WEBVTT 00:00:00.000 --> 00:00:03.560 How do you, how do you genuinely stop guessing 00:00:03.560 --> 00:00:05.960 in the stock market? Pete, I mean, what if you 00:00:05.960 --> 00:00:09.400 could take a simple idea, a hypothesis, use AI 00:00:09.400 --> 00:00:12.119 tools, you know, consumer level stuff, and find 00:00:12.119 --> 00:00:14.660 companies that did way better than the S &P 500, 00:00:15.160 --> 00:00:17.379 like more than double? That is exactly the puzzle. 00:00:17.559 --> 00:00:19.620 Yeah. That's what we're digging into today. We're 00:00:19.620 --> 00:00:22.600 looking at a specific framework that historically, 00:00:22.760 --> 00:00:26.829 anyway. generated 22 .4 % returns. Compare that 00:00:26.829 --> 00:00:30.629 to the S &P 500's 9%. And that difference, that 00:00:30.629 --> 00:00:33.170 edge, it came from using tools you probably have 00:00:33.170 --> 00:00:36.149 opened right now. like on your desktop. So welcome, 00:00:36.289 --> 00:00:38.350 everyone, to this deep dive. We're focusing on 00:00:38.350 --> 00:00:41.049 these AI -driven investing frameworks. Our mission, 00:00:41.070 --> 00:00:43.030 really, is to cut through the hype and show you 00:00:43.030 --> 00:00:46.409 exactly how tools like Chat GPT and Google Sheets, 00:00:46.490 --> 00:00:48.390 things most people can access, can actually run 00:00:48.390 --> 00:00:50.670 some serious systematic research. Right. We'll 00:00:50.670 --> 00:00:52.549 start with the core concept. First, this NFL 00:00:52.549 --> 00:00:54.969 strategy. Then we'll get into the nitty gritty, 00:00:55.030 --> 00:00:57.170 the data filtering, the scoring, and then the 00:00:57.170 --> 00:00:59.329 really practical bit. We're giving you these 00:00:59.329 --> 00:01:01.509 deep multi -layered prompts you can take and 00:01:01.509 --> 00:01:03.869 use right away for your own research. Okay, let's 00:01:03.869 --> 00:01:08.170 start with the hypothesis itself. NEFL. The big 00:01:08.170 --> 00:01:11.629 idea seems pretty straightforward. Combine strong 00:01:11.629 --> 00:01:14.650 business protection with dedicated management. 00:01:15.329 --> 00:01:17.870 So network effects plus founder -led companies. 00:01:18.069 --> 00:01:21.109 That's NEFL. Exactly. And the network effect 00:01:21.109 --> 00:01:24.450 part, that's what builds the moat, the economic 00:01:24.450 --> 00:01:27.209 protection. It's that simple idea. the product 00:01:27.209 --> 00:01:29.549 gets better, more valuable as more people use 00:01:29.549 --> 00:01:32.090 it. That distinction is really key, isn't it? 00:01:32.629 --> 00:01:35.609 Like, Visa or MasterCard, every time a new store 00:01:35.609 --> 00:01:38.170 accepts the card, it's better for everyone who 00:01:38.170 --> 00:01:40.489 has the card. Mm -hmm, and vice versa. It's this 00:01:40.489 --> 00:01:43.010 loop, this self -reinforcing thing that makes 00:01:43.010 --> 00:01:45.189 it incredibly tough for someone new to just jump 00:01:45.189 --> 00:01:47.510 in and compete at the same scale. Yeah, it builds 00:01:47.510 --> 00:01:49.569 this huge wall around the business. Yeah. Protects 00:01:49.569 --> 00:01:52.049 that market position potentially for, you know, 00:01:52.290 --> 00:01:53.909 decades. That's the kind of structural advantage 00:01:53.909 --> 00:01:56.120 you're looking for. And the second piece, founder 00:01:56.120 --> 00:01:58.599 led. That anchors the vision, you said. Why does 00:01:58.599 --> 00:02:00.920 who's running the show matter so much here? Well, 00:02:01.060 --> 00:02:02.640 founders just have more skin in the game, right? 