WEBVTT 00:00:00.000 --> 00:00:03.960 So I want you to picture a scenario. Just think 00:00:03.960 --> 00:00:06.419 about a typical day at work. OK. Or maybe you're 00:00:06.419 --> 00:00:08.480 just at home troubleshooting a weird noise in 00:00:08.480 --> 00:00:10.839 your car. Yeah. You're staring at the engine, 00:00:11.140 --> 00:00:12.660 trying to figure out the problem, and suddenly 00:00:12.660 --> 00:00:16.019 you find an explanation. Right. It fits perfectly. 00:00:16.160 --> 00:00:18.879 It makes complete sense. It explains the noise. 00:00:19.179 --> 00:00:21.179 Or if you're at work, it explains the drop in 00:00:21.179 --> 00:00:24.019 sales. Whatever it is, what is the very next 00:00:24.019 --> 00:00:28.480 thing you do? Well, usually. You stop looking. 00:00:28.559 --> 00:00:30.239 You stop looking. You totally stop. You pat yourself 00:00:30.239 --> 00:00:32.299 on the back, you close the ticket, and you just 00:00:32.299 --> 00:00:34.259 move on. Yeah, it's that moment where you stop 00:00:34.259 --> 00:00:36.100 being an investigator and you kind of start being 00:00:36.100 --> 00:00:39.200 a defense attorney for your own idea. Exactly. 00:00:39.579 --> 00:00:42.119 And that is exactly what we are unpacking in 00:00:42.119 --> 00:00:44.520 today's Deep Dive. We're looking at a source 00:00:44.520 --> 00:00:47.820 that argues that specific moment, that feeling 00:00:47.820 --> 00:00:51.719 of, aha, I found it, is actually a massive cognitive 00:00:51.719 --> 00:00:53.820 trap. It really is. And the source we have for 00:00:53.820 --> 00:00:57.259 this is fascinating. It's a bit of a... A source 00:00:57.259 --> 00:00:59.640 within a source, actually. Yeah, it's layered. 00:01:00.219 --> 00:01:02.299 We're looking at a blog post by Gary Ackerman, 00:01:02.719 --> 00:01:05.840 published in December 2025 on hackscience .education. 00:01:05.980 --> 00:01:09.219 It's titled On Multiple Working Hypotheses. But 00:01:09.219 --> 00:01:11.700 the backstory here is what's really wild. Ackerman 00:01:11.700 --> 00:01:13.840 shares this personal story about a paper he read 00:01:13.840 --> 00:01:17.140 as an undergrad back in 1985. Right, and it completely 00:01:17.140 --> 00:01:19.500 changed his life. It did, but then he lost his 00:01:19.500 --> 00:01:22.510 copy of it. He spent 40 years looking for this 00:01:22.510 --> 00:01:25.750 specific paper. 40 years? Yeah, and he finally 00:01:25.750 --> 00:01:28.430 tracked it down recently using AI to search for 00:01:28.430 --> 00:01:31.250 the concepts. And the kicker, the paper wasn't 00:01:31.250 --> 00:01:34.790 from 1985 at all. No, it was from 1890. 1890, 00:01:35.069 --> 00:01:37.390 written by a geologist named T .C. Chamberlain. 00:01:37.530 --> 00:01:39.689 I just love the archaeology of this idea. We 00:01:39.689 --> 00:01:43.049 were dusting off this 135 year old document to 00:01:43.049 --> 00:01:45.129 answer a really modern question. Which is basically 00:01:45.129 --> 00:01:47.390 why the standard scientific method and, you know, 00:01:47.409 --> 00:01:49.430 the one we all learned in fifth grade might actually 00:01:49.430 --> 00:01:51.959 be terrible for our critical thinking. It's a 00:01:51.959 --> 00:01:54.239 really bold claim. I mean we're usually taught 00:01:54.239 --> 00:01:56.909 that forming a hypothesis and then testing it 00:01:56.909 --> 00:01:59.989 is the absolute peak of intellectual rigor. Right, 00:02:00.030 --> 00:02:02.549 but Chamberlain argues it's actually a step down 00:02:02.549 --> 00:02:04.870 from the real gold standard and Ackerman does 00:02:04.870 --> 00:02:07.170 a great job breaking