WEBVTT 00:00:00.000 --> 00:00:03.819 You know, there's this fear I hear constantly 00:00:03.819 --> 00:00:06.299 these days. It's almost become this mantra in 00:00:06.299 --> 00:00:09.779 coffee shops, in boardrooms. Be careful. Using 00:00:09.779 --> 00:00:12.419 AI is going to make you stupid. Right. Or lazy. 00:00:12.720 --> 00:00:14.740 That your brain's just going to atrophy. Exactly. 00:00:14.820 --> 00:00:17.879 This terrifying image of us just slowly forgetting 00:00:17.879 --> 00:00:20.179 how to think because the machines are doing it 00:00:20.179 --> 00:00:22.760 all for us. It really is the prevailing anxiety 00:00:22.760 --> 00:00:25.620 of our time. But here's the twist. And this is 00:00:25.620 --> 00:00:27.300 really the core of what we're looking at today. 00:00:28.070 --> 00:00:31.510 AI doesn't inherently make you worse. It doesn't 00:00:31.510 --> 00:00:34.170 rot your brain. It's the way you use it that 00:00:34.170 --> 00:00:37.049 keeps you stagnant. The danger isn't the tool. 00:00:37.270 --> 00:00:39.890 It's using it as a crutch instead of a turbocharger. 00:00:40.229 --> 00:00:43.189 That is a critical distinction. So welcome to 00:00:43.189 --> 00:00:45.929 the deep dive. Today we are tackling mastering 00:00:45.929 --> 00:00:48.270 the mind in the AI age. We're exploring how to 00:00:48.270 --> 00:00:51.579 move from that passive consumption to to active 00:00:51.579 --> 00:00:53.600 engagement. Exactly. And we've got a roadmap. 00:00:53.719 --> 00:00:56.079 We're going to look at why AI lies to us and 00:00:56.079 --> 00:00:58.619 why it's not technically lying. We'll tackle 00:00:58.619 --> 00:01:01.200 this illusion of fluency, which is a fascinating 00:01:01.200 --> 00:01:03.759 psychological trap. Then we'll break down a kind 00:01:03.759 --> 00:01:06.180 of re -imagined pyramid of thinking to see what 00:01:06.180 --> 00:01:08.599 you should outsource and what you absolutely 00:01:08.599 --> 00:01:11.299 have to keep. And the practical part. And finally, 00:01:11.620 --> 00:01:15.280 a specific four -step routine to learn with AI, 00:01:15.700 --> 00:01:17.879 not by AI. I like the sound of that. Learning 00:01:17.879 --> 00:01:21.359 with, not by. So let's start at that foundational 00:01:21.359 --> 00:01:24.359 level, the nature of the beast. Why is the information 00:01:24.359 --> 00:01:27.120 we get from these large language models so often 00:01:27.120 --> 00:01:30.799 just wrong? Yeah, it really comes down to probability. 00:01:30.939 --> 00:01:33.939 We tend to anthropomorphize these models. We 00:01:33.939 --> 00:01:35.879 think they know the truth. That they don't. They 00:01:35.879 --> 00:01:38.359 don't. Whether it's ChatGPT or Claude, they're 00:01:38.359 --> 00:01:40.400 just predicting what word should reasonably come 00:01:40.400 --> 00:01:42.719 next to make a sentence sound good. So it's optimizing 00:01:42.719 --> 00:01:46.239 for plausibility, not for truth. It's basically 00:01:46.239 --> 00:01:49.430 a... super advanced autocomplete. Precisely. 00:01:49.750 --> 00:01:52.189 The source material calls this the hallucination 00:01:52.189 --> 00:01:55.269 phenomenon. The AI is trained to be helpful and 00:01:55.269 --> 00:01:56.870 confident, so if it doesn't know an answer, it 00:01:56.870 --> 00:01:58.349 might just make something up that looks like 00:01:58.349 --> 00:02:01.230 an answer, just to please you. It's like a people 00:02:01.230 --> 00:02:02.849 -pleasing intern who's afraid to say, I don't 00:02:02.849 --> 00:02:06.349 know. That is the perfect analogy. It wants that 00:02:06.349 --> 00:02:09.770 gold star, so it invents a citation. And there's 00:02:09.770 --> 00:02:11.870 another layer here. It's called source blindness. 