00:02:02.739 --> 00:02:06.299 They're not just a hired CEO maybe looking at 00:02:06.299 --> 00:02:08.680 the next quarter's bonus. They actually built 00:02:08.680 --> 00:02:11.060 the company. They often own a big chunk of it. 00:02:11.159 --> 00:02:13.560 So they're naturally thinking about the 10 year, 00:02:13.620 --> 00:02:16.300 the 20 year future, not just hitting the next 00:02:16.300 --> 00:02:18.639 earnings target. You think of someone like maybe 00:02:18.639 --> 00:02:21.259 Mark Zuckerberg at Metta. Still driving that 00:02:21.259 --> 00:02:23.740 long term vision, even when the market gets noisy 00:02:23.740 --> 00:02:27.199 in the short term. Exactly. So the NFL idea is 00:02:27.199 --> 00:02:29.340 combine that structural defense, the network 00:02:29.340 --> 00:02:32.400 effect, with that kind of dedicated long view 00:02:32.400 --> 00:02:37.610 leadership. combo should theoretically lead to 00:02:37.610 --> 00:02:40.229 companies that consistently outperform. OK, but 00:02:40.229 --> 00:02:42.210 if these companies have such obvious strengths, 00:02:42.449 --> 00:02:44.610 these moats and dedicated founders, why doesn't 00:02:44.610 --> 00:02:46.449 the market just price that in immediately? Why 00:02:46.449 --> 00:02:48.270 aren't they always expensive? Well, because the 00:02:48.270 --> 00:02:50.949 market often misses the nuance. It gets caught 00:02:50.949 --> 00:02:53.530 up in short -term stuff. So that inefficiency, 00:02:53.530 --> 00:02:56.169 that's the opportunity. Precisely. And that's 00:02:56.169 --> 00:02:59.169 where AI becomes this massive time saver. Which 00:02:59.169 --> 00:03:01.490 brings us to step one, the initial screening. 00:03:01.659 --> 00:03:04.840 Right. Instead of spending weeks, maybe months, 00:03:05.319 --> 00:03:07.759 reading through hundreds of company reports manually, 00:03:08.460 --> 00:03:10.560 we used a single prompt. Just natural language. 00:03:10.740 --> 00:03:13.439 Yeah. We basically asked the AI, find me all 00:03:13.439 --> 00:03:15.479 the S &P 500 companies that have both network 00:03:15.479 --> 00:03:17.979 effects, A &D, or still founder -led. Simple 00:03:17.979 --> 00:03:20.259 as that. And boom, it came back with 26 companies 00:03:20.259 --> 00:03:22.840 instantly. That speed, that efficiency gain, 00:03:22.840 --> 00:03:25.819 that's the first big win. Yeah. And I still wrestle 00:03:25.819 --> 00:03:28.120 with prompt drift myself sometimes, getting the 00:03:28.120 --> 00:03:31.680 AI to stay on track. But for just raw data gathering, 00:03:32.060 --> 00:03:34.900 pulling lists based on clear criteria, AI is 00:03:34.900 --> 00:03:36.979 just fantastic. It saves so much time. Totally. 00:03:37.479 --> 00:03:39.879 So then we set up the back test scenario. OK, 00:03:39.879 --> 00:03:42.560 what if we put $100 into each of those 26 companies 00:03:42.560 --> 00:03:44.560 right when they went public on their IPO date? 00:03:44.900 --> 00:03:48.319 Total imaginary investment. $600. And then you 00:03:48.319 --> 00:03:50.240 compare that portfolio's performance against 00:03:50.240 --> 00:03:52.719 the S &P 500 index over the exact same time frame. 00:03:52.919 --> 00:03:54.439 It's crucial to note, though, the results are 00:03:54.439 --> 00:03:56.419 all over the place for individual stocks, right? 00:03:56.780 --> 00:03:59.560 Like Netflix was astronomical. $100 turned into, 00:03:59.560 --> 00:04:01.699 what, over $100 ,000? Yeah, insane. But then 00:04:01.699 --> 00:04:04.599 you had others, like Snap, where the $100 actually 00:04:04.599 --> 00:04:07.500 lost value. And that variance that really highlights 00:04:07.500 --> 00:04:09.780 why you need more filtering. You need to go deeper, 00:04:10.060 --> 00:04:12.139 which is exactly what the broad market often 00:04:12.139 --> 00:04:15.240 fails to do. So how critical was nailing down 00:04:15.240 --> 00:04:18.879 that exact historical IPO date data? Was that 00:04:18.879 --> 00:04:20.980 tough to get right for the backtest accuracy? 