down Chamberlain's core 00:02:07.170 --> 00:02:09.610 framework. Yeah, let's get into that framework 00:02:09.610 --> 00:02:13.449 because Chamberlain identified three distinct 00:02:13.449 --> 00:02:16.270 types of learning or three ways we try to discover 00:02:16.270 --> 00:02:18.330 new truths. Right. And the first one he points 00:02:18.330 --> 00:02:21.129 out is what he calls the ruling theory. Yeah, 00:02:21.150 --> 00:02:23.270 the ruling theory. This is essentially the bottom 00:02:23.270 --> 00:02:26.789 of the ladder. It's dogma. Just strict imitation. 00:02:27.110 --> 00:02:30.610 Exactly. You accept an explanation simply because 00:02:30.610 --> 00:02:33.370 an authority figure, a teacher, a boss, some 00:02:33.370 --> 00:02:35.930 ruling voice told you it's the truth. You aren't 00:02:35.930 --> 00:02:37.930 investigating anything. You're just complying. 00:02:38.129 --> 00:02:40.629 It sounds a lot like the because I said so style 00:02:40.629 --> 00:02:42.810 of management. It totally is. And the danger 00:02:42.810 --> 00:02:45.330 is that you start interpreting any new facts 00:02:45.330 --> 00:02:48.090 specifically to make them fit into that old theory 00:02:48.090 --> 00:02:50.129 you were handed. Which is efficient, I guess. 00:02:50.120 --> 00:02:52.039 but it's completely stagnant. You never learn 00:02:52.039 --> 00:02:54.939 anything new. Right. So eventually, thinkers 00:02:54.939 --> 00:02:57.419 realized this wasn't great, and they moved to 00:02:57.419 --> 00:03:00.199 the second tier, which Chamberlain calls the 00:03:00.199 --> 00:03:02.379 working hypothesis. And this is the one most 00:03:02.379 --> 00:03:05.000 of us are familiar with. You see a problem, you 00:03:05.000 --> 00:03:07.439 come up with a likely explanation, your one working 00:03:07.439 --> 00:03:09.699 hypothesis, and then you test it to see if it's 00:03:09.699 --> 00:03:12.159 true. That's the one. But this is where I got 00:03:12.159 --> 00:03:13.960 a bit stuck when I was first reading the source 00:03:13.960 --> 00:03:16.659 material, because that sounds, well, correct. 00:03:16.889 --> 00:03:19.310 That's how science works, right? You have an 00:03:19.310 --> 00:03:21.990 idea and you test it. It seems logical. So why 00:03:21.990 --> 00:03:23.849 does Chamberlain and Ackerman, by extension, 00:03:24.569 --> 00:03:27.590 think this method is actually dangerous? It's 00:03:27.590 --> 00:03:30.509 all about human psychology, really. Chamberlain 00:03:30.509 --> 00:03:33.069 had this incredible psychological insight way 00:03:33.069 --> 00:03:36.030 ahead of his time. He argues that the flaw isn't 00:03:36.030 --> 00:03:38.550 in the testing, it's in when you stop thinking. 00:03:38.849 --> 00:03:41.409 What do you mean? Well, as soon as you formulate 00:03:41.409 --> 00:03:43.810 that single working hypothesis to answer the 00:03:43.810 --> 00:03:46.789 question, how did this happen, your mindset shifts. 00:03:47.169 --> 00:03:49.330 You stop being an impartial observer. Because 00:03:49.330 --> 00:03:51.960 now I have like... skin in the game. Exactly. 00:03:52.199 --> 00:03:54.780 You're invested. Chamberlain uses this brilliant, 00:03:55.060 --> 00:03:57.159 honestly kind of funny phrase for it. He says 00:03:57.159 --> 00:03:59.740 the moment you create a single theory, it becomes 00:03:59.740 --> 00:04:03.379 your intellectual offspring. Intellectual offspring. 00:04:03.979 --> 00:04:06.560 Oh, man. So we basically turned into helicopter 00:04:06.560 --> 00:04:09.759 parents for our own ideas. We do. A sort of parental 00:04:09.759 --> 00:04:12.520 affection kicks in. And because of that affection, 00:04:13.219 --> 00:04:15.819 even if you are actively trying to remain objective, 00:04:16.240 --> 00:04:19.220 you unconsciously start steering the investigation. 