, which is reframing failure, and helps provide, I think, an excellent model and taxonomy for thinking about how today, if you step back, failure is a pejorative. 00:15:52.736 --> 00:15:54.407 When you fail, it is bad. 00:15:55.106 --> 00:15:59.206 Versus in an experimental frame, failure is just feedback. 00:15:59.986 --> 00:17:04.287 And The F Word does an extraordinarily good job, I think, best I've seen, in all of the things that I've read and run across, of helping in an organizational or business context, place the word failure in a context where you can deconstruct it and apply it in more generative ways so that you can look at more experimental versus less experimental environments, more innovation or creative environments versus more mission-critical environments, and look at how"failure," or AKA feedback, is or is not appropriate and what you might do to take advantage of the sort of inevitable nature of failure, which is inevitable and out of our control, and apply it in ways that are more valuable. 00:17:04.326 --> 00:17:05.136 So that was one. 00:17:05.747 --> 00:17:21.826 And the other was, of course, the pace layers model, where in Brand's book, How Buildings Learn, he unpacks that at the most foundational layers, things change very slowly in a building, and at the outer layers, things change relatively quickly. 00:17:23.106 --> 00:17:44.287 The paint and the trim can change relatively easily and quickly versus the walls change more slowly versus the electrical infrastructure, HVAC tend to change even more slowly, and you get down to the foundation and probably it doesn't change much at all unless there is a major renovation taking place. 00:17:44.707 --> 00:17:51.586 But these are all ways of thinking about these kinds of structures. 00:17:51.615 --> 00:17:57.446 And then turning it around, of course, I was recalling, like in the advent of the telephone, there was one... 00:17:57.846 --> 00:18:00.497 You picked up the phone and somebody was on the other end. 00:18:01.119 --> 00:18:03.619 And there was no number to dial right? 00:18:04.059 --> 00:18:07.779 Because there was only a point-to-point connection at the advent of the telephone. 00:18:07.819 --> 00:18:09.160 And now we have 11 digits. 00:18:09.200 --> 00:18:19.740 And it's gone through this evolution of one digit, call the operator, to two digits, which was quite common a long time ago, to seven digits. 00:18:19.839 --> 00:18:23.299 Now it's seven plus three plus a country code, so we have 11 digits. 00:18:23.740 --> 00:18:29.355 And it's this model for abstracting away the underlying details. 00:18:29.376 --> 00:18:35.256 And now we just tap a thing on our phone that has a name, and underneath it is all of that detail, right? 00:18:36.407 --> 00:18:43.804 The primary reason why I brought this reading is I'm interested in the legibility of what is happening, right? 00:18:43.903 --> 00:18:47.964 So, if you want to manage something, you need like you were saying, you need feedback. 00:18:48.045 --> 00:18:57.066 You need to know how the system is responding to whatever decisions you've made, incentives you've created, et cetera. 00:18:57.205 --> 00:19:07.247 And the stage that we're in, at least for many organizations, this is one of those not evenly distributed problems. 00:19:07.717 --> 00:19:17.166 But for at least some of the folks that I've been talking to, we seem to be moving from a stage where the incentives have been around adoption of the technology. 00:19:18.527 --> 00:19:26.267 And I'm sure that you've seen people boasting, like they have competitions internally as to who's using the most tokens. 00:19:29.376 --> 00:19:40.769 That feels to me like that's the unit of measurement you would want to have visibility on if what you care about is usage of a particular technology. 00:19:41.809 --> 00:19:47.192 And I think that many organizations, they might be shifting away from that model. 00:19:47.893 --> 00:19:52.311 And the question is like, what is the visibility model that we need to move toward? 00:19:53.491 --> 00:19:59.093 Is it how many new services we are creating thanks to this technology? 00:19:59.152 --> 00:20:07.198 Is it the degree to which we are increasing our customer satisfaction scores? 00:20:07.278 --> 00:20:07.838 Is it... 00:20:09.128 --> 00:20:26.709 I'm wanting to move to a point where the things that we are measuring are more relevant to actual business goals as opposed to mere adoption of a new technology. 00:20:27.750 --> 00:20:35.528 And I'm hearing from folks that I'm talking with, there's another complication that this is happening in an environment of anxiety. 00:20:36.086 --> 00:20:37.145 You talked about failure. 00:20:38.665 --> 00:20:59.961 Part of the anxiety here is that for a lot of folks in these organizations, they're in a damned if they do, damned if they don't situation because if what is being measured is the adoption of the technology, and they're being incentivized to have their teams adopt the technology, then that's the number they wanna drive up. 00:21:00.101 --> 00:21:04.461 But then the media is telling them that this is a technology that is going to eliminate jobs. 00:21:05.121 --> 00:21:20.181 So it's,"Okay, so we're succeeding in deploying in an organization this thing that is

