WEBVTT 00:00:00.000 --> 00:00:02.180 Welcome back to the Deep Dive. We're here to 00:00:02.180 --> 00:00:04.240 cut through the noise and get you insights that, 00:00:04.240 --> 00:00:06.639 well, really matter. And today we're tackling 00:00:06.639 --> 00:00:08.580 something that's definitely buzzing under the 00:00:08.580 --> 00:00:10.699 surface. Yeah, it's that feeling in offices and 00:00:10.699 --> 00:00:13.000 maybe even in your home office right now. That 00:00:13.000 --> 00:00:16.059 quiet, nagging question. The one everyone's asking, 00:00:16.379 --> 00:00:19.399 even if subconsciously. Is my job. Is my role. 00:00:19.699 --> 00:00:23.640 Secure. Especially with AI advancing so, so quickly. 00:00:23.800 --> 00:00:27.160 It's a genuine anxiety. And our sources today 00:00:27.160 --> 00:00:29.620 really hit this head on. The key thing. the big 00:00:29.620 --> 00:00:33.159 insight. It's that the real fight isn't about 00:00:33.159 --> 00:00:36.539 trying to be faster than AI. That's a losing 00:00:36.539 --> 00:00:39.079 game for most knowledge workers. Right. That's 00:00:39.079 --> 00:00:41.119 the trap, isn't it? This idea that we just need 00:00:41.119 --> 00:00:43.439 to be more efficient, quicker to somehow out 00:00:43.439 --> 00:00:46.439 -machine the machine. Exactly. But the sources 00:00:46.439 --> 00:00:48.789 we're diving into today are pretty clear. That 00:00:48.789 --> 00:00:51.390 is just not a winning strategy. You're never 00:00:51.390 --> 00:00:54.030 gonna beat AI at pure computation or storing 00:00:54.030 --> 00:00:57.210 facts or just following rules flawlessly. It's 00:00:57.210 --> 00:00:59.210 like trying to race a car on foot. You're competing 00:00:59.210 --> 00:01:01.570 on their turf, playing their game. Yeah, bringing 00:01:01.570 --> 00:01:04.090 a knife to a, well, a laser fight, as you said 00:01:04.090 --> 00:01:07.170 earlier. It's destined to fail. So why is that? 00:01:07.310 --> 00:01:10.170 Why is it such a losing battle? Well... It'd 00:01:10.170 --> 00:01:13.090 come down to core strengths. AI's power lies 00:01:13.090 --> 00:01:16.269 in processing massive data sets, running complex 00:01:16.269 --> 00:01:20.069 algorithms perfectly. Speed. Things humans just 00:01:20.069 --> 00:01:22.150 aren't built for at that scale. It's not about 00:01:22.150 --> 00:01:24.290 effort, then. It's about fundamental design. 00:01:24.549 --> 00:01:27.209 Precisely. It's a subtle, but really profound 00:01:27.209 --> 00:01:29.810 difference in how we operate versus how they 00:01:29.810 --> 00:01:33.049 operate. OK, so if we can't win on speed and 00:01:33.049 --> 00:01:36.129 efficiency, what's left? What's our advantage? 00:01:36.290 --> 00:01:39.370 What can humans do that machines fundamentally 00:01:39.370 --> 00:01:41.390 struggle with. Ah, now that's what gets really 00:01:41.390 --> 00:01:43.370 interesting. Our source material today, drawing 00:01:43.370 --> 00:01:46.250 from discussions around six books to outwit AI, 00:01:46.870 --> 00:01:49.989 points to a uniquely human superpower. A superpower. 00:01:50.010 --> 00:01:52.010 I like the sound of that. It's interdisciplinary 00:01:52.010 --> 00:01:54.469 thinking. Interdisciplinary thinking. Okay, unpack 00:01:54.469 --> 00:01:56.370 that a bit. Think about connecting ideas from 00:01:56.370 --> 00:01:58.890 totally different fields, like using insights 00:01:58.890 --> 00:02:01.209 from psychology to solve a tricky business problem, 00:02:01.689 --> 00:02:03.930 or applying lessons from how nature builds things 00:02:03.930 --> 00:02:06.700 biomimicry to design better systems. So it's 00:02:06.700 --> 00:02:09.379 about bridging gaps, seeing connections others 00:02:09.379 --> 00:02:13.939 miss. Exactly that. This ability to link distant 00:02:13.939 --> 00:02:17.340 concepts, to see patterns across domains, that's 00:02:17.340 --> 00:02:19.460 incredibly hard for machines. They're usually 00:02:19.460 --> 00:02:22.039 trained on specific data sets, highly specialized. 00:02:22.159 --> 00:02:24.639 They stay in their lane, so to speak. Right. 00:02:24.819 --> 00:02:27.620 They lack that intuitive feel for the wider world, 00:02:27.780 --> 00:02:30.599 that ability to make creative leaps between different 00:02:30.599 --> 00:02:33.120 areas of knowledge. OK, so the mission for us 00:02:33.120 --> 00:02:37.150 today, for you listening, is to figure out How 00:02:37.150 --> 00:02:40.610 do we actually train this human skill? How do 00:02:40.610 --> 00:02:43.009 we learn to think in ways machines just can't 00:02:43.009 --> 00:02:45.389 copy? Think of the sources we're using, these 00:02:45.389 --> 00:02:48.669 six books, as tools for building a kind of mental 00:02:48.669 --> 00:02:51.330 gymnasium. A cognitive gymnasium. Okay, yeah, 00:02:51.349 --> 00:02:54.250 each book, each thinking model, strengthens different 00:02:54.250 --> 00:02:56.389 cognitive muscles. They aren't just abstract 00:02:56.389 --> 00:02:58.669 theories, they're practical frameworks. They 00:02:58.669 --> 00:03:01.069 help you build a new cognitive operating system, 00:03:01.430 --> 00:03:03.830 one designed for connection -making and deep 00:03:03.900 --> 00:03:06.020 human thinking. Alright, I'm ready for the workout. 