Bản chép lời đầy đủ của PAPod 611 - Command–Adapt Paradox: Rethinking Safety for a Riskier World - David Woods - PART TWO

a percent improvement Not an order 00:04:54.609 --> 00:04:55.729 of magnitude improvement. 00:04:55.729 --> 00:05:00.389 And we're looking at harm that is two orders of magnitude. And as I said to 00:05:00.389 --> 00:05:03.499 you last time, measles is a simple example. 00:05:03.667 --> 00:05:08.834 And by the way, when I talked to you last time, we're now in the seventh month, 00:05:09.467 --> 00:05:11.799 right? And we have set a record, right? 00:05:12.339 --> 00:05:15.199 We are way past the max number. 00:05:15.759 --> 00:05:19.190 We're way on our way to three orders of magnitude. We might hit, 00:05:19.579 --> 00:05:22.836 we're definitely going to hit two and a half orders of magnitude in measles cases. 00:05:24.779 --> 00:05:32.849 As a canary in the coal mine on how the exposure to harm is increasing at the 00:05:32.849 --> 00:05:35.402 same time that our systems grow in capability. 00:05:35.728 --> 00:05:37.231 How will we break that cycle? 00:05:38.697 --> 00:05:42.256 Well, the irony is we can rock the boat with a science base. 00:05:43.509 --> 00:05:48.222 But it is kind of the answer in a way, right? Or it's a vital part of the answer. 00:05:48.851 --> 00:05:54.662 Yes. I mean, we have to have a generalizable set of what goes on. 00:05:55.869 --> 00:06:01.359 Because the normal context and pressures and constraints in which we do practical 00:06:01.359 --> 00:06:03.311 safety programs, they're very narrow. 00:06:03.590 --> 00:06:05.860 We get lost in the details of every setting. 00:06:06.295 --> 00:06:08.286 And it's driven very reactively. 00:06:08.849 --> 00:06:12.919 Even the, you know, the, our proactive, we get proactive successes, 00:06:12.919 --> 00:06:14.858 but it's hard to sustain. 00:06:15.217 --> 00:06:19.889 It's hard to have them organically multiply. It's a uphill climb each time we 00:06:19.889 --> 00:06:22.897 pull one off and that doesn't scale. 00:06:23.263 --> 00:06:26.670 And we've seen that when, if you take the long look, you go, yeah. 00:06:27.117 --> 00:06:32.999 So I want to go back to this issue of everything in the human experience pretty 00:06:32.999 --> 00:06:35.244 much is encompassed in the label safety, 00:06:35.859 --> 00:06:39.159 success, failure, reward, harm, fear, blame, responsibility, 00:06:39.159 --> 00:06:44.301 ethics, courage, disaster, heroic saves, the drama of the very near miss, right? 00:06:44.746 --> 00:06:47.765 It covers the C-suite. It covers regulatory policy. 00:06:48.509 --> 00:06:52.949 It goes all the way down to people who have to operate critical systems under 00:06:52.949 --> 00:06:56.868 enormous pressures to do amazing things at the sharp end. 00:06:57.488 --> 00:07:02.669 We cover a new technology that can do more and extend human reach, 00:07:02.669 --> 00:07:06.273 human vision, human action to new scales and distances. 00:07:06.911 --> 00:07:08.786 They have an enormous range of things. 00:07:09.512 --> 00:07:16.609 But in the end, you know, a mantra for proactive

safety that I articulated 26 00:07:16.609 --> 00:07:24.269 years ago, been in print 21 years, is this our fundamental obligation as safety people, right? 00:07:24.269 --> 00:07:28.199 If you're engaged in creating safety, our fundamental obligation is to create 00:07:28.199 --> 00:07:32.645 foresight about the changing shape of risk before harm occurs. Mm-hmm. 00:07:33.774 --> 00:07:38.083 This came out first in the patient safety movement around 2000, 00:07:38.083 --> 00:07:43.024 and then very significantly in the Columbia space shuttle accident. 00:07:44.655 --> 00:07:54.843 And so now those two things, this huge encompassing of human systems issues and this obligation, 00:07:55.450 --> 00:08:01.847 for foresight about the changing shape of risk, means we've got a lot. 00:08:02.347 --> 00:08:06.625 I mean, what scope? I mean, what fun we get to have over the years, right? You and me. 00:08:07.234 --> 00:08:09.873 All the places we get to go, all the things we get to interact, 00:08:09.873 --> 00:08:13.803 all the scary places we've been, right? That's what we talk about over beers. 