00:04:19.360 --> 00:04:21.680 Make your kid look good. Right. You start cherry 00:04:21.680 --> 00:04:24.899 picking data that supports your idea. Yeah. And 00:04:24.899 --> 00:04:28.060 maybe even worse, you naturally start to ignore 00:04:28.060 --> 00:04:30.660 or downplay evidence that supports any other 00:04:30.660 --> 00:04:32.860 theory. I can completely see that happening. 00:04:32.899 --> 00:04:35.579 It's so relatable. If I go to my boss and say, 00:04:35.939 --> 00:04:38.279 hey, I think the problem with our sales is the 00:04:38.279 --> 00:04:41.420 new marketing budget. Now I am highly motivated 00:04:41.420 --> 00:04:43.439 to find evidence that proves it's the marketing 00:04:43.439 --> 00:04:47.220 budget. Yes. You press the facts to fit your 00:04:47.220 --> 00:04:50.040 theory, just like a parent making excuses for 00:04:50.040 --> 00:04:51.980 their child's bad behavior at the playground. 00:04:52.160 --> 00:04:53.860 Like, oh, he didn't mean to push that kid. He's 00:04:53.860 --> 00:04:56.040 just expressing his boundary. Exactly. The data 00:04:56.040 --> 00:04:59.300 isn't an outlier. It's just a special case. That 00:04:59.300 --> 00:05:01.980 is so true. So the working hypothesis doesn't 00:05:01.980 --> 00:05:04.879 just guide your research. It actively biases 00:05:04.879 --> 00:05:07.779 you as the researcher. You're no longer testing 00:05:07.779 --> 00:05:10.399 to find the truth. You're testing to prove yourself 00:05:10.399 --> 00:05:13.509 right. It's a subtle distinction, but it's massive 00:05:13.509 --> 00:05:16.069 when it comes to critical thinking. OK, so if 00:05:16.069 --> 00:05:19.189 having one idea is a trap, what's the alternative? 00:05:19.589 --> 00:05:21.589 Do we just go in with zero ideas? No, you go 00:05:21.589 --> 00:05:23.649 in with the third tier, the method of multiple 00:05:23.649 --> 00:05:26.410 working hypotheses. Multiple working hypotheses. 00:05:26.910 --> 00:05:29.370 OK, so that just sounds like having backup plans. 00:05:29.730 --> 00:05:32.589 It's a bit more profound than that. It's a fundamental 00:05:32.589 --> 00:05:35.610 shift in your starting line. Instead of coming 00:05:35.610 --> 00:05:38.350 up with one explanation and testing it, you force 00:05:38.350 --> 00:05:41.949 yourself to explain and test every rational explanation 00:05:41.949 --> 00:05:45.509 you can conjure up all simultaneously. Wow. You 00:05:45.509 --> 00:05:48.230 don't just have one child. You bring a whole 00:05:48.230 --> 00:05:50.550 family of intellectual offspring into the world 00:05:50.550 --> 00:05:53.089 at the exact same time. I see. So if I have five 00:05:53.089 --> 00:05:55.370 hypotheses, I can't really play favorites. Right. 00:05:55.509 --> 00:05:57.670 If you have five kids, you have to be fair. It 00:05:57.670 --> 00:05:59.550 neutralizes your bias because your affection 00:05:59.550 --> 00:06:02.009 is divided. Let me try an analogy here to see 00:06:02.009 --> 00:06:03.790 if I'm tracking the difference between the single 00:06:03.790 --> 00:06:08.149 working hypothesis and the multiple method. It 00:06:08.149 --> 00:06:09.910 feels a bit like the difference between betting 00:06:09.910 --> 00:06:12.509 on a horse and just watching the race. Oh, I 00:06:12.509 --> 00:06:14.569 like that. Go on. If I'm using the single working 00:06:14.569 --> 00:06:16.449 hypothesis, I'm basically putting all my money 00:06:16.449 --> 00:06:19.509 on one horse. Let's call him Glacier. OK, Glacier. 