00:03:06.180 --> 00:03:07.719 Where do we start? Let's jump into the first 00:03:07.719 --> 00:03:10.360 one, systems thinking. Ugh, systems thinking. 00:03:10.800 --> 00:03:13.199 Okay, I think I know this feeling. You fix one 00:03:13.199 --> 00:03:14.919 problem right and then suddenly two new ones 00:03:14.919 --> 00:03:17.740 pop up somewhere else. Precisely, the classic 00:03:17.740 --> 00:03:20.340 whack -a -mole scenario. We're often trained 00:03:20.340 --> 00:03:23.199 to look at problems in neat little boxes. Like 00:03:23.199 --> 00:03:26.219 you boost sales figures, but then customer satisfaction 00:03:26.219 --> 00:03:29.680 takes a dive. Or you streamline a process. And 00:03:29.680 --> 00:03:32.800 suddenly team morale just plummets. It happens 00:03:32.800 --> 00:03:36.219 because we often treat symptoms without seeing 00:03:36.219 --> 00:03:38.439 the underlying connections. The ripple effects. 00:03:38.680 --> 00:03:41.340 Exactly. the feedback loops, everything's connected. 00:03:41.759 --> 00:03:44.199 Our solutions fail when they ignore these connections. 00:03:45.060 --> 00:03:47.719 Now, AI is great at optimizing within clear boundaries. 00:03:47.900 --> 00:03:50.379 But throw in unpredictable system interactions. 00:03:50.500 --> 00:03:52.979 And it struggles. Yeah. They can't easily anticipate 00:03:52.979 --> 00:03:55.879 those unintended consequences. Danela Meadows 00:03:55.879 --> 00:03:58.580 in Thinking in Systems peaches us to see those 00:03:58.580 --> 00:04:00.780 invisible threads. So it's not just solving problems, 00:04:00.860 --> 00:04:03.539 but understanding why they happen and why solutions 00:04:03.539 --> 00:04:06.039 sometimes backfire. Right. And the real game 00:04:06.039 --> 00:04:09.460 changer she introduces is the idea of Leverage 00:04:09.460 --> 00:04:11.319 points. Leverage points. Okay, what are those? 00:04:11.539 --> 00:04:13.759 They're places in a system where a small shift, 00:04:14.060 --> 00:04:16.720 a small change, can create a really big impact. 00:04:17.240 --> 00:04:19.680 She even ranks them. Ranks them from weak to 00:04:19.680 --> 00:04:22.000 strong. Yeah. The weakest is often just changing 00:04:22.000 --> 00:04:24.459 numbers, parameters, like budgets or targets. 00:04:24.740 --> 00:04:28.040 The strongest. Changing the paradigm, the fundamental 00:04:28.040 --> 00:04:31.339 beliefs or goals of the system itself. Wow. That 00:04:31.339 --> 00:04:33.420 sounds incredibly powerful, like finding the 00:04:33.420 --> 00:04:35.699 master switch. It really can be. Think about, 00:04:35.759 --> 00:04:38.550 say, city traffic. The usual response is what? 00:04:38.790 --> 00:04:41.370 Build more roads. Yeah, seems logical. But that's 00:04:41.370 --> 00:04:43.490 often a low leverage action. Meadows calls it 00:04:43.490 --> 00:04:45.930 changing a parameter. It often leads to induced 00:04:45.930 --> 00:04:49.009 demand. The new roads just fill up. Congestion 00:04:49.009 --> 00:04:51.269 returns. Okay, so what's the high leverage approach? 00:04:51.670 --> 00:04:54.589 A systems thinker questions the paradigm. Is 00:04:54.589 --> 00:04:57.529 driving a personal car the only or best way to 00:04:57.529 --> 00:04:59.569 get around? Then you look at higher leverage 00:04:59.569 --> 00:05:02.709 points. Investing massively in great public transport. 00:05:02.990 --> 00:05:05.089 Designing walkable cities. Promoting remote work. 00:05:05.339 --> 00:05:08.079 changing the goal or the mindset of the system. 