00:02:12.110 --> 00:02:14.830 Source blindness. Yeah, the AI reads billions 00:02:14.830 --> 00:02:17.590 of websites, but it struggles to tell the difference 00:02:17.590 --> 00:02:21.530 between a peer -reviewed article and a random 00:02:21.530 --> 00:02:24.169 angry comment on a forum. It just ingests it 00:02:24.169 --> 00:02:27.650 all. That sounds dangerous if you're trying to 00:02:27.650 --> 00:02:30.229 learn something complex. Yeah. If it's just blending 00:02:30.229 --> 00:02:32.650 fact and fiction, how do we use it safely? You 00:02:32.650 --> 00:02:35.189 need a strategy. The guide we're looking at suggests 00:02:35.189 --> 00:02:37.810 a risk classification system. Think of it like 00:02:37.810 --> 00:02:40.689 triage. First, you've got low -risk topics. OK, 00:02:40.770 --> 00:02:43.520 like what? Simple stuff. How to cook rice. What 00:02:43.520 --> 00:02:45.560 is Newton's second law? Things where there's 00:02:45.560 --> 00:02:47.780 a huge consensus. So the data is overwhelmingly 00:02:47.780 --> 00:02:49.599 consistent. Right. You can probably trust it. 00:02:49.599 --> 00:02:52.719 But then you have high -risk topics, new fields, 00:02:53.139 --> 00:02:56.759 deep political debates, specific medical advice. 00:02:57.080 --> 00:02:59.419 For these, the AI should only be for suggestions. 00:02:59.479 --> 00:03:01.400 You have to check the output against textbooks 00:03:01.400 --> 00:03:04.360 or professionals. So this implies the strategy 00:03:04.360 --> 00:03:06.460 isn't just about fact -checking everything after 00:03:06.460 --> 00:03:09.419 the fact. No. It's about triaging information 00:03:09.419 --> 00:03:12.219 based on complexity. before you even type the 00:03:12.219 --> 00:03:14.740 prompt. You have to ask, if this is wrong, will 00:03:14.740 --> 00:03:18.460 I know? That's a great heuristic. But let's say 00:03:18.460 --> 00:03:20.960 the AI gets it right. The answer is accurate. 00:03:22.159 --> 00:03:24.780 There's still this problem. The source calls 00:03:24.780 --> 00:03:28.250 it the illusion of fluency. And this one. This 00:03:28.250 --> 00:03:30.110 really resonated with me. Oh, this is the trap, 00:03:30.189 --> 00:03:32.550 the big one. It's that feeling of being incredibly 00:03:32.550 --> 00:03:34.729 smart while you're chatting with the bot. Yeah. 00:03:35.009 --> 00:03:36.770 You ask a question. It gives this brilliant, 00:03:36.990 --> 00:03:38.770 nuanced answer. And you're nodding along, thinking, 00:03:38.770 --> 00:03:40.889 yes, exactly. I get it. You feel fluent. But 00:03:40.889 --> 00:03:43.889 then you close the laptop. And the mind goes 00:03:43.889 --> 00:03:46.930 blank. A complete blank page. You mistook access 00:03:46.930 --> 00:03:49.409 to the answer for mastery of the answer. The 00:03:49.409 --> 00:03:51.409 source makes a really hard distinction here between 00:03:51.409 --> 00:03:54.030 productivity and learning. OK. Productivity is 00:03:54.030 --> 00:03:56.270 getting an essay written in two minutes. Task 00:03:56.270 --> 00:04:00.080 done. But your growth? Zero. It's like watching 00:04:00.080 --> 00:04:02.020 a workout video and thinking you got stronger. 00:04:02.439 --> 00:04:05.180 Yes. You watched the heavy lifting. You didn't 00:04:05.180 --> 00:04:08.199 do it. Real learning needs that mental sweat. 