00:04:21.199 --> 00:04:23.339 Oh, absolutely. Getting that correct historical 00:04:23.339 --> 00:04:26.439 data is key. Otherwise, the comparison isn't 00:04:26.439 --> 00:04:29.759 fair. OK, so that brings us to the maybe the 00:04:29.759 --> 00:04:33.850 most interesting part. Step two. Adding nuance 00:04:33.850 --> 00:04:36.629 with this 1 to 10 scoring system, because like 00:04:36.629 --> 00:04:38.949 you said, the real world isn't just yes or no 00:04:38.949 --> 00:04:41.129 for something like network effect. Exactly. That's 00:04:41.129 --> 00:04:42.850 a really powerful point. We had to figure out, 00:04:43.050 --> 00:04:45.470 OK, which companies have a kind of weak, maybe 00:04:45.470 --> 00:04:47.189 questionable network effect, let's say a 3 or 00:04:47.189 --> 00:04:50.029 4, versus those with a truly dominant, almost 00:04:50.029 --> 00:04:52.589 unbreakable one, a 9 or 10. And asking for a 00:04:52.589 --> 00:04:54.470 score, not just a yes or no, that helps with 00:04:54.470 --> 00:04:56.649 accuracy too, right? Less risk of the AI just 00:04:56.649 --> 00:04:58.649 making something up. That's the thinking, yeah. 00:04:59.259 --> 00:05:02.259 when you forced it into a binary, yes, no, the 00:05:02.259 --> 00:05:05.259 risk of hallucination, as they call it, seems 00:05:05.259 --> 00:05:08.160 higher. It might give a confident yes based on 00:05:08.160 --> 00:05:10.980 weak evidence, but asking for a score like one 00:05:10.980 --> 00:05:14.600 to 10, it forces the AI to weigh different factors 00:05:14.600 --> 00:05:18.399 to give a more measured judgment. We found that 00:05:18.399 --> 00:05:20.399 dramatically increased the reliability. And the 00:05:20.399 --> 00:05:22.839 tool for this wasn't some super expensive finance 00:05:22.839 --> 00:05:25.519 terminal. It was just Google Sheets add -ons. 00:05:25.819 --> 00:05:27.740 Pretty much, yeah. Free add -ons that let you 00:05:27.740 --> 00:05:29.480 run AI prompts right there in the spreadsheet 00:05:29.480 --> 00:05:32.000 cells. Super accessible. OK, so what did the 00:05:32.000 --> 00:05:33.920 prompt look like? It was basically a formula. 00:05:34.639 --> 00:05:37.300 Tell me on a scale of 1 to 10 how much company 00:05:37.300 --> 00:05:39.839 name benefits from network effects. 1 is not 00:05:39.839 --> 00:05:42.699 at all. 10 is very strong network effects. Simple. 00:05:42.839 --> 00:05:45.560 Whoa. I mean, just imagine scaling that, running 00:05:45.560 --> 00:05:48.339 that kind of analysis across all 500 SMT companies 00:05:48.339 --> 00:05:51.339 yourself would take forever. But doing it with 00:05:51.339 --> 00:05:54.639 these tools? That's incredible power for an individual 00:05:54.639 --> 00:05:56.660 investor. Yeah, it took about 30 minutes total. 00:05:56.680 --> 00:05:58.759 Yeah. And gave us these detailed ratings for 00:05:58.759 --> 00:06:00.879 every single S &P 500 company that costs with 00:06:00.879 --> 00:06:03.879 maybe 15 bucks in API credits. It's just insane 00:06:03.879 --> 00:06:06.259 data efficiency. But hang on, how did you check 00:06:06.259 --> 00:06:09.579 those scores? How do you know the AI wasn't just 00:06:09.579 --> 00:06:12.279 guessing a number? How do you validate that 1 00:06:12.279 --> 00:06:14.540 to 10 data point before you relied on it? Ah, 00:06:14.620 --> 00:06:16.920 good question. That's always the challenge with 00:06:16.920 --> 00:06:19.199 generative AI and finance, right? Validation. 00:06:19.680 --> 00:06:22.019 We did it by spot -checking the extremes. We 00:06:22.019 --> 00:06:24.139 looked at the companies the AI scored highest, 00:06:24.399 --> 00:06:27.860 like Visa, Adobe, the obvious ones, and checked 00:06:27.860 --> 00:06:30.920 that against, you know, consensus expert opinions, 00:06:31.220 --> 00:06:33.160 financial reports. Okay. And then we checked 00:06:33.160 --> 00:06:35.779 the low -scoring ones, too. We found the AI was 00:06:35.779 --> 00:06:37.620 generally pretty good at applying the definitions 00:06:37.620 --> 00:06:39.980 we gave it in the prompt. It wasn't perfect, 00:06:40.240 --> 00:06:42.399 but the scores directionally made sense. Gotcha. 