00:05:08.220 --> 00:05:09.699 That's a completely different way of looking 00:05:09.699 --> 00:05:12.360 at it. Okay, so how can you, listening right 00:05:12.360 --> 00:05:16.420 now, apply this? Try a mini systems audit. Pick 00:05:16.420 --> 00:05:19.319 a recurring problem at work. A bottleneck, maybe, 00:05:19.399 --> 00:05:21.800 or a difficult team dynamic. Okay. First, just 00:05:21.800 --> 00:05:25.259 map it out. Who's involved, what processes, what 00:05:25.259 --> 00:05:28.899 tools or policies. Then, and this is key, draw 00:05:28.899 --> 00:05:30.939 the connections. How does one thing influence 00:05:30.939 --> 00:05:33.540 another? Like, maybe constant time pressure leads 00:05:33.540 --> 00:05:36.079 to rushed, lower quality work. Right. And that 00:05:36.079 --> 00:05:38.240 lower quality work maybe leads to more meetings 00:05:38.240 --> 00:05:40.740 to fix things. Which eats up time, leading to 00:05:40.740 --> 00:05:43.680 less deep work. Which increases the time pressure 00:05:43.680 --> 00:05:46.800 again. A vicious cycle. Exactly. A feedback loop. 00:05:46.970 --> 00:05:49.230 Once you see that loop, you can look for leverage 00:05:49.230 --> 00:05:51.490 points. Instead of just saying less meetings, 00:05:51.930 --> 00:05:53.990 maybe the leverage point is a rule change, like 00:05:53.990 --> 00:05:56.449 every meeting needs a clear agenda and objective 00:05:56.449 --> 00:05:58.949 set beforehand. Or changing a goal, like shifting 00:05:58.949 --> 00:06:02.389 focus from just task completion speed to, say, 00:06:02.870 --> 00:06:05.750 long -term value creation. Precisely. It's about 00:06:05.750 --> 00:06:07.889 finding where a small nudge can redirect the 00:06:07.889 --> 00:06:09.829 whole system. I actually did this with my morning 00:06:09.829 --> 00:06:12.310 routine once realized hitting snooze wasn't more 00:06:12.310 --> 00:06:14.850 rest, it was just more stress. Seeing the system 00:06:14.850 --> 00:06:18.379 helps. Okay, so systems thinking helps us map 00:06:18.379 --> 00:06:21.579 the territory, but the territory is often foggy, 00:06:21.779 --> 00:06:24.540 uncertain, right? How do we make good decisions 00:06:24.540 --> 00:06:27.660 then? Ah, that brings us perfectly to the next 00:06:27.660 --> 00:06:31.360 cognitive muscle, probabilistic thinking. This 00:06:31.360 --> 00:06:34.019 draws heavily from Annie Duke's Thinking in Bats. 00:06:34.259 --> 00:06:36.579 Annie Duke, the poker player. The very same. 00:06:37.019 --> 00:06:39.899 And poker teaches you something crucial. The 00:06:39.899 --> 00:06:42.629 world isn't black and white. It's shades of gray. 00:06:42.810 --> 00:06:45.649 It's probabilities. Our brains crave certainty, 00:06:46.170 --> 00:06:48.790 but reality rarely delivers. We want to know 00:06:48.790 --> 00:06:51.990 if a decision was right or wrong. Exactly. But 00:06:51.990 --> 00:06:53.829 Duke argues that's often the wrong question. 00:06:54.269 --> 00:06:56.769 She learned the hard way at the poker table that 00:06:56.769 --> 00:06:58.730 a good outcome doesn't automatically mean it 00:06:58.730 --> 00:07:00.509 was a good decision. And a bad outcome doesn't 00:07:00.509 --> 00:07:02.670 mean the decision was bad either. Precisely. 00:07:02.689 --> 00:07:04.470 You have to separate the quality of the decision 00:07:04.470 --> 00:07:07.470 process from the quality of the result. AI can 00:07:07.470 --> 00:07:10.050 crunch probabilities, sure. but it struggles 00:07:10.050 --> 00:07:12.750 with the ambiguity, the context, the hidden information 00:07:12.750 --> 00:07:16.569 of the real world. So is less, was I right, and 00:07:16.569 --> 00:07:18.990 more? Was my process sound, given what I knew 00:07:18.990 --> 00:07:21.949 at the time? Duke calls the trap of judging decisions 00:07:21.949 --> 00:07:24.529 solely by outcomes resulting. Resulting? Okay, 00:07:24.670 --> 00:07:26.889 give us an example. Imagine your team decides, 00:07:27.329 --> 00:07:29.810 based on solid data and research at that moment, 00:07:30.290 --> 00:07:33.300 to cut a product feature. Six months later, oops, 00:07:33.439 --> 00:07:35.500 a competitor launches something similar, and 00:07:35.500 --> 00:07:38.019 it's a huge hit. The immediate reaction is, oh 00:07:38.019 --> 00:07:41.139 no, we messed up. Terrible decision. That's resulting. 00:07:41.939 --> 00:07:45.300 The probabilistic thinker asks, OK, hold on. 00:07:45.579 --> 00:07:48.100 Given the information we had back then, was cutting 00:07:48.100 --> 00:07:52.060 the feature a reasonable bet? Maybe it was. Perhaps 00:07:52.060 --> 00:07:54.839 the market shifted unexpectedly or the competitor 00:07:54.839 --> 00:07:57.120 had different information. So the lesson isn't 00:07:57.120 --> 00:08:00.629 necessarily regret, but... Improvement. Exactly. 00:08:00.730 --> 00:08:02.750 How can we improve our information gathering? 00:08:03.189 --> 00:08:05.089 How can we refine our decision -making process 00:08:05.089 --> 00:08:07.170 for the next bet, regardless of how this one 00:08:07.170 --> 00:08:08.910 turned out? Okay, practical application time. 00:08:08.930 --> 00:08:11.269 How do we practice this? Conduct a decision debrief. 00:08:11.269 --> 00:08:14.089 After a big project, a key decision, win or lose, 00:08:14.269 --> 00:08:16.610 get the team together, or even just reflect yourself. 