00:08:14.903 --> 00:08:19.837 Who went to the scariest place and came out alive? 00:08:20.875 --> 00:08:28.763 The so the key is building science base but that science base is about it has 00:08:28.763 --> 00:08:31.762 these funny constraints on it how do you do, 00:08:32.598 --> 00:08:36.543 lab how do you do what science people when people say science they usually think 00:08:36.543 --> 00:08:41.753 of lab experiments and definitive results with laws and you go big laws to hold 00:08:41.753 --> 00:08:43.215 everything up yeah i'm with you, 00:08:43.846 --> 00:08:48.920 and well when safety encompasses so much of human experience you have to look at the real world, 00:08:49.574 --> 00:08:54.393 and how do you and are you understanding human aspirations are you understanding 00:08:54.393 --> 00:08:57.793 why we build technology why we operate in certain ways why we're willing to 00:08:57.793 --> 00:09:02.563 invest in some aspects more automation less people maybe sometimes we should 00:09:02.563 --> 00:09:04.458 grow human expertise and have 00:09:04.847 --> 00:09:08.123 and maybe that's a better investment in the long run sometimes, 00:09:09.103 --> 00:09:11.244 always be one way or the other right. 00:09:12.462 --> 00:09:18.896 The and and and so you end up with this thing of we can't really do what our 00:09:18.896 --> 00:09:21.046 social norm of what science is. 00:09:21.046 --> 00:09:25.076 We have to do something that looks at the real world. So all of those practical 00:09:25.076 --> 00:09:28.716 experiences you and I have been in, we've treated we were just joking about 00:09:28.716 --> 00:09:32.182 it as research opportunities. Right. We're not just helping them. 00:09:32.634 --> 00:09:36.924 Each time we do it, we're going, I'm looking for patterns and understanding, 00:09:37.499 --> 00:09:40.886 whether that's at the sharp end, whether that's about the technology or the 00:09:40.886 --> 00:09:46.178 human technology interaction. whether that's about the organizational beliefs, how do you move them? 00:09:46.636 --> 00:09:49.388 How do you change their models and biases? 00:09:50.227 --> 00:09:55.286 I mean, another one is that makes this hard and different than usual areas of 00:09:55.286 --> 00:10:00.016 science is how do you do science when the people trying to do the science are 00:10:00.016 --> 00:10:02.887 part of the processes the scientists are trying to study? 00:10:03.445 --> 00:10:06.394 If you're trying to understand safety, you're influencing safety. 00:10:06.748 --> 00:10:12.076 You can't study it without being related to the process that's trying to build 00:10:12.076 --> 00:10:13.834 safety or sustain safety. 00:10:14.142 --> 00:10:16.220 You're studying how it breaks, right? 00:10:16.551 --> 00:10:21.398 You're part of the thing you study. You're not an outside objective observer. 00:10:21.538 --> 00:10:24.556 You're in something that's changing all the time. 00:10:25.172 --> 00:10:31.419 So the third thing is creating safety is about all kinds of skills, 00:10:31.791 --> 00:10:37.196 but they're all about balancing competing interests, right? In a changing world, 00:10:37.196 --> 00:10:39.165 not a state, not a category. 00:10:39.531 --> 00:10:44.146 It's about rebalancing how activities need to shift relative to goals, 00:10:44.146 --> 00:10:51.519 how goals need to shift relative to risk, how resources need to shift relative to uncertainty and risk, 00:10:52.406 --> 00:10:54.865 and new technology and new opportunities. 00:10:55.753 --> 00:11:01.791 It's a dynamic balancing And it's very hard To get organizations In today's world, 00:11:02.633 --> 00:11:06.216 That's why I use a lot of military examples Because they have to be good at 00:11:06.216 --> 00:11:12.953 balancing Dynamically Otherwise they lose Even if they start in the best position And they stay there. 