likely going to replace us," and that, that creates an environment that is, let's say like the opposite of joyful. 00:21:21.060 --> 00:21:23.641 I can only imagine that it's really stressful, right? 00:21:24.821 --> 00:21:37.849 So anyway, visibility, like part of this book is about visibility and how do you take a really complex and unmanaged situation and abstract it in such a way that you can manage it. 00:21:37.890 --> 00:21:40.789 And I think that we have to be really mindful about how we do that 00:21:41.499 --> 00:21:51.518 There are so many examples of measuring one side without identifying and then measuring the countervailing metric. 00:21:53.219 --> 00:22:08.997 I think about, I really don't remember if it was Southwest or one of the epic examples of trying to get to on-time status with number of airplanes, and what they missed was the number of bags that were getting left behind or something. 00:22:09.376 --> 00:22:11.416 And customer satisfaction went off the chart. 00:22:11.416 --> 00:22:15.337 So it was like, what is the countervailing metric for anything that you're doing? 00:22:15.376 --> 00:22:21.846 And then there was, the other epic one, which I remember conceptually, but not specifically, around trying to... 00:22:22.468 --> 00:22:22.807 What was it? 00:22:22.866 --> 00:22:28.326 Trying to get rid of, trying to get rid of snakes in India or something, and so that... 00:22:28.346 --> 00:22:29.846 Yes, the Cobra effect. 00:22:30.957 --> 00:22:32.146 Yeah, people started breeding them. 00:22:33.326 --> 00:22:37.477 And getting clear about what's on both sides of the equation. 00:22:37.477 --> 00:22:46.346 And if joy is a thing that we need to start measuring, then what are the leading indicators of that? 00:22:46.406 --> 00:22:51.267 And how do we start paying attention to that when we're measuring tokens on one side? 00:22:52.598 --> 00:22:56.259 Yeah, I don't want to suggest that joy is what the business is gonna be after. 00:22:56.318 --> 00:22:58.298 I think that would be naive. 00:22:59.398 --> 00:23:12.818 But I'm saying that the choice of measures, the choice of what we are incentivizing and how we're going to be tracking progress is being done in an environment that is not... 00:23:13.199 --> 00:23:15.859 in a lot of our conversations we talk about acting... 00:23:16.979 --> 00:23:19.759 seeing clearly and responding skillfully. 00:23:20.499 --> 00:23:27.179 And I think that we're doing this, we're trying to determine what to measure and what we're driving toward. 00:23:27.582 --> 00:23:32.442 I'm just trying to acknowledge that we're doing it in an environment of high anxiety for people. 00:23:33.082 --> 00:23:37.031 And it's very important that we get this right because we want to... 00:23:39.862 --> 00:23:54.761 It's almost I'm asking the questions like, make sure that you're clear on what it is that you're looking to gain traction on'cause, you might discover that you have gained traction, but you have gained traction on the wrong thing, right? 00:23:54.801 --> 00:24:12.461 So like if your manager is asking you to measure how many people are using AI and I don't know, how many agents are being created or what have you, it's like I don't know that's the long term, that's the best measurement of something that will ultimately create success for the business. 00:24:14.060 --> 00:24:29.770 Just understand the broader context in which you're being asked to do this and try to do right by what the business needs, but also maybe what your society needs more broadly beyond that. 00:24:31.092 --> 00:24:40.031 And like Marshall Goldsmith, who wrote What Got You Here Won't Get You There, asserts is that we need to be measuring effort over outcomes. 00:24:40.371 --> 00:24:45.056 And, I think the corollary to that is I think we need to be measuring progress over productivity. 00:24:45.855 --> 00:24:47.953 And, maybe joy was the bad example. 00:24:48.115 --> 00:24:52.998 Maybe productivity is actually the thing that people used to focus on. 00:24:53.018 --> 00:24:55.298 But progress, how do you measure progress? 00:24:56.038 --> 00:24:57.898 Is it a function of a zero-sum system? 00:24:57.958 --> 00:24:58.558 I hope not. 00:24:59.458 --> 00:25:05.478 But I know we've gotta wrap it today, and this is a very important topic. 00:25:06.738 --> 00:25:18.638 I'm definitely gonna go back and reread that book in my copious spare time, and I so appreciate that you brought it today, and I think anybody that's willing to invest the time in reading this book will see the world a different way when they're done with it. 00:25:19.073 --> 00:25:19.874 I'm not exaggerating. 00:25:19.913 --> 00:25:21.756 It is one of the most important books I've read. 00:25:22.176 --> 00:25:24.217 it certainly informed my worldview deeply. 00:25:24.996 --> 00:25:32.175 We do have more to talk about in this, and you've mentioned several books that I need to check out that might help us deepen this conversation. 00:25:32.237 --> 00:25:32.896 So thank you, Harry. 00:25:33.816 --> 00:25:34.017 Yeah. 00:25:34.076 --> 00:25:34.635 Thank you, Jorge. 00:25:37.201 --> 00:25:41.641 Thank you for listening to Traction Heroes with Harry Max and Jorge Arango. 00:25:42.300 --> 00:25:49.980 Check out the show notes at tractionheroes.com and if you enjoyed the show, please leave us a rating in Apple's podcasts app. 00:25:50.340 --> 00:25:50.790 Thanks.