00:08:16.670 --> 00:08:18.689 What questions do you ask? What did we actually 00:08:18.689 --> 00:08:20.709 know when we made the call? What were the unknowns? 00:08:21.110 --> 00:08:23.110 What alternatives did we seriously consider? 00:08:23.329 --> 00:08:25.949 Was the discussion open? Were different viewpoints 00:08:25.949 --> 00:08:28.360 welcomed, even encouraged? And crucially, What 00:08:28.360 --> 00:08:31.279 can we learn to make better bets next time? It's 00:08:31.279 --> 00:08:33.720 not about assigning blame for the past. It's 00:08:33.720 --> 00:08:35.860 about improving judgment for the future. All 00:08:35.860 --> 00:08:37.379 right. We've mapped the system. We're learning 00:08:37.379 --> 00:08:40.139 to navigate uncertainty. Now let's talk about 00:08:40.139 --> 00:08:42.399 getting to the actual root of problems. This 00:08:42.399 --> 00:08:45.460 feels like logic 101, but maybe it's trickier 00:08:45.460 --> 00:08:47.840 than it looks. It often is. We're drawing here 00:08:47.840 --> 00:08:50.659 from Russell Ackoff and the art of problem solving. 00:08:51.639 --> 00:08:54.539 Logic seems basic, but it's surprising how often 00:08:54.539 --> 00:08:59.620 even very smart people make logical errors. Confusing 00:08:59.620 --> 00:09:02.639 correlation with causation is a classic. Jumping 00:09:02.639 --> 00:09:05.120 to conclusions. Assuming things without checking. 00:09:05.779 --> 00:09:08.460 Right. Ackhoff's big idea was that many, maybe 00:09:08.460 --> 00:09:11.059 most, workplace issues aren't really technical 00:09:11.059 --> 00:09:13.879 problems. They're problems of logic or definition 00:09:13.879 --> 00:09:16.679 in disguise. Logic or definition. Interesting. 00:09:16.919 --> 00:09:19.500 And his really provocative idea is that you shouldn't 00:09:19.500 --> 00:09:21.840 just aim to solve problems, you should aim to 00:09:21.840 --> 00:09:23.840 dissolve them. Dissolve them? Like, make them 00:09:23.840 --> 00:09:26.200 vanish? Essentially, yes. He argued that many 00:09:26.200 --> 00:09:28.340 problems only exist because of how we frame them 00:09:28.340 --> 00:09:30.220 in the first place. Change the frame, change 00:09:30.220 --> 00:09:32.039 the definition, and the problem itself might 00:09:32.039 --> 00:09:37.080 just disappear. Wow. Can AI do that? Step back 00:09:37.080 --> 00:09:39.379 and question the frame. That's precisely what 00:09:39.379 --> 00:09:42.159 it struggles with. AI is brilliant at solving 00:09:42.159 --> 00:09:44.860 clearly defined problems within given constraints. 00:09:45.240 --> 00:09:48.070 But asking, hold on. Are we even working on the 00:09:48.070 --> 00:09:51.009 right problem here? That's a deeply human step. 00:09:51.149 --> 00:09:53.409 OK, I need an example of dissolving a problem. 00:09:53.509 --> 00:09:55.490 Think about the common management question. How 00:09:55.490 --> 00:09:57.610 do we motivate our employees? Yeah, seems like 00:09:57.610 --> 00:10:00.190 a standard problem to solve. But notice the assumption 00:10:00.190 --> 00:10:03.429 baked in, that employees lack motivation and 00:10:03.429 --> 00:10:06.210 need it externally supplied. Ackhoff would flip 00:10:06.210 --> 00:10:08.990 this. He'd dissolve the problem by reframing 00:10:08.990 --> 00:10:12.610 it. Oh. He'd ask, What is our system, our policies, 00:10:12.769 --> 00:10:15.769 our processes, our culture doing that demotivates 00:10:15.769 --> 00:10:18.990 people who are likely already motivated? Ah, 00:10:19.169 --> 00:10:21.149 that completely shifts the focus. It's not about 00:10:21.149 --> 00:10:23.269 fixing the people. It's about fixing the environment 00:10:23.269 --> 00:10:24.990 around them. Exactly. Suddenly you're looking 00:10:24.990 --> 00:10:27.549 at bureaucracy, bad management, unclear goals, 00:10:27.850 --> 00:10:30.210 lack of autonomy, entirely different avenues 00:10:30.210 --> 00:10:32.289 for action. That's powerful. So for the listener, 00:10:32.370 --> 00:10:35.669 how can they practice this reframing? Try ACOF's 00:10:35.669 --> 00:10:38.429 five refrains exercise. Take a problem you're 00:10:38.429 --> 00:10:41.460 wrestling with. Let's say it's sales for Product 00:10:41.460 --> 00:10:43.879 X are declining. Okay, standard business problem. 00:10:44.419 --> 00:10:47.000 Now rewrite that problem statement in five different 00:10:47.000 --> 00:10:49.779 ways, each forcing a different perspective. Like 00:10:49.779 --> 00:10:51.879 how? You could frame it from the customer's view. 00:10:52.539 --> 00:10:55.460 How are the needs Product X used to meet changing? 00:10:56.059 --> 00:10:58.919 Or the competitor's view? What alternatives are 00:10:58.919 --> 00:11:02.340 customers choosing instead and why? Or the system 00:11:02.340 --> 00:11:05.279 view? What internal factors might be impacting 00:11:05.279 --> 00:11:08.840 Product X's performance? or value? Is product 00:11:08.840 --> 00:11:11.940 X's core value still relevant? Or even the inverse? 