00:11:14.177 --> 00:11:18.356 The best position doesn't stay best because there's somebody else adapting to. 00:11:19.074 --> 00:11:26.474 And so that gets us to this current situation we've been talking about of people 00:11:26.474 --> 00:11:28.704 looking back at 50 years and saying, 00:11:29.291 --> 00:11:34.374 yeah, it's just a lot of noise, a lot of labels, not much progress. Sloganeering. 00:11:34.852 --> 00:11:38.589 Pre-scientific sloganeering, right? 00:11:39.222 --> 00:11:43.414 And I think we want to blow away any confusion That there is, 00:11:43.414 --> 00:11:46.774 in fact, fantastic progress And, 00:11:47.834 --> 00:11:52.567 that the world keeps changing And that progress is necessary And needs to be, 00:11:53.190 --> 00:11:55.036 sustained and sped up even, 00:11:56.086 --> 00:12:01.174 And that it's easy, given the pressures Limited resources, huge variety of situations 00:12:01.174 --> 00:12:06.413 And distractions we have to face in safety It's easy not to know a lot about what we know,

00:12:07.734 --> 00:12:08.627 It's hard to keep up. 00:12:08.952 --> 00:12:12.290 It's hard to be part of this. It's hard to synthesize. 00:12:14.149 --> 00:12:17.574 God, that's interesting. You know that's an incredibly interesting thing you 00:12:17.574 --> 00:12:19.257 just said. That's remarkable. 00:12:20.894 --> 00:12:24.894 I mean, so we're getting a little bit to some of the punchlines. 00:12:24.894 --> 00:12:30.834 One is got to put effective and safe, safe and effective up front in order to 00:12:30.834 --> 00:12:34.594 be viable in both the short run and the long run, to be adaptive both in the 00:12:34.594 --> 00:12:35.754 short run and the long run. 00:12:36.374 --> 00:12:42.494 That are put faster, cheaper first assumes a stable world. Right. 00:12:43.325 --> 00:12:47.784 Where it's just a quick race. But when the landscape changes and the rules change 00:12:47.784 --> 00:12:52.784 and the destination changes and the trophy at the end changes, you got to adapt. Yeah. 00:12:53.974 --> 00:12:59.342 So what I want to highlight with you and encourage everybody is to spend some 00:12:59.342 --> 00:13:05.462 more time investigating, arguing, challenging, revising, extending, limiting. 00:13:05.462 --> 00:13:09.002 I'm not even going to say the scope of what we know is more limited than we 00:13:09.002 --> 00:13:10.678 thought, building more. 00:13:11.061 --> 00:13:15.362 But I want to give a couple examples that we have generalizable findings that 00:13:15.362 --> 00:13:19.744 are both concrete to guide how to work in a particular situation. 00:13:20.069 --> 00:13:23.610 Right. One general generalization doesn't apply everywhere. 00:13:23.964 --> 00:13:30.251 But it applies in lots more places. In fact, the reason we know it's a generalization or a law even, 00:13:30.790 --> 00:13:35.732 or a theorem even more that we can prove, is because we see it in many different 00:13:35.732 --> 00:13:40.102 settings play out, even though there are differences because of the unique features 00:13:40.102 --> 00:13:42.627 of those settings or the layer at which we observe it. 00:13:43.461 --> 00:13:49.816 And so I want to give one, right? It's really simple. It's a short paper under 4,000 words. 00:13:50.170 --> 00:13:52.185 It's a synthesis of syntheses. 00:13:52.800 --> 00:13:58.642 It's called the Command-Adapt Paradox Paper And it was written in frustration, 00:13:59.002 --> 00:14:04.372 I admit it Because people kept coming back and saying Well, what about compliance? 00:14:04.372 --> 00:14:09.572 And how much pressure for compliance should we be putting on our subordinate 00:14:09.572 --> 00:14:15.992 managers and personnel To just comply with the policy, to comply with the plans, 00:14:15.992 --> 00:14:17.605 to comply with the rules? 00:14:18.372 --> 00:14:22.314 And I kind of went, oh my God Are we re-litigating this again? 