00:06:19.750 --> 00:06:21.269 I'm watching the race, but I'm only watching 00:06:21.269 --> 00:06:25.129 Glacier. If he stumbles, I make excuses for the 00:06:25.129 --> 00:06:27.350 track conditions. If he speeds up, I'm cheering. 00:06:27.990 --> 00:06:30.199 But I don't really care what the other horses 00:06:30.199 --> 00:06:32.959 are doing unless they get in his way. My observation 00:06:32.959 --> 00:06:35.519 is entirely narrowed to that one lane. That's 00:06:35.519 --> 00:06:37.079 a perfect way to put it. You're blind to the 00:06:37.079 --> 00:06:39.819 rest of the field. Right. But with multiple working 00:06:39.819 --> 00:06:42.600 hypotheses, I haven't placed a bet or I guess 00:06:42.600 --> 00:06:44.839 I've placed an equal bet on every single horse 00:06:44.839 --> 00:06:46.759 in the race. So I'm just watching the whole field 00:06:46.759 --> 00:06:48.879 to see who is actually the fastest. And that 00:06:48.879 --> 00:06:52.490 allows you to see the race as it truly is. rather 00:06:52.490 --> 00:06:54.689 than how you want it to be. You're a judge, not 00:06:54.689 --> 00:06:57.730 a gambler. It's a powerful reframing. And it's 00:06:57.730 --> 00:06:59.430 not even about being smarter than everyone else. 00:06:59.449 --> 00:07:02.629 It's just a hack to trick your own ego. Exactly. 00:07:02.709 --> 00:07:05.509 And we have to remember the context here. Chamberlain 00:07:05.509 --> 00:07:07.529 wasn't just sitting in a room thinking about 00:07:07.529 --> 00:07:10.310 abstract philosophy or horse racing. He was a 00:07:10.310 --> 00:07:13.389 working geologist. Right, in 1890. Yes. So he 00:07:13.389 --> 00:07:16.350 was dealing with massive, incredibly complex 00:07:16.350 --> 00:07:19.430 natural systems. Specifically, he uses the example 00:07:19.430 --> 00:07:22.089 of the formation of the Great Lakes. The source 00:07:22.089 --> 00:07:25.509 material mentions this. I imagine in the late 00:07:25.509 --> 00:07:28.649 1800s, there were some pretty heated debates 00:07:28.649 --> 00:07:30.930 about how those legs got there. Oh, incredibly 00:07:30.930 --> 00:07:33.769 heated. One camp of geologists would say, glaciers 00:07:33.769 --> 00:07:35.670 carve them out. Another camp would say, no, it 00:07:35.670 --> 00:07:38.089 was ancient river erosion. And if you're using 00:07:38.089 --> 00:07:40.430 the single working hypothesis, you pick your 00:07:40.430 --> 00:07:43.189 side, you pick glaciers, and you go out looking 00:07:43.189 --> 00:07:45.290 for glacial scratches on the rocks. And guess 00:07:45.290 --> 00:07:48.250 what? You will absolutely find them. Because 00:07:48.250 --> 00:07:50.310 they're there. And then you declare victory. 00:07:50.410 --> 00:07:53.189 See? I told you it was the glaciers. But you 00:07:53.189 --> 00:07:55.470 completely missed the fact that the river erosion 00:07:55.470 --> 00:07:58.230 is also there. Chamberlain points out that in 00:07:58.230 --> 00:08:00.589 the real world, especially with complex phenomena, 00:08:01.050 --> 00:08:03.610 there is almost never just one simple cause. 00:08:03.629 --> 00:08:06.209 Right. The Great Lakes were formed by glaciers 00:08:06.209 --> 00:08:08.889 and pre -existing erosion and tectonic shifts, 00:08:09.189 --> 00:08:11.829 all of them interacting together over time. So 00:08:11.829 --> 00:08:15.029 having a single hypothesis forces this weird 00:08:15.029 --> 00:08:17.990 true false binary onto a world that is actually 00:08:17.990 --> 00:08:20.889 like all of the above. Exactly. That is the danger 00:08:20.889 --> 00:08:22.569 he's worrying us about. If you are looking for 00:08:22.569 --> 00:08:24.629 the cause, singular, you will definitely find 00:08:24.629 --> 00:08:27.290 a cause. But you'll miss the whole system. Which 00:08:27.290 --> 00:08:30.230 means your solution is going to be flawed. Right. 