00:11:12.399 --> 00:11:14.440 How can we accelerate the decline of product 00:11:14.440 --> 00:11:17.399 X to make space for something better? Each one 00:11:17.399 --> 00:11:19.620 opens up totally different potential solutions, 00:11:19.840 --> 00:11:21.940 doesn't it? Completely. It breaks you out of 00:11:21.940 --> 00:11:23.840 tunnel vision and forces you to question the 00:11:23.840 --> 00:11:25.860 initial framing of the problem itself. Okay, 00:11:25.940 --> 00:11:28.379 this is fascinating. We're going deep. Now, let's 00:11:28.379 --> 00:11:30.240 broaden out. You mentioned interdisciplinary 00:11:30.240 --> 00:11:32.960 thinking earlier. This next one, broad thinking 00:11:32.960 --> 00:11:35.620 from David Epstein's range, things right up that 00:11:35.620 --> 00:11:37.960 alley. It absolutely is. And it pushes back against 00:11:37.960 --> 00:11:40.379 some really common advice, doesn't it? The whole 00:11:40.379 --> 00:11:44.580 niche down, specialize, find your lane mantra. 00:11:45.080 --> 00:11:48.419 Epstein argues that might actually be. Counterproductive. 00:11:48.820 --> 00:11:51.820 In many situations, yes. Especially in complex, 00:11:52.059 --> 00:11:54.460 unpredictable fields, which is, let's face it, 00:11:54.620 --> 00:11:57.519 most fields today. His research suggests that 00:11:57.519 --> 00:11:59.980 generalists, people with range, often outperform 00:11:59.980 --> 00:12:02.399 narrow specialists when faced with novel problems. 00:12:02.639 --> 00:12:04.639 But AI is the ultimate specialist, isn't it? 00:12:04.799 --> 00:12:07.080 It knows everything about its narrow domain. 00:12:07.340 --> 00:12:10.049 Precisely. And that's its limitation. It lacks 00:12:10.049 --> 00:12:12.730 that broad feel for the world, the intuition 00:12:12.730 --> 00:12:15.690 that comes from varied experiences. Breakthroughs 00:12:15.690 --> 00:12:18.070 often come from what Epstein calls cognitive 00:12:18.070 --> 00:12:21.090 bees. Cognitive bees. I like that. Explain. People 00:12:21.090 --> 00:12:22.970 who flip between different fields, different 00:12:22.970 --> 00:12:25.169 disciplines, picking up ideas here, pollinating 00:12:25.169 --> 00:12:28.200 concepts there, they create novel hybrids. Combinations 00:12:28.200 --> 00:12:30.120 that someone stuck in a single silo would never 00:12:30.120 --> 00:12:32.740 conceive of. So deep specialization might make 00:12:32.740 --> 00:12:35.299 you efficient at known tasks, but range makes 00:12:35.299 --> 00:12:37.639 you adaptable and innovative for unknown challenges. 00:12:37.840 --> 00:12:39.519 That's a great way to put it. So practically 00:12:39.519 --> 00:12:41.919 speaking, how do we cultivate this range? How 00:12:41.919 --> 00:12:45.620 do we become cognitive bees? Start actively collecting 00:12:45.620 --> 00:12:48.179 mental models, solutions, and ways of thinking 00:12:48.179 --> 00:12:50.779 from fields completely unrelated to your own. 00:12:51.200 --> 00:12:53.200 Be curious. Give me some examples. How could 00:12:53.200 --> 00:12:55.279 that work? OK, say you're struggling with company 00:12:55.279 --> 00:12:58.179 culture. Don't just read management books. How 00:12:58.179 --> 00:13:01.100 do marine biologists think about complex, evolving 00:13:01.100 --> 00:13:03.299 ecosystems? Maybe there are metaphors there. 00:13:03.480 --> 00:13:06.220 Interesting. Or maybe leading an innovation team 00:13:06.220 --> 00:13:08.659 under pressure. Look at how Michelin -starred 00:13:08.659 --> 00:13:11.539 chefs manage intense creativity and execution 00:13:11.539 --> 00:13:14.860 in a chaotic kitchen environment. Or designing 00:13:14.860 --> 00:13:17.720 better customer experiences. How do video game 00:13:17.720 --> 00:13:20.419 designers keep players hooked and engaged for 00:13:20.419 --> 00:13:23.059 hundreds of hours? So you deliberately look for 00:13:23.059 --> 00:13:25.600 analogies and frameworks in unexpected places. 