00:04:08.759 --> 00:04:10.939 When you write an essay yourself, you're wrestling 00:04:10.939 --> 00:04:13.960 with words. You're structuring arguments. You're 00:04:13.960 --> 00:04:16.720 literally building neural pathways. And if you 00:04:16.720 --> 00:04:19.060 let the AI do that, you're just robbing yourself 00:04:19.060 --> 00:04:21.560 of the cognitive workout. You are. The source 00:04:21.560 --> 00:04:24.660 breaks this into two types of reliance. There's 00:04:24.660 --> 00:04:27.860 positive reliance. The booster. OK. That's where 00:04:27.860 --> 00:04:31.160 you outsource low value stuff. Summarizing a 00:04:31.160 --> 00:04:34.639 long transcript, fixing grammar, it saves your 00:04:34.639 --> 00:04:36.660 mental energy for the real analysis. And the 00:04:36.660 --> 00:04:39.579 other one. Negative reliance. The crutch. That's 00:04:39.579 --> 00:04:41.779 letting the AI think from start to finish. I 00:04:41.779 --> 00:04:43.699 think we've all been guilty of the crutch approach 00:04:43.699 --> 00:04:46.040 when we're tired or rushing. Just write this 00:04:46.040 --> 00:04:48.259 email for me. So how do we catch ourselves? Don't 00:04:48.259 --> 00:04:51.459 count tasks completed. Track your internal understanding. 00:04:51.800 --> 00:04:53.939 The source suggests the five -year -old test. 00:04:54.480 --> 00:04:56.480 Can you explain the topic to a child without 00:04:56.480 --> 00:04:58.980 looking at a screen? Ah, the Feynman technique. 00:04:59.319 --> 00:05:01.379 Exactly. If you get stuck and you have to open 00:05:01.379 --> 00:05:04.399 chat GPT to find the words, you don't really 00:05:04.399 --> 00:05:06.139 know it. You're just renting the knowledge. You 00:05:06.139 --> 00:05:08.259 don't own it. So does this mean we should just 00:05:08.259 --> 00:05:11.819 stop using AI for any kind of output? No, not 00:05:11.819 --> 00:05:13.740 at all. Just don't let it replace the thinking 00:05:13.740 --> 00:05:16.480 process itself. OK, let's unpack that, the thinking 00:05:16.480 --> 00:05:19.240 process. Because thinking is such a broad term. 00:05:19.480 --> 00:05:22.819 The source material introduces a hierarchy, a 00:05:22.819 --> 00:05:26.279 sort of reimagined Bloons taxonomy for the AI 00:05:26.279 --> 00:05:28.939 age. Right. Picture a pyramid. The bottom three 00:05:28.939 --> 00:05:32.160 levels are remember, understand, and basic apply. 00:05:32.699 --> 00:05:36.360 So facts, basic meanings, following steps. Exactly. 00:05:36.680 --> 00:05:39.000 And the argument is, let the robots have these. 00:05:39.120 --> 00:05:41.660 Don't waste your limited cognitive load memorizing 00:05:41.660 --> 00:05:44.000 figures you can look up in three seconds. AI 00:05:44.000 --> 00:05:46.100 is great at the bottom of the pyramid. It's perfect 00:05:46.100 --> 00:05:48.779 recall. Perfect recall. But the top three levels, 00:05:49.019 --> 00:05:50.980 that's the human advantage. OK, walk us through 00:05:50.980 --> 00:05:54.060 those. What's left for us? First, analyze. This 00:05:54.060 --> 00:05:56.259 is about connecting the dots. So don't just ask 00:05:56.259 --> 00:05:59.040 for a definition of SEO. Ask yourself, how is 00:05:59.040 --> 00:06:02.100 SEO different from Facebook ads in the specific 00:06:02.100 --> 00:06:04.600 context of my new coffee shop? The context is 00:06:04.600 --> 00:06:07.050 the key. The AI knows the definition, but I know 00:06:07.050 --> 00:06:10.430 my business. Right. Then above that is evaluate. 00:06:11.310 --> 00:06:14.329 Judgment. An AI can give you a list of pros and 00:06:14.329 --> 00:06:16.709 cons for a decision, but it can't tell you what's 00:06:16.709 --> 00:06:19.370 suitable for your values, your risk tolerance. 00:06:19.949 --> 00:06:22.129 It lacks skin in the game. And at the very top 00:06:22.129 --> 00:06:24.509 of the pyramid. Create. Making something new. 00:06:25.009 --> 00:06:28.490 Now AI creates by remixing old data. It's a synthesis 00:06:28.490 --> 00:06:32.430 engine. But humans. Humans can have breakthroughs 00:06:32.430 --> 00:06:35.550 that defy the data. because we have lived experience. 