00:14:22.633 --> 00:14:26.012 How many times have we re-litigated this? Why are we doing this? 00:14:26.012 --> 00:14:29.942 This is an example of us We're in the washing machine going around and around 00:14:29.942 --> 00:14:33.186 again I'm tired of getting an ounce in your head off the wall You know, 00:14:33.985 --> 00:14:38.482 And so I reviewed the papers and syntheses To remind people about the basics 00:14:38.482 --> 00:14:45.339 And tried to summarize the synthesis of the syntheses You know, in plain, plain terms, 00:14:46.280 --> 00:14:51.074 compliance says there's plan and conform we often turn the plan into automation 00:14:51.074 --> 00:14:53.251 to try to make sure there's conformance right, 00:14:53.899 --> 00:14:59.554 right and the pressure to plan to behave by plan and conform reduces adaptive 00:14:59.554 --> 00:15:06.717 capacity this is laid out and a variety of things and i went through them in a ways like. 00:15:07.744 --> 00:15:12.844 Here are i had it up but anyway you should go look at it and i'll put the paper 00:15:12.844 --> 00:15:14.614 i'll put the paper in the notes for this. 00:15:14.614 --> 00:15:20.291 You know, it's like one of the findings, one, two, three, four, finding one. 00:15:20.645 --> 00:15:23.048 Okay. Well, the, the modern, right. 00:15:25.120 --> 00:15:30.494 There will always be surprised to challenge your plans. The modern date on that 00:15:30.494 --> 00:15:32.882 is in the 1980s on that finding. 00:15:33.387 --> 00:15:39.190 How far does it go back? It goes back to 1836 in Clausewitz. 00:15:40.741 --> 00:15:49.615 Another one is people outside models of work are miscalibrated. 00:15:50.684 --> 00:15:56.584 Your model of how someone else confronts the world is not the same of the model 00:15:56.584 --> 00:16:00.249 of the person confronting uncertainty, complexity in the world. 00:16:01.909 --> 00:16:08.434 That model confusion really dates back to again, some developments and papers 00:16:08.434 --> 00:16:12.660 and findings in our 50-year history. 00:16:13.832 --> 00:16:18.168 But you know how far it goes back? It goes back to 1890. 00:16:19.892 --> 00:16:24.347 William James, and he called it the psychologist fallacy. 00:16:25.067 --> 00:16:30.393 When the person studying human behavior substitutes their model for the situation, 00:16:30.944 --> 00:16:34.456 for the person's model of the situation, they have to act it. 00:16:35.136 --> 00:16:38.767 And he says, your job is to discover the model the person brings, 00:16:38.767 --> 00:16:42.287 not because it's right, but because they'll behave according, 00:16:42.287 --> 00:16:44.853 that model will influence their behavior. 00:16:45.387 --> 00:16:48.272 And then you can start to see what they're missing, what they're getting right. 00:16:48.568 --> 00:16:51.320 And you'll see that both of your models are wrong. 00:16:52.069 --> 00:16:56.767 Both of your models are incomplete and the risk of miscalibration exists, 00:16:56.767 --> 00:17:00.467 whether the person acting in the world and the person trying to understand the 00:17:00.467 --> 00:17:01.891

person acting in the world. 00:17:03.048 --> 00:17:09.267 And that's in Behind Human Error. That's in the 1983 paper at the Bellagio, 00:17:09.887 --> 00:17:14.792 the first formal official meeting on going behind the label human error in 1983. 00:17:16.287 --> 00:17:21.237 Those were, that fundamental is there. And that comes out in how we deal with 00:17:21.237 --> 00:17:24.767 hindsight bias and the techniques and a whole bunch of different things. 00:17:24.767 --> 00:17:26.335 And we can go on and on in there. 00:17:26.989 --> 00:17:31.331 Very quickly, we see research coming out. And again, we could date it earlier, 00:17:31.575 --> 00:17:37.833 but 1986, Lucy Suchman, plans are resources for action. Plans aren't specifications of action. 00:17:38.359 --> 00:17:41.348 All right. And then we go through how that developed. 00:17:42.153 --> 00:17:47.694 And what that means and what you end up with, oh, and that plans are necessary 00:17:47.694 --> 00:17:51.084 to recognize anomalies and surprise, right? 00:17:51.084 --> 00:17:54.304 If I don't have a model of what's typical, what's supposed to work, 00:17:54.304 --> 00:17:58.625 how things usually work, how we usually handle disruption and variation, 00:17:58.811 --> 00:18:01.464 I can't recognize anomalies, right? 