00:13:25.820 --> 00:13:29.279 Yes. Maybe commit, say, 20 % of your reading 00:13:29.279 --> 00:13:32.320 or learning time to exploring areas totally outside 00:13:32.320 --> 00:13:34.960 your professional domain. It builds a unique 00:13:34.960 --> 00:13:37.580 mental toolkit. The connections might not be 00:13:37.580 --> 00:13:39.879 immediate, but when you face a truly new challenge, 00:13:40.240 --> 00:13:42.159 you'll have this diverse set of tools that no 00:13:42.159 --> 00:13:44.759 specialized AI can match. All right, range gives 00:13:44.759 --> 00:13:47.580 us diverse inputs. Now, how do we generate actual 00:13:47.580 --> 00:13:49.460 ideas from that? Let's talk design thinking, 00:13:49.679 --> 00:13:51.740 drawing from idea flow. All right, because it's 00:13:51.740 --> 00:13:54.639 easy to fall into traps here, too. One common 00:13:54.639 --> 00:13:56.899 trap is solving problems backward. Backward, 00:13:56.940 --> 00:13:58.820 what do you mean? Starting with a solution you 00:13:58.820 --> 00:14:00.980 already know, or like maybe a piece of technology 00:14:00.980 --> 00:14:04.080 or a familiar process, and then looking for a 00:14:04.080 --> 00:14:06.639 problem it can solve, or... Or just jumping on 00:14:06.639 --> 00:14:08.679 the very first solution that comes to mind for 00:14:08.679 --> 00:14:11.620 a problem. Exactly. We converge too quickly. 00:14:11.950 --> 00:14:15.029 Idea Flow emphasizes a more systematic process, 00:14:15.269 --> 00:14:17.490 drawing from design thinking principles. It's 00:14:17.490 --> 00:14:19.450 about deliberately expanding the possibilities. 00:14:20.009 --> 00:14:22.809 First, the divergence phase. Brainstorming widely, 00:14:23.009 --> 00:14:25.070 no judgment. Yes, generating lots of options, 00:14:25.210 --> 00:14:28.429 even wild ones. Then, you systematically narrow 00:14:28.429 --> 00:14:31.549 down and refine the convergence phase. AI can 00:14:31.549 --> 00:14:34.309 generate variations on a theme, but it struggles 00:14:34.309 --> 00:14:36.850 with that initial, wide -open exploration and 00:14:36.850 --> 00:14:39.250 questioning the fundamental need. It won't easily 00:14:39.250 --> 00:14:42.080 ask, do we even need an idea here? So how do 00:14:42.080 --> 00:14:44.779 we ensure we diverge properly? One powerful tool 00:14:44.779 --> 00:14:48.159 they highlight is the How Might We or HMW framework. 00:14:49.059 --> 00:14:51.440 It's a way to phrase questions that inherently 00:14:51.440 --> 00:14:54.200 opens up possibilities. Okay, let's use an example. 00:14:54.820 --> 00:14:57.779 Say a local bookstore struggling against online 00:14:57.779 --> 00:15:01.379 giants. The obvious maybe backward solution is 00:15:01.529 --> 00:15:04.230 Offer discounts. Right. A race to the bottom, 00:15:04.250 --> 00:15:07.669 they probably can't win. But using HMW questions 00:15:07.669 --> 00:15:11.049 reframes the challenge. How might we? What? How 00:15:11.049 --> 00:15:13.970 might we turn the bookstore into the community's 00:15:13.970 --> 00:15:16.850 third living room? How might we create a book 00:15:16.850 --> 00:15:19.409 discovery experience that an algorithm simply 00:15:19.409 --> 00:15:22.549 can't replicate? How might we become a hub for 00:15:22.549 --> 00:15:25.289 local cultural events and book clubs? Ah, see, 00:15:25.289 --> 00:15:27.409 each of those questions points towards completely 00:15:27.409 --> 00:15:30.090 different types of solutions. Community, experience, 00:15:30.389 --> 00:15:33.659 events. not just price. Exactly. They open up 00:15:33.659 --> 00:15:35.980 the solution space dramatically, focusing on 00:15:35.980 --> 00:15:38.039 human connection and experience, areas where 00:15:38.039 --> 00:15:40.419 the physical store has an advantage. So the practical 00:15:40.419 --> 00:15:42.320 application for you listening? Take a problem 00:15:42.320 --> 00:15:44.639 you're stuck on. Instead of jumping to solutions, 00:15:44.799 --> 00:15:46.980 try generating five or ten different how -might 00:15:46.980 --> 00:15:49.220 -we questions about it. See how it shifts your 00:15:49.220 --> 00:15:51.539 perspective and what new avenues it opens up. 00:15:51.679 --> 00:15:53.820 It forces you to think wider before you narrow 00:15:53.820 --> 00:15:56.500 down. Okay, we've covered a lot of ground. Mapping 00:15:56.500 --> 00:15:59.820 systems, betting under uncertainty, reframing 00:15:59.820 --> 00:16:02.820 problems, thinking broadly, generating ideas. 00:16:03.259 --> 00:16:05.899 What's our final workout? We end with maybe the 00:16:05.899 --> 00:16:08.360 most fundamental and perhaps challenging mode. 00:16:09.120 --> 00:16:12.440 First, principles thinking. This is heavily inspired 00:16:12.440 --> 00:16:15.860 by Peter Thiel's zero to one. Zero to one, meaning 00:16:15.860 --> 00:16:18.740 creating something entirely new, not just improving 00:16:18.740 --> 00:16:22.740 what exists. Exactly. Most people, most companies 00:16:22.740 --> 00:16:25.320 operate in the one to end space. They take something 00:16:25.320 --> 00:16:27.080 that already exists and make it incrementally 00:16:27.080 --> 00:16:29.580 better. That's reasoning by analogy. Doing what 00:16:29.580 --> 00:16:31.539 others are doing, but maybe slightly faster or 00:16:31.539 --> 00:16:33.980 cheaper? Right. But truly disruptive