00:06:35.730 --> 00:06:37.750 So looking at this pyramid, do you think AI can 00:06:37.750 --> 00:06:40.110 eventually climb that ladder? Can it ever truly 00:06:40.110 --> 00:06:43.230 create? Maybe one day. But right now, it lacks 00:06:43.230 --> 00:06:46.209 that real experience and context. It mimics creation, 00:06:46.529 --> 00:06:48.370 but it doesn't understand the emotional weight 00:06:48.370 --> 00:06:50.610 behind it. That brings us to the how. We've got 00:06:50.610 --> 00:06:52.569 the theory. We know we need to stay at the top 00:06:52.569 --> 00:06:55.110 of the pyramid. But the source outlines a very 00:06:55.110 --> 00:06:58.230 specific four -step routine for smart learning. 00:06:58.389 --> 00:07:00.389 Yeah, this routine is gold. It's designed to 00:07:00.389 --> 00:07:02.189 force that mental sweat we were talking about. 00:07:02.329 --> 00:07:05.170 Step one is background and big picture. You use 00:07:05.170 --> 00:07:07.790 the AI to scan a ton of data and just summarize 00:07:07.790 --> 00:07:10.689 the main concept. OK, so a prompt like, summarize 00:07:10.689 --> 00:07:12.990 behavioral economics with daily life examples. 00:07:13.509 --> 00:07:16.000 Right. get the lay of the land. But then, and 00:07:16.000 --> 00:07:18.920 this is so important, step two is the analog 00:07:18.920 --> 00:07:22.339 step. Analog, you mean like paper? Physical paper, 00:07:22.639 --> 00:07:25.839 a pen. You actually stop using the AI. You step 00:07:25.839 --> 00:07:28.079 away from the screen and you draw a mind map. 00:07:28.379 --> 00:07:31.000 You force your own brain to find the connections. 00:07:31.519 --> 00:07:35.459 How does concept A relate to concept B? Why is 00:07:35.459 --> 00:07:37.860 that physical paper step so important? I mean, 00:07:37.939 --> 00:07:39.399 couldn't I just do that in a different window 00:07:39.399 --> 00:07:41.720 or something? You could, but the physical act 00:07:41.720 --> 00:07:44.600 changes things. It forces active recall without 00:07:44.600 --> 00:07:47.100 digital crutches. When the screen is off, your 00:07:47.100 --> 00:07:49.519 brain has to struggle to retrieve the information, 00:07:49.779 --> 00:07:52.259 and that struggle is where the memory gets cemented. 00:07:52.360 --> 00:07:54.399 Okay, so we struggled. We have our messy handwritten 00:07:54.399 --> 00:07:57.120 mind map. What's step three? Step three is the 00:07:57.120 --> 00:08:00.139 Socratic Challenger. You go back to the AI, but 00:08:00.139 --> 00:08:02.740 you don't ask for answers. You tell it... I'm 00:08:02.740 --> 00:08:04.779 explaining the anchoring effect for negotiation, 00:08:05.279 --> 00:08:07.600 act as a tough expert, and ask me three difficult 00:08:07.600 --> 00:08:10.360 questions to test my logic. That's intimidating. 00:08:10.439 --> 00:08:12.339 You're inviting it to criticize you. It should 00:08:12.339 --> 00:08:15.180 be. Whoa. I mean, imagine having Socrates in 00:08:15.180 --> 00:08:17.259 your pocket, ready to poke holes in your logic 00:08:17.259 --> 00:08:19.740 at any moment. That is powerful. Right. Most 00:08:19.740 --> 00:08:21.939 people use AI to tell them they're right. The 00:08:21.939 --> 00:08:24.139 smart learner says, tell me where I'm wrong. 00:08:24.920 --> 00:08:28.769 And step four. Refine and create. You take the 00:08:28.769 --> 00:08:31.370 feedback from that Socratic session and you fix 00:08:31.370 --> 00:08:34.909 your mistakes. This is the create level. By now, 00:08:34.909 --> 00:08:36.529 you're not just repeating what the bot said. 00:08:36.570 --> 00:08:38.529 You've processed it, you've connected it, you've 00:08:38.529 --> 00:08:40.750 defended it, and you've refined it. You own it 00:08:40.750 --> 00:08:44.610 now. It effectively gamifies the learning process. 00:08:44.669 --> 00:08:46.450 It really just changes the dynamic completely. 