00:18:01.464 --> 00:18:04.914 And therefore, I can't revise models. I can't revise tactics. 00:18:04.914 --> 00:18:06.764 I can't learn new strategies. 00:18:07.037 --> 00:18:10.560 I don't understand and diagnose the anomalies and where they come from. 00:18:11.202 --> 00:18:16.234 So learning and change comes from recognizing anomalies And to do that, 00:18:16.234 --> 00:18:21.651 I need a baseline of typicality that comes from plans So plans help me revise, 00:18:22.364 --> 00:18:27.075 And that's an old finding Eisenhower said that one in the 1950s. 00:18:27.944 --> 00:18:33.394 As well as scientists in the 80s and 90s So that led us to the idea of It's 00:18:33.394 --> 00:18:39.121 not plan and conform, it's plan and revise But revision isn't freelance, 00:18:39.578 --> 00:18:43.414 So then we run through what's the interaction between, right, 00:18:43.414 --> 00:18:46.569 what do we want to achieve? How are we trying to achieve it? 00:18:47.167 --> 00:18:52.826 Variability, surprise, and opportunity will arise outside the boundaries of that scope. 00:18:53.268 --> 00:18:58.132 And how do we revise relative to that? How do we recognize, revise, 00:18:58.319 --> 00:19:03.286 reframe, revitalize, right, head in new directions? 00:19:03.594 --> 00:19:06.544 And we've seen that now that we've lived decades in this. 00:19:06.544 --> 00:19:11.184 We see that over and over again Where safety programs, safety innovations Were 00:19:11.184 --> 00:19:16.884 turned into programs The programs, these were all fights The programs eventually 00:19:16.884 --> 00:19:18.646 were widespread in the industry, 00:19:19.335 --> 00:19:22.630 Even transferred to other industries And then you go and look. 00:19:23.384 --> 00:19:27.820 A while later, and you recognize they're not fit for purpose anymore. 00:19:28.473 --> 00:19:33.781 They're not doing the job. They've lost vitality, right? They need to be revitalized. 00:19:34.182 --> 00:19:38.744 Some of those were the successes in teamwork at crew resource management. 00:19:38.744 --> 00:19:43.782 And then you go into places that have crew resource management or other teamwork programs and you go, 00:19:44.419 --> 00:19:47.604 yeah, but when you actually look at the organization closely, 00:19:47.604 --> 00:19:52.221 you're finding a lot of breakdowns in teamwork, a lot of working across purposes, 00:19:52.581 --> 00:19:57.144 a lack of synchronization over different organizational boundaries within, 00:19:57.144 --> 00:19:58.644 you know, within the organization. 00:19:58.644 --> 00:20:01.416 And you go, there's not great teamwork here, guys. 00:20:01.851 --> 00:20:05.954 You need to revitalize the program and expectations around this. 00:20:05.954 --> 00:20:11.657 And it become routinized and rote and fine and a distraction. 00:20:12.034 --> 00:20:16.376 I get to go off and take my team training, but nothing effective came from it. 00:20:17.022 --> 00:20:21.824 So we have to re-energize the pursuit of safety We have to revitalize That should 00:20:21.824 --> 00:20:27.204 be seen as a basic in safety Not only do you need to You're in a place of. 00:20:28.984 --> 00:20:33.427 Re-energizing, revitalizing safety We want to do this before bad accidents happen 00:20:33.747 --> 00:20:38.324 But down the road, it's not one and done It's a phase And then you're going 00:20:38.324 --> 00:20:43.164 to have to re-energize and revitalize And some of that will be on different aspects, 00:20:44.144 --> 00:20:49.074 That weren't critical before And, you know, were okay to handle With the other 00:20:49.074 --> 00:20:52.934 programs you had But now needed to be more the focus and central thing Because 00:20:52.934 --> 00:20:53.994 the world keeps changing, 00:20:54.693 --> 00:20:59.167 So anyway, so if somebody comes and says to you, it's all noise and it's pre-scientific, 00:20:59.735 --> 00:21:03.024 read the Command-Adapt Paradox paper, right? 