innovation, 00:16:34.279 --> 00:16:36.840 the zero to one leaps, often come from reasoning 00:16:36.840 --> 00:16:39.120 from first principles. And AI struggles with 00:16:39.120 --> 00:16:42.899 this. Massively. AI is, in its essence, an analogy 00:16:42.899 --> 00:16:45.220 machine. It learns from vast amounts of existing 00:16:45.220 --> 00:16:47.740 data and examples. It's brilliant at finding 00:16:47.740 --> 00:16:50.639 patterns in what is. But it can't easily tear 00:16:50.639 --> 00:16:52.940 down existing assumptions and build something 00:16:52.940 --> 00:16:55.200 completely new from the ground up because its 00:16:55.200 --> 00:16:58.139 entire foundation is based on precedent. So what 00:16:58.139 --> 00:17:00.039 does thinking from first principles actually 00:17:00.039 --> 00:17:02.899 involve? It means breaking a problem down to 00:17:02.899 --> 00:17:05.839 its absolute most fundamental truths. The things 00:17:05.839 --> 00:17:08.420 you know are true, like the laws of physics or 00:17:08.420 --> 00:17:11.259 basic human needs. You strip away all the assumptions, 00:17:11.460 --> 00:17:13.259 all the conventions, all the way things are usually 00:17:13.259 --> 00:17:16.119 done. And then you build back up from only those 00:17:16.119 --> 00:17:19.039 fundamental truths. Precisely. Ignoring how it's 00:17:19.039 --> 00:17:21.480 done now. Let's take an example. Improving employee 00:17:21.480 --> 00:17:24.690 training. Okay. The typical approach is, how 00:17:24.690 --> 00:17:27.549 can we make our workshops better? Or should we 00:17:27.549 --> 00:17:30.650 buy a new online course platform? That's reasoning 00:17:30.650 --> 00:17:33.910 by analogy, improving existing models. First 00:17:33.910 --> 00:17:35.990 principles thinking starts differently. Ask, 00:17:36.589 --> 00:17:39.690 what is the fundamental truth here? Employees 00:17:39.690 --> 00:17:42.309 need certain skills to do their jobs well, and 00:17:42.309 --> 00:17:45.150 maybe. Learning is most effective when it's relevant 00:17:45.150 --> 00:17:47.609 and applied quickly. Good. Now, what's the current 00:17:47.609 --> 00:17:49.650 assumption? that workshops or online courses 00:17:49.650 --> 00:17:52.450 are the best or only way to deliver that skill 00:17:52.450 --> 00:17:55.029 transfer. Okay, now ignore that assumption. Based 00:17:55.029 --> 00:17:57.329 only on the fundamental truths needing skills, 00:17:57.609 --> 00:18:00.390 learning by doing, how would you design the absolute 00:18:00.390 --> 00:18:03.690 best way to transfer skills from scratch? Maybe 00:18:03.690 --> 00:18:06.109 it's not a course at all. Maybe it's a structured 00:18:06.109 --> 00:18:10.109 mentorship program. Or guided real -world projects 00:18:10.109 --> 00:18:13.509 with immediate feedback. Or an internal platform 00:18:13.509 --> 00:18:16.569 where experts share knowledge directly as needed. 00:18:17.439 --> 00:18:21.460 See? You've just opened the door to potentially 00:18:21.460 --> 00:18:24.660 far superior solutions simply by questioning 00:18:24.660 --> 00:18:27.220 the ingrained assumption and building up from 00:18:27.220 --> 00:18:29.859 bedrock truths. That's a powerful mental reset. 00:18:30.160 --> 00:18:31.940 So the application for listeners. Take a big 00:18:31.940 --> 00:18:34.319 challenge you're facing. Really grill yourself. 00:18:34.660 --> 00:18:37.299 What are the absolute undeniable truths here? 00:18:37.359 --> 00:18:39.000 What are the assumptions I'm making just because 00:18:39.000 --> 00:18:41.539 that's how it's always been done? And crucially, 00:18:41.720 --> 00:18:44.200 How would I rebuild this from scratch based only 00:18:44.200 --> 00:18:46.660 on those truths? Wow. Okay. That's six powerful 00:18:46.660 --> 00:18:50.420 ways of thinking systems That's logic range design 00:18:50.420 --> 00:18:52.859 first principles. So what happens when you put 00:18:52.859 --> 00:18:54.680 them all together? What's the big picture? The 00:18:54.680 --> 00:18:56.480 big picture is that these aren't just isolated 00:18:56.480 --> 00:18:58.380 tools. They work together They reinforce each 00:18:58.380 --> 00:19:00.720 other think of it as installing a new more powerful 00:19:00.720 --> 00:19:03.720 cognitive operating system a human OS upgrade 00:19:03.720 --> 00:19:06.759 for the AI something like that Their real power 00:19:06.759 --> 00:19:08.460 comes from their synergy. How do they connect? 00:19:08.819 --> 00:19:11.269 We'll think about it Systems thinking gives you 00:19:11.269 --> 00:19:13.950 the map. It shows you the complex landscape of 00:19:13.950 --> 00:19:16.049 the problem, all the interconnections. Okay, 00:19:16.049 --> 00:19:18.890 I see the map. Probabilistic thinking is your 00:19:18.890 --> 00:19:21.470 compass for navigating the inherent uncertainty 00:19:21.470 --> 00:19:24.529 on that map. It helps you make smarter bets when 00:19:24.529 --> 00:19:27.690 the path isn't clear. Got my map and compass. 