00:08:30.430 --> 00:08:33.409 The method of multiple working hypotheses forces 00:08:33.409 --> 00:08:36.570 you to look at the interaction of causes. Chamberlain 00:08:36.570 --> 00:08:39.889 said this leads to better and truer observations 00:08:39.889 --> 00:08:42.950 because you simply aren't forcing the data to 00:08:42.950 --> 00:08:45.169 fit a neat little narrative. That makes so much 00:08:45.169 --> 00:08:47.809 sense. And this transitions perfectly into the 00:08:47.809 --> 00:08:50.870 second half of our deep dive today. Because Gary 00:08:50.870 --> 00:08:53.490 Ackerman, the guy who wrote the 2025 blog post, 00:08:53.929 --> 00:08:56.570 he's not a geologist. No, his field is education. 00:08:56.850 --> 00:08:59.570 Right. And he argues that a modern classroom 00:08:59.570 --> 00:09:03.070 is just as complex, if not more complex, than 00:09:03.070 --> 00:09:05.070 the geology of the Great Lakes. Oh, without a 00:09:05.070 --> 00:09:07.659 doubt. Think about it. 30 distinct personalities 00:09:07.659 --> 00:09:10.000 in a room, different backgrounds, different learning 00:09:10.000 --> 00:09:12.320 styles, different moods on any given day, all 00:09:12.320 --> 00:09:14.419 interacting in real time. It's an ecosystem. 00:09:14.820 --> 00:09:17.559 It is. And Ackerman points out that the intellectual 00:09:17.559 --> 00:09:20.679 offspring trap hits the education sector incredibly 00:09:20.679 --> 00:09:22.639 hard. How does that play out in the school setting? 00:09:22.919 --> 00:09:25.179 Well, look at the debates teachers and administrators 00:09:25.179 --> 00:09:27.220 constantly have. They're always searching for 00:09:27.220 --> 00:09:29.539 the best method. Oh, yeah. Like, is it better 00:09:29.539 --> 00:09:32.700 to use tablets or traditional textbooks? Exactly. 00:09:33.059 --> 00:09:35.779 Is it better to lecture, or should everything 00:09:35.779 --> 00:09:39.279 be project -based group work? Is phonics the 00:09:39.279 --> 00:09:42.419 best way to teach reading, or is it whole language? 00:09:42.519 --> 00:09:44.580 You're constantly looking for the ruling theory, 00:09:44.840 --> 00:09:46.980 that one silver bullet that's going to fix everyone's 00:09:46.980 --> 00:09:49.620 test scores. And Ackerman says that is fundamentally 00:09:49.620 --> 00:09:53.580 the wrong question to ask. If you apply Chamberlain's 00:09:53.580 --> 00:09:56.299 multiple working hypotheses to teaching, you 00:09:56.299 --> 00:09:58.919 stop asking, what is the best method? And what 00:09:58.919 --> 00:10:01.039 do you ask instead? You start asking, what are 00:10:01.039 --> 00:10:03.720 the special values of these different methods? 00:10:03.940 --> 00:10:06.720 Oh, wow. What are the special values? That is 00:10:06.720 --> 00:10:09.879 a huge shift in perspective. It changes a teacher's 00:10:09.879 --> 00:10:12.559 entire toolkit. Instead of trying to prove that 00:10:12.559 --> 00:10:14.779 tablets are the future of learning, you hold 00:10:14.779 --> 00:10:17.480 multiple hypotheses at once. You acknowledge 00:10:17.480 --> 00:10:19.320 that tablets help with interactive engagement, 00:10:19.740 --> 00:10:22.440 but textbooks help with sustained focus, and 00:10:22.440 --> 00:10:24.620 lectures are great for providing structure. You 00:10:24.620 --> 00:10:26.700 keep all those options open and active. Right. 