00:21:03.024 --> 00:21:06.904 It synthesizes lots of research. It encompasses a broad set of things, 00:21:06.904 --> 00:21:10.904 and it ends up with an actionable strategy, and it's being carried out. 00:21:10.904 --> 00:21:15.524 Guided adaptability is being used by pragmatic people working with everyday 00:21:15.524 --> 00:21:20.344 organizations like Dave Provan and Bethleh and their book coming out on this, 00:21:20.785 --> 00:21:25.713 and they're turning it into practical action so that you can meet those demands. 00:21:25.992 --> 00:21:28.889 Do it now, do it fast, and don't cost me money. That's right. 00:21:29.624 --> 00:21:31.095 And make it better, make it better. 00:21:31.904 --> 00:21:36.134 What a great conversation. Thank you, my friend. Well,

the other ones, 00:21:36.134 --> 00:21:41.244 I mean, there's other ones I want to mention quickly to remind people of this 00:21:41.244 --> 00:21:42.829 before we bring it to a close. 00:21:43.724 --> 00:21:48.164 Just to highlight to people, I already mentioned the fluency law and now that 00:21:48.164 --> 00:21:50.440 goes back to the psychologist fallacy. 00:21:51.349 --> 00:21:56.028 But the fluency law and all that stuff is critical to our analysis of how things 00:21:56.028 --> 00:21:58.048 go right and how things can go wrong. 00:21:58.645 --> 00:22:02.388 You have to go in and understand how people are adapting. And to do that, 00:22:02.388 --> 00:22:06.134 that's where in resilience engineering, we said study how things go right. Why? 00:22:06.511 --> 00:22:10.198 Because people are doing fluent adaptations to make things go right. 00:22:10.198 --> 00:22:12.129 So go study how things go right. 00:22:12.558 --> 00:22:17.088 So make sure you understand the fluency law. It's critical out there in the 00:22:17.088 --> 00:22:19.780 last 25 years of how we pursue safety. 00:22:20.308 --> 00:22:24.858 And forget the big labels and arguments over this And putting numbers on it, 00:22:24.858 --> 00:22:28.980 whatever There's a fundamental law, right? 00:22:29.236 --> 00:22:34.958 Now, it's an empirical law But we can derive this from some first principles 00:22:34.958 --> 00:22:39.545 That gives it extra strength It's not just empirical that we see it a lot It 00:22:39.777 --> 00:22:41.676 kind of has to work this way Right. 00:22:42.958 --> 00:22:46.808 And so then when you understand it You can make use of it in a variety of settings 00:22:46.808 --> 00:22:50.628 To help understand and direct your action and also understand how sometimes 00:22:50.628 --> 00:22:52.045 people get on the wrong track. 00:22:52.702 --> 00:22:56.468 The other one is, this one came, there were some things anticipating it, 00:22:56.468 --> 00:23:00.911 but it was first, the label was first used in 98, it's the law of stretch systems. 00:23:01.513 --> 00:23:05.621 Every system is stretched to operate at its capacity as soon as there is some improvement. 00:23:06.638 --> 00:23:10.318 When a new form of technology comes along, the capabilities will be exploited 00:23:10.718 --> 00:23:13.291 to achieve a new intensity and tempo of activity. 00:23:13.925 --> 00:23:20.029 The statement of the law is actually a paraphrase of an after-action report from a U.S. 00:23:20.319 --> 00:23:27.708 Military operation that was Kordsman, Anthony Kordsman, summarized the use of 00:23:27.708 --> 00:23:30.682 new modern military technology in the Gulfs. 00:23:31.618 --> 00:23:35.454 To do more, do it faster in more complex ways. It didn't make the work on the, 00:23:36.054 --> 00:23:37.580 sharp end operator easier. 00:23:38.160 --> 00:23:41.364 It enabled them to do more, do it faster in more complex ways. 00:23:41.364 --> 00:23:45.811 What did that mean? You could do more. It was good. A