00:19:28.029 --> 00:19:30.569 Ackoff's logic ensures you're not just wandering, 00:19:30.930 --> 00:19:33.450 but actually trying to get to the right destination 00:19:33.450 --> 00:19:36.130 by dissolving misleading problems and asking 00:19:36.130 --> 00:19:37.890 the right questions. Making sure I'm heading 00:19:37.890 --> 00:19:40.650 the right way. Range gives you potential shortcuts 00:19:40.650 --> 00:19:43.170 and alternative routes by letting you borrow 00:19:43.170 --> 00:19:46.410 ideas and solutions from distant, unexpected 00:19:46.410 --> 00:19:49.950 places on the map. Finding clever pathways. Design 00:19:49.950 --> 00:19:52.470 thinking is like your sketchbook. It helps you 00:19:52.470 --> 00:19:55.150 rapidly generate and explore dozens of potential 00:19:55.150 --> 00:19:57.509 routes before committing to one. Exploring the 00:19:57.509 --> 00:19:59.849 options. And first principles thinking. That's 00:19:59.849 --> 00:20:02.150 the ultimate power tool. It gives you the ability 00:20:02.150 --> 00:20:05.640 to say, this whole map is wrong. tear it up, 00:20:06.000 --> 00:20:08.019 and draw a fundamentally new and better one from 00:20:08.019 --> 00:20:11.839 scratch. Map, compass, destination check, shortcuts, 00:20:12.240 --> 00:20:14.480 route sketching, and the power to redraw the 00:20:14.480 --> 00:20:17.599 map. That's quite a toolkit. When you start combining 00:20:17.599 --> 00:20:20.259 these modes of thought, you develop a way of 00:20:20.259 --> 00:20:23.200 seeing and solving problems that is deeply human 00:20:23.200 --> 00:20:25.660 and incredibly difficult for any current AI to 00:20:25.660 --> 00:20:28.380 replicate. That's the edge. That's how you become 00:20:28.380 --> 00:20:31.819 indispensable. Hashtag, hashtag, tag, atro. So 00:20:31.819 --> 00:20:33.980 the message here seems pretty clear. I think 00:20:33.980 --> 00:20:36.140 so. The future really does belong to minds that 00:20:36.140 --> 00:20:39.180 can think in ways machines can't. If your value 00:20:39.180 --> 00:20:42.420 is primarily in speed, efficiency, or just executing 00:20:42.420 --> 00:20:45.500 known procedures, well, AI is getting very good 00:20:45.500 --> 00:20:47.799 at that. Those roles are likely at risk. But 00:20:47.799 --> 00:20:51.059 AI cannot easily replicate the ability to connect 00:20:51.059 --> 00:20:54.200 disparate ideas, to navigate profound uncertainty 00:20:54.200 --> 00:20:57.220 with good judgment, to reframe and dissolve complex 00:20:57.220 --> 00:20:59.900 problems, or to imagine something truly new. 00:21:00.039 --> 00:21:02.339 So our job isn't to become more like machines. 00:21:02.539 --> 00:21:04.559 No. It's to become more human in our thinking, 00:21:04.740 --> 00:21:07.019 to lean into these cognitive strengths that are 00:21:07.019 --> 00:21:09.660 uniquely ours. And the great thing is we've outlined 00:21:09.660 --> 00:21:12.180 a clear path to start doing that based on these 00:21:12.180 --> 00:21:14.470 sources. Where should someone begin? Don't try 00:21:14.470 --> 00:21:17.430 to master all six at once. That's overwhelming. 00:21:17.789 --> 00:21:20.650 Pick the one thinking model, the one book that 00:21:20.650 --> 00:21:23.529 resonates most with you right now, or that addresses 00:21:23.529 --> 00:21:25.930 what you feel is your biggest current weakness. 00:21:26.410 --> 00:21:28.569 Just to recap the titles for everyone, we talked 00:21:28.569 --> 00:21:31.089 about Thinking in Systems by Daniella Meadows. 00:21:31.289 --> 00:21:34.539 Thinking in Bets by Annie Duke. The Art of Problem 00:21:34.539 --> 00:21:37.599 Solving by Russell Ackoff. Range by David Epstein. 00:21:38.039 --> 00:21:40.880 Idea Flow by Jeremy Utley and Perry Kleban. And 00:21:40.880 --> 00:21:43.000 Zero to One by Peter Thiel and Blake Masters. 00:21:43.200 --> 00:21:46.059 Pick one, dive in, and start exercising that 00:21:46.059 --> 00:21:49.460 particular cognitive muscle. Exactly. This deep 00:21:49.460 --> 00:21:51.640 dive wasn't just about understanding the challenge. 00:21:51.960 --> 00:21:54.740 It was about giving you practical tools, a way 00:21:54.740 --> 00:21:57.460 forward, a way to not just survive, but actually 00:21:57.460 --> 00:22:00.380 thrive by leveraging your unique human intelligence 00:22:00.380 --> 00:22:03.390 in the age of AI. Don't just listen and start 00:22:03.390 --> 00:22:05.630 thinking differently. Start practicing. Your 00:22:05.630 --> 00:22:07.029 future thinking starts now.