00:10:26.980 --> 00:10:29.759 So when a student is struggling, you aren't stuck 00:10:29.759 --> 00:10:32.139 trying to force your pet theory to work. You're 00:10:32.139 --> 00:10:34.100 not sitting there saying, well, my tablet method 00:10:34.100 --> 00:10:35.960 is the best, so if this kid isn't learning, the 00:10:35.960 --> 00:10:38.480 kid must be the problem. Exactly, you have a 00:10:38.480 --> 00:10:40.759 full array of reasons why a lesson might work 00:10:40.759 --> 00:10:43.799 or might fail. Ackerman writes that when educators 00:10:43.799 --> 00:10:46.500 understand this full array, they are so much 00:10:46.500 --> 00:10:48.659 more likely to find a solution that actually 00:10:48.659 --> 00:10:51.200 helps the student in front of them. It prevents 00:10:51.200 --> 00:10:53.620 you from becoming a hammer that only sees nails. 00:10:53.840 --> 00:10:56.580 Yes. It stops you from being rigid. And honestly, 00:10:56.799 --> 00:11:00.340 this applies to way more than just teaching or 00:11:00.340 --> 00:11:03.500 geology. Oh, absolutely. Troubleshooting code, 00:11:03.980 --> 00:11:06.179 resolving a fight with your spouse, figuring 00:11:06.179 --> 00:11:08.779 out business strategy. If your business is failing, 00:11:09.080 --> 00:11:11.000 you shouldn't just say, well, the market is down. 00:11:11.240 --> 00:11:14.440 Right. You have to ask, is the market down? Or 00:11:14.440 --> 00:11:16.679 is my product outdated? Or is my marketing just 00:11:16.679 --> 00:11:18.549 boring? You have to test all of them. I love 00:11:18.549 --> 00:11:21.549 this concept. I really do. But I have to play 00:11:21.549 --> 00:11:23.529 devil's advocate for a second here. Go for it. 00:11:23.590 --> 00:11:26.169 This sounds completely exhausting. It is. It 00:11:26.169 --> 00:11:28.850 really is. I mean, just cognitively. Human brains 00:11:28.850 --> 00:11:31.429 crave certainty. We want to close the open loops. 00:11:31.490 --> 00:11:33.809 If I'm in a management meeting and I propose 00:11:33.809 --> 00:11:36.169 five conflicting reasons why we're losing money 00:11:36.169 --> 00:11:38.409 and I say we need to investigate all of them 00:11:38.409 --> 00:11:40.389 equally, people are going to look at me like 00:11:40.389 --> 00:11:42.970 I'm crazy or at least indecisive. You're not 00:11:42.970 --> 00:11:44.970 wrong. And Chamberlain actually knew this would 00:11:44.970 --> 00:11:48.389 be a hard sell. He was very upfront about the 00:11:48.389 --> 00:11:50.850 challenges of this method. And he highlighted 00:11:50.850 --> 00:11:53.610 two big downsides. Okay, what's the first one? 00:11:54.009 --> 00:11:57.210 Communication. It is incredibly difficult to 00:11:57.210 --> 00:12:00.669 express multiple complex, conflicting ideas at 00:12:00.669 --> 00:12:03.529 the same time to a general audience. Because 00:12:03.529 --> 00:12:06.350 people just want the bottom line. Yes. They want 00:12:06.350 --> 00:12:09.169 clarity. They don't want a messy list of possibilities. 00:12:09.590 --> 00:12:11.929 They want a hero and a villain. They want the 00:12:11.929 --> 00:12:15.200 one true cause. It's complicated. is rarely a 00:12:15.200 --> 00:12:18.100 winning campaign slogan. Exactly. And the second 00:12:18.100 --> 00:12:20.299 challenge he noted is what he called the problem 00:12:20.299 --> 00:12:22.659 of the uninitiated. Meaning people who are new 00:12:22.659 --> 00:12:25.700 to a topic. Right. He worried that young students 00:12:25.700 --> 00:12:27.779 or adults who are completely unfamiliar with 00:12:27.779 --> 00:12:30.340 a specific field would just get overwhelmed by 00:12:30.340 --> 00:12:32.559 the