capability also meant it was riskier. 00:23:46.578 --> 00:23:49.132 Things could go wrong. When they went wrong, harm was bigger. 00:23:49.312 --> 00:23:52.533 Right. And some of that were friendly fire incidents. 00:23:53.642 --> 00:23:58.994 So notice this highlights how everybody's adapting and how when people say we're 00:23:58.994 --> 00:24:05.324 going to improve the system in this way, I'm sorry, but that's not what's going to happen. 00:24:06.008 --> 00:24:09.824 Do I know what's going to happen? No. What I know is what we think is going 00:24:09.824 --> 00:24:11.774 to happen is not what's going to happen. 00:24:12.554 --> 00:24:17.144 Because people are adaptive. And what they'll see as an advantage is different 00:24:17.144 --> 00:24:19.013 than what we thought it would be. 00:24:19.517 --> 00:24:24.234 We'll think more narrowly and linearly. They'll see it relative to their goals. 00:24:24.234 --> 00:24:26.124 Remember that psychologist fallacy. 00:24:26.443 --> 00:24:31.464 They'll see it in a new light. And they'll go, look what this will do for me. 00:24:32.156 --> 00:24:37.104 And that means maybe it won't do something so good for a different role at a 00:24:37.104 --> 00:24:40.103 different layer of the operation or organization. 00:24:40.798 --> 00:24:44.694 And we see this in the adaptations around AI coding. 00:24:46.414 --> 00:24:51.144 We just see it all the time in all of this. one that is really, 00:24:51.144 --> 00:24:56.054 really important and has been there since 19. 00:24:56.054 --> 00:25:01.230 I first said it in 83 at the Bellagio meeting, and then it's in the 94 Behind Human Error, is how, 00:25:01.843 --> 00:25:05.934 hindsight, knowledge of outcome changes your perspective and changes your ability 00:25:05.934 --> 00:25:11.224 to understand what happens when people confront situations with uncertainty and risk. 00:25:11.912 --> 00:25:15.499 And so I gave it a new name a few years ago, the novelty inequality. 00:25:16.109 --> 00:25:20.994 I think this is a mathematically provable situation. I have not found a partner 00:25:20.994 --> 00:25:23.644 to help me pull off a mathematical Bayesian proof. 00:25:24.997 --> 00:25:30.001 Well, you know, I've always argued with Bayes. Now I want to use Bayes to help me. 00:25:30.334 --> 00:25:33.820 I haven't quite found a way to do this. Bayes is kind of intransigent. 00:25:33.820 --> 00:25:37.460 You know, I'm not sure I'm on your side, Dave. You've criticized me so much. 00:25:38.990 --> 00:25:44.610 But the novelty inequality, when you're in a situation facing uncertainty and 00:25:44.610 --> 00:25:51.662 risk, the potential novelty in front of you is much, much greater than the known novelty, 00:25:52.179 --> 00:25:54.379 that's determined after the fact. 00:25:55.182 --> 00:25:59.472 Right. Afterwards, we can hone in on this is exactly what was different. 00:26:00.064 --> 00:26:04.520 This is exactly where the rules didn't apply. This is exactly the surprising 00:26:04.520 --> 00:26:08.951 nature. This is exactly what needed to be adapted relative to normal practice. 00:26:09.578 --> 00:26:14.400 But beforehand, you're not sure There's multiple forms This was shown in the 00:26:14.400 --> 00:26:18.120 hindsight cartoon That Richard Cook came up with And we used in the patient 00:26:18.120 --> 00:26:23.840 safety movement In the 90s Based on the Behind Human Error book Trying to explain what this, 00:26:24.666 --> 00:26:30.691 meant And literally the cartoon used It's in the 2010 book You can see, 00:26:31.418 --> 00:26:36.810 The range of uncertainty factors And trade-offs that the Sharp end person confronts 00:26:36.810 --> 00:26:43.620 In the face of uncertainty and risk versus the after-the-fact analysis can hone in on only a few. 00:26:44.007 --> 00:26:47.860 And as I always joked about Three Mile Island, they had 10 minutes to make a 00:26:47.860 --> 00:26:53.684 decision for a situation they had not been trained for and didn't know even could exist. Right. 00:26:54.950 --> 00:27:03.120 When the after-action analysis took six months and $6 million of the best engineers 00:27:03.120 --> 00:27:08.060 in the field and $60 million in 1980 terms is a hell of a lot of money. 00:27:08.060 --> 00:27:09.337 Yeah, yeah, that's some big jack. 00:27:09.686 --> 00:27:13.100 And I think a lot of us would make decisions better with $60 million, 00:27:13.900 --> 00:27:18.980 even in today's dollars, and access to the best engineers or analysts on that 00:27:18.980 --> 00:27:20.726 particular topic, right? 