00:08:46.490 --> 00:08:49.990 You remain the boss of your own brain. We are 00:08:49.990 --> 00:08:52.330 back. I want to shift gears a little bit to the 00:08:52.330 --> 00:08:54.470 professional side of this. There's a section 00:08:54.470 --> 00:08:57.960 in the source material about career sabotage, 00:08:58.399 --> 00:09:01.799 which it sounds dramatic, but the logic really 00:09:01.799 --> 00:09:03.840 holds up. It's a serious warning. Companies aren't 00:09:03.840 --> 00:09:05.940 going to pay high salaries for average reports 00:09:05.940 --> 00:09:08.779 that ChatGPT can generate for free. If your output 00:09:08.779 --> 00:09:11.299 looks exactly like the AI's output, you are. 00:09:11.440 --> 00:09:13.779 You're redundant. So the value shifts from doing 00:09:13.779 --> 00:09:17.120 the work to judging the work. Exactly. Critical 00:09:17.120 --> 00:09:19.559 thinking becomes the strongest habit. You have 00:09:19.559 --> 00:09:22.759 to constantly doubt the AI. Ask, why did it choose 00:09:22.759 --> 00:09:25.960 this word? What perspective is missing? You need 00:09:25.960 --> 00:09:28.259 to be able to smell when the AI is wrong, and 00:09:28.259 --> 00:09:30.539 that requires deep expertise. You know, I have 00:09:30.539 --> 00:09:34.179 to admit, I still wrestle with prompt drift myself. 00:09:34.320 --> 00:09:36.200 I'll start with really good intentions, but then 00:09:36.200 --> 00:09:38.940 I get lazy and just type, you know, write a budget 00:09:38.940 --> 00:09:41.460 for me. I think we all do, but that is the weak 00:09:41.460 --> 00:09:44.080 professional move. The source outlines a formula. 00:09:44.259 --> 00:09:47.799 It's role plus context plus task plus goal, but 00:09:47.799 --> 00:09:50.179 with a specific twist they recommend. What's 00:09:50.179 --> 00:09:52.759 the twist? Constraints, specifically telling 00:09:52.759 --> 00:09:56.149 the AI to wait. Wait? Yeah. Instead of make me 00:09:56.149 --> 00:09:58.690 a budget, you try this. You are a senior financial 00:09:58.690 --> 00:10:01.230 consultant. I earn $1 ,000 a month, and I want 00:10:01.230 --> 00:10:04.190 to save $5 ,000. Do not give me a plan yet. First, 00:10:04.309 --> 00:10:06.690 ask me five critical questions about my habits 00:10:06.690 --> 00:10:09.330 before giving a plan. Oh, that's good. Do not 00:10:09.330 --> 00:10:11.730 give me a plan yet. You're forcing it to gather 00:10:11.730 --> 00:10:13.990 the context it doesn't have. That is the key. 00:10:14.149 --> 00:10:16.830 You are treating the AI like a talented intern. 00:10:17.149 --> 00:10:19.769 If you just say make a budget, the intern guesses. 00:10:20.149 --> 00:10:22.870 If you say interview me first, the intern learns. 00:10:23.210 --> 00:10:25.690 So the quality of the answer depends entirely 00:10:25.690 --> 00:10:29.230 on the setup? 100%. Treat AI like a talented 00:10:29.230 --> 00:10:32.750 intern, not a magic button. Okay, let's ground 00:10:32.750 --> 00:10:35.370 this with some quick fire examples. The source 00:10:35.370 --> 00:10:38.570 breaks this down for vocabulary, reading, and 00:10:38.570 --> 00:10:41.230 coding. Yeah, let's take vocabulary. The lazy 00:10:41.230 --> 00:10:43.870 way is to ask for a definition. You read it, 00:10:43.870 --> 00:10:46.750 you forget it. The smart way. Ask the AI to write 00:10:46.750 --> 00:10:49.110 a funny story with the word, but leave a blank 00:10:49.110 --> 00:10:51.009 space where the word should be. You have to fill 00:10:51.009 --> 00:10:53.269 it in. Active recall again. You're participating. 