cognitive load. That makes sense. If I don't 00:12:32.559 --> 00:12:34.759 know the basics of geology and you start telling 00:12:34.759 --> 00:12:39.460 me it's a complex mix of glaciers, rivers, tectonics, 00:12:39.539 --> 00:12:41.639 and maybe a meteor strike, I'm just going to 00:12:41.639 --> 00:12:44.200 check out. Right. Chamberlain thought that maybe 00:12:44.200 --> 00:12:46.639 people need to master the basics of the simple 00:12:46.639 --> 00:12:49.299 ruling theories before they can attempt this 00:12:49.299 --> 00:12:52.539 kind of high wire act of holding multiple hypotheses. 00:12:53.320 --> 00:12:55.360 So there's a real risk of analysis paralysis. 00:12:55.899 --> 00:12:58.460 If you keep every option open all the time, maybe 00:12:58.460 --> 00:13:00.879 you never actually learn or decide anything. 00:13:01.159 --> 00:13:04.480 That was his fear in 1890. But interestingly, 00:13:04.919 --> 00:13:07.279 Ackerman brings in a counterpoint in his blog 00:13:07.279 --> 00:13:09.620 post from a more modern scientific heavyweight. 00:13:09.980 --> 00:13:14.360 Steven J. Gould. Ah, the evolutionary biologist. 00:13:14.460 --> 00:13:16.620 I remember reading some of his essays years ago. 00:13:16.960 --> 00:13:19.720 Yes. Gould wrote extensively for Natural History 00:13:19.720 --> 00:13:21.840 Magazine, and Ackerman points out that Gould 00:13:21.840 --> 00:13:24.600 fundamentally believed we underestimate the public. 00:13:24.759 --> 00:13:27.820 Really? Yeah. Gould believed that lay readers 00:13:27.820 --> 00:13:30.440 are perfectly capable of understanding highly 00:13:30.440 --> 00:13:33.220 sophisticated, nuanced topics. So Gould's basically 00:13:33.220 --> 00:13:35.289 saying, don't dumb it down. Essentially, yes. 00:13:35.669 --> 00:13:37.809 And Ackerman seems to side with Gould on this. 00:13:38.149 --> 00:13:40.289 He suggests that the reason people struggle with 00:13:40.289 --> 00:13:42.389 complexity today isn't because their brains can't 00:13:42.389 --> 00:13:44.190 handle it, it's because we don't teach them how 00:13:44.190 --> 00:13:47.500 to handle it. We feed people. ruling theories 00:13:47.500 --> 00:13:50.720 and single hypotheses for so long that their 00:13:50.720 --> 00:13:53.259 complexity muscles just completely atrophy. That's 00:13:53.259 --> 00:13:55.139 a great way to put it. That is quite a condemning 00:13:55.139 --> 00:13:58.000 thought, actually. We train people from childhood 00:13:58.000 --> 00:14:00.580 to look for the one simple answer, and then as 00:14:00.580 --> 00:14:03.000 adults, we wonder why they can't handle nuance 00:14:03.000 --> 00:14:05.480 in politics or science. And in the age of AI 00:14:05.480 --> 00:14:08.500 and endless information overload, that inability 00:14:08.500 --> 00:14:12.059 to handle nuance is genuinely dangerous. If you 00:14:12.059 --> 00:14:14.059 stop at the first answer that feels good your 00:14:14.059 --> 00:14:16.789 first intellectual offspring, you are very likely 00:14:16.789 --> 00:14:19.230 stopping at a half -truth. It's so interesting 00:14:19.230 --> 00:14:21.610 to think about Ackerman's own journey with this. 00:14:22.129 --> 00:14:24.789 He spends 40 years looking for this one paper 00:14:24.789 --> 00:14:26.870 because he remembered it being life -changing. 00:14:27.190 --> 00:14:29.389 When he finally unearths it, does he feel it 00:14:29.389 --> 00:14:31.909 still holds up today? He absolutely does. In 00:14:31.909 --> 00:14:33.970 fact, he argues it is more relevant now than 00:14:33.970 --> 00:14:36.529 it was in 1890. And there is a real poetry in 00:14:36.529 --> 00:14:38.990 the fact that he used AI to find it. How so? 00:14:39.110 --> 00:14:41.730 Well, think about what AI is. It's the ultimate 00:14:41.730 --> 00:14:44.070 answer engine. You type in a prompt, and it gives 00:14:44.070 --> 00:14:46.570 you an answer. Usually just one definitive answer. 