00:27:22.102 --> 00:27:26.110 And stuff. So we have lots and lots of stuff. 00:27:26.812 --> 00:27:30.992 So we wanted to come back. You asked me, well, and this is going to go back 00:27:30.992 --> 00:27:34.181 to, we're trying to stimulate our friends. 00:27:34.663 --> 00:27:40.592 Yeah. That, you know, in a time of turbulence and change where the value of 00:27:40.592 --> 00:27:47.372 pursuing safety is in decline, the attention and willingness to invest in efforts to improve safety. 00:27:47.372 --> 00:27:51.506 In fact, it's the opposite, are faster, cheaper. 00:27:52.197 --> 00:27:55.820 Who gives a damn about what consequences play out along the way? 00:27:56.174 --> 00:27:58.524 We have to move fast and we have to move cheaper. 00:27:59.186 --> 00:28:03.268 And so how do we stand up for safety in this period of turbulence? 00:28:03.859 --> 00:28:05.212 We all need to pull together. 00:28:05.929 --> 00:28:10.660 And as you asked at the beginning, how do I deal with it and how do I personally get frustrated? 00:28:11.275 --> 00:28:16.284 Well, Voltaire said it for me in the 1700s. 00:28:17.018 --> 00:28:24.268 Mon Dieu, gardez-moi de mes amis. Quand à mes ennemis, je m'en charge. 00:28:25.132 --> 00:28:31.548 Dear, dear God, protect me from my friends. I can contend with my adversaries on my own. 00:28:32.888 --> 00:28:36.730 Beautiful. Thank you, my friend. We all need to pull together. 00:28:37.822 --> 00:28:42.466 And that is synthesize. Put safely and effective first. 00:28:43.802 --> 00:28:51.579 Build the momentum for continuous energy investment, re-energizing, revitalization. 00:28:51.933 --> 00:28:54.722 We are in a period of desperate need for revitalization. 00:28:55.323 --> 00:28:59.596 That requires innovation, that can draw on the science base. 00:28:59.926 --> 00:29:04.310 That's going to change the science base. It's going to expand the science base. 00:29:04.692 --> 00:29:08.750 We need to do it together because we are like the sharp end operator. 00:29:09.017 --> 00:29:12.892 We got to pull off something when we're under constraints. They're tying one 00:29:12.892 --> 00:29:17.121 arm behind our back, and they're trying to cut us off at the knees at the same time. 00:29:17.759 --> 00:29:22.990 We got to show a way, get some tools in there to show we can still dance, 00:29:23.204 --> 00:29:24.905 and we can help them dance too. 00:29:25.265 --> 00:29:30.992 Because we all are part of the balancing act that creates and sustains safety 00:29:30.992 --> 00:29:32.879 while we do things effectively. 00:29:33.355 --> 00:29:36.422 And that's always been the case, and it will always be the case. Thank you. 00:29:40.947 --> 00:29:45.629 And there you have it. Part two, finito, complete, terminar. 00:29:45.629 --> 00:29:47.627 It's over. We've done it. 00:29:48.178 --> 00:29:50.589 And thanks for hanging in there and listening to both parts. 00:29:50.589 --> 00:29:54.053 I know it's made a difference. It's a very interesting conversation. 00:29:54.308 --> 00:29:58.755 And I kind of hope it does rock a boat. I'll be good with it rocking a boat. 00:29:59.196 --> 00:30:02.759 But more importantly, I hope it builds a pretty strong case for we're better 00:30:02.759 --> 00:30:05.819 collectively together than we are separate. 00:30:06.486 --> 00:30:10.381 And I think that's a really important message. And that's not coming from me, baby. 00:30:10.770 --> 00:30:15.066 That's coming from Professor David Woods, the Ohio State University, the man. 00:30:15.664 --> 00:30:19.599 He's been around along 47 of the 50 years. That's a long time, 00:30:19.599 --> 00:30:21.010 dude. That is a long time. 00:30:21.428 --> 00:30:24.439 Learn something new every single day. Have as much fun as you possibly can. 00:30:24.439 --> 00:30:26.789 Be good to each other. Be kind to each other. Take care of each other, 00:30:26.789 --> 00:30:29.993 even the least among us. And for goodness sakes, be safe.