00:10:53.470 --> 00:10:55.990 Exactly. For reading books. Don't just ask for 00:10:55.990 --> 00:10:57.750 a summary. Pick the hardest chapter, the one 00:10:57.750 --> 00:11:00.289 you didn't quite get, and ask the AI to explain 00:11:00.289 --> 00:11:03.529 it using metaphors, like explain quantum entanglement 00:11:03.529 --> 00:11:06.070 using a pair of dice. That connects the new hard 00:11:06.070 --> 00:11:08.529 information to something you already know. And 00:11:08.529 --> 00:11:11.840 what about coding? high stakes. The super prompt 00:11:11.840 --> 00:11:14.740 for coding is fantastic. Usually people paste 00:11:14.740 --> 00:11:17.299 their error and say fix this. The smart prompt 00:11:17.299 --> 00:11:20.919 is here is my code. Do not give me the correct 00:11:20.919 --> 00:11:23.740 code yet. Tell me where my logic is wrong and 00:11:23.740 --> 00:11:26.639 suggest three technical keywords so I can research 00:11:26.639 --> 00:11:30.000 the fix. That is strict. It forces you to actually 00:11:30.000 --> 00:11:32.000 go do the research yourself. It prevents you 00:11:32.000 --> 00:11:34.039 from just copy pasting your way into a broken 00:11:34.039 --> 00:11:36.120 product. You have to understand the why behind 00:11:36.120 --> 00:11:38.720 the bug. What stands out to me here is the common 00:11:38.720 --> 00:11:41.750 thread. whether it's vocabulary or budgeting 00:11:41.750 --> 00:11:45.070 or coding. Yeah. What is it? They all force the 00:11:45.070 --> 00:11:47.610 brain to do the heavy lifting. Every single one 00:11:47.610 --> 00:11:49.850 of these strategies inserts a pause where the 00:11:49.850 --> 00:11:52.269 human has to think. It really reframes the whole 00:11:52.269 --> 00:11:54.750 relationship. We think of AI as an accelerator, 00:11:54.750 --> 00:11:57.649 you know, to speed us up. But these strategies 00:11:57.649 --> 00:11:59.970 are about deliberately slowing down to make sure 00:11:59.970 --> 00:12:02.470 you're learning. Speed without direction is just 00:12:02.470 --> 00:12:05.070 getting lost faster. So let's bring this all 00:12:05.070 --> 00:12:07.220 together. What is the big takeaway? for someone 00:12:07.220 --> 00:12:09.240 listening who just feels overwhelmed by all this. 00:12:09.480 --> 00:12:12.480 I think it's that core metaphor. AI is a powerful 00:12:12.480 --> 00:12:15.240 engine. It's got incredible horsepower. But you 00:12:15.240 --> 00:12:17.019 must be the driver holding the steering wheel. 00:12:17.159 --> 00:12:19.639 If you let go, you're either going to crash or 00:12:19.639 --> 00:12:21.240 you're just going to end up wherever the algorithm 00:12:21.240 --> 00:12:23.139 takes you. And to keep your hands on the wheel, 00:12:23.340 --> 00:12:27.659 we have those golden rules. Right. Use AI for 00:12:27.659 --> 00:12:31.639 the heavy, tedious work, the summaries, the grammar. 00:12:32.139 --> 00:12:36.220 Use it for simple topics, but never ever Let 00:12:36.220 --> 00:12:39.220 AI decide for you on deep analysis or career 00:12:39.220 --> 00:12:41.960 choices, and always check the facts. It's a call 00:12:41.960 --> 00:12:44.179 to action, really. Don't just use AI to finish 00:12:44.179 --> 00:12:46.179 your work. Use it to get better at your work. 00:12:46.320 --> 00:12:48.559 That's it. The future belongs to those who can 00:12:48.559 --> 00:12:51.399 combine machine speed with human deep thinking. 00:12:52.100 --> 00:12:54.000 Be the hybrid. And I want to leave you with a 00:12:54.000 --> 00:12:56.720 thought to mull over. We talked about AI as a 00:12:56.720 --> 00:12:59.659 tutor or an intern, but what if you treated it 00:12:59.659 --> 00:13:02.470 as a rival? What if, for one week, you're trying 00:13:02.470 --> 00:13:04.750 to outthink the machine on every single topic, 00:13:05.070 --> 00:13:06.889 using it only to grade your own performance? 00:13:07.610 --> 00:13:09.590 How sharp do you think you would get? That is 00:13:09.590 --> 00:13:11.570 a fascinating experiment. Thanks for diving in 00:13:11.570 --> 00:13:13.110 with us. We'll see you next time. See ya.