00:14:47.129 --> 00:14:49.950 Exactly. And if we aren't careful, we just treat 00:14:49.950 --> 00:14:54.070 that AI output as our new ruling theory. We accept 00:14:54.070 --> 00:14:56.269 it because the machine authority gave it to us. 00:14:56.909 --> 00:15:00.590 Ackerman is using this 135 -year -old paper To 00:15:00.590 --> 00:15:02.970 remind us that we have to be the ones generating 00:15:02.970 --> 00:15:04.909 the multiple hypotheses Otherwise, we're just 00:15:04.909 --> 00:15:07.049 outsourcing our critical thinking to an algorithm 00:15:07.049 --> 00:15:09.590 an algorithm that might just be Hallucinating 00:15:09.590 --> 00:15:12.590 its own single hypothesis anyway Wow, so bring 00:15:12.590 --> 00:15:15.049 this all back together. We have a geology paper 00:15:15.049 --> 00:15:18.169 from the 1890s teaching us a vital lesson about 00:15:18.169 --> 00:15:21.350 modern education AI and just daily critical thinking 00:15:21.350 --> 00:15:24.870 and the core lesson really seems to be about 00:15:24.870 --> 00:15:27.730 humility. Scientific humility. It's having the 00:15:27.730 --> 00:15:29.929 bravery to say, I have a really great idea, but 00:15:29.929 --> 00:15:31.870 I'm going to deliberately look for reasons why 00:15:31.870 --> 00:15:33.990 I might be wrong, and I'm going to actively look 00:15:33.990 --> 00:15:36.090 for three other explanations at the exact same 00:15:36.090 --> 00:15:38.330 time. It's about being willing to kill your darlings. 00:15:39.090 --> 00:15:41.429 Or, to use Chamberlain's amazing phrase, you 00:15:41.429 --> 00:15:43.490 have to be willing to neglect your intellectual 00:15:43.490 --> 00:15:45.750 offspring. You do, for the greater good of finding 00:15:45.750 --> 00:15:48.330 the actual truth. It's tainful, it goes against 00:15:48.330 --> 00:15:51.210 our nature, but it's entirely necessary. So here 00:15:51.210 --> 00:15:53.070 is the challenge we want to leave you with today. 00:15:53.769 --> 00:15:55.610 We always try to give you something practical 00:15:55.610 --> 00:15:58.289 to do with this information. The next time you 00:15:58.289 --> 00:16:01.789 find yourself entirely sure that you know the 00:16:01.789 --> 00:16:03.970 answer to something, whether it's why your car 00:16:03.970 --> 00:16:07.230 won't start or why a major project at work failed 00:16:07.230 --> 00:16:10.450 or even why a friend is mad at you, I want you 00:16:10.450 --> 00:16:13.289 to just stop. Catch yourself right in that moment 00:16:13.289 --> 00:16:16.929 of absolute certainty. Ask yourself, am I just 00:16:16.929 --> 00:16:19.830 protecting my pet theory right now? And then 00:16:19.929 --> 00:16:22.429 force yourself to write down at least two other 00:16:22.429 --> 00:16:25.289 possible explanations, even if they sound totally 00:16:25.289 --> 00:16:28.110 unlikely at first. Just the physical act of writing 00:16:28.110 --> 00:16:31.070 them down breaks the spell of the single hypothesis. 00:16:31.230 --> 00:16:33.929 It might hurt your ego a little bit, but it might 00:16:33.929 --> 00:16:36.110 also be the only way to actually solve the problem 00:16:36.110 --> 00:16:38.450 in front of you. Thank you so much for diving 00:16:38.450 --> 00:16:40.870 deep with us today. Keep looking for the alternative. 00:16:41.029 --> 00:16:44.230 Keep asking why and what else. See you next time.