00:06:42.579 --> 00:06:45.089 Okay, that makes sense. So once you had that 00:06:45.089 --> 00:06:47.629 reliable scoring, what about steps three and 00:06:47.629 --> 00:06:50.430 four, building the portfolio tracker app? Was 00:06:50.430 --> 00:06:53.860 that coding? Nope, no coding required. That part 00:06:53.860 --> 00:06:55.920 was mostly about visualizing the data and being 00:06:55.920 --> 00:06:58.139 able to filter it easily. We used simple prompts 00:06:58.139 --> 00:07:00.220 and a no -code tool to build a basic web app. 00:07:00.319 --> 00:07:02.540 It took maybe, I don't know, 45 minutes of tweaking 00:07:02.540 --> 00:07:04.300 with just text commands to get the interface 00:07:04.300 --> 00:07:06.300 working how we wanted. OK, so you've got the 00:07:06.300 --> 00:07:08.120 score database. You've got the app to filter 00:07:08.120 --> 00:07:11.000 it. What was the final cutoff? What did you filter 00:07:11.000 --> 00:07:13.759 for to get the highest conviction ideas? We went 00:07:13.759 --> 00:07:16.120 for the absolute strongest candidates based on 00:07:16.120 --> 00:07:18.759 the hypothesis. So network effects score between 00:07:18.759 --> 00:07:20.980 8 and 10, and founder -led score between 8 and 00:07:20.980 --> 00:07:24.019 10. really stringent, and only six companies 00:07:24.019 --> 00:07:26.420 actually met both of those high bars. Only six 00:07:26.420 --> 00:07:28.660 out of the original 26. Out of the whole S &P 00:07:28.660 --> 00:07:31.519 500, yeah. Well, out of the 26 initially identified, 00:07:31.680 --> 00:07:33.680 then scored against the whole 500s of a very 00:07:33.680 --> 00:07:35.879 narrow filter. Right, and the result of focusing 00:07:35.879 --> 00:07:38.480 just on that top tier, that highest quality slice? 00:07:38.720 --> 00:07:41.759 That specific, highly selective portfolio. It 00:07:41.759 --> 00:07:45.660 returned 22 .4 % over the back test period, compared 00:07:45.660 --> 00:07:48.920 to, remember, 9 % for the S &P 500 overall. The 00:07:48.920 --> 00:07:51.459 key takeaway wasn't just use AI. It was using 00:07:51.459 --> 00:07:54.699 AI to apply this deep nuance, this scoring, to 00:07:54.699 --> 00:07:58.120 a simple, clear, testable idea. That NEFL example 00:07:58.120 --> 00:08:00.420 really shows the power of thinking systematically. 00:08:01.360 --> 00:08:03.540 Okay, let's pivot now. Let's talk about how listeners 00:08:03.540 --> 00:08:05.839 can apply this kind of thinking. Let's share 00:08:05.839 --> 00:08:07.819 some of those deeper prompts you mentioned, the 00:08:07.819 --> 00:08:10.439 ones for customizing your own research. Absolutely. 00:08:10.500 --> 00:08:13.060 Because the real value here isn't just copying 00:08:13.060 --> 00:08:15.639 the NFL strategy. It's about defining your own 00:08:15.639 --> 00:08:19.050 investment goals and then using AI. to handle 00:08:19.050 --> 00:08:22.290 the complex filtering and constraints. So yeah, 00:08:22.290 --> 00:08:25.110 we've got three pretty powerful multi -layered 00:08:25.110 --> 00:08:27.470 prompts as examples. OK, let's start with prompt 00:08:27.470 --> 00:08:29.829 one. This one's focused on value investing, right? 00:08:30.129 --> 00:08:32.730 Finding cheap cash machines, looking for those 00:08:32.730 --> 00:08:34.750 hidden gems. Exactly. We want companies that 00:08:34.750 --> 00:08:37.370 are just spewing cash. So we need to target the 00:08:37.370 --> 00:08:40.750 top 20 % based on free cash flow yield. And for 00:08:40.750 --> 00:08:42.850 anyone listening, free cash flow yield is basically 00:08:42.850 --> 00:08:45.190 just how much cash the business generates compared 00:08:45.190 --> 00:08:47.570 to its stock price. It's a raw measure of cash 00:08:47.570 --> 00:08:49.990 efficiency. But just being a cash machine isn't 00:08:49.990 --> 00:08:52.610 enough. There's a crucial constraint. The market 00:08:52.610 --> 00:08:55.350 has to be undervaluing them still. Yes. That's 00:08:55.350 --> 00:08:58.029 key. So we add the condition. Their price to 00:08:58.029 --> 00:09:00.929 earnings ratio, the PE, has to be 15 % lower 00:09:00.929 --> 00:09:04.309 than their industry peers. The AI needs to check 00:09:04.309 --> 00:09:06.970 both things at once. Companies like Qualcomm 00:09:06.970 --> 00:09:09.269 showed up in those results sometimes. It works 00:09:09.269 --> 00:09:11.230 because you're finding that sweet spot. High 00:09:11.230 --> 00:09:13.669 cash generation and a relatively low valuation. 00:09:13.990 --> 00:09:17.519 OK, cool. Next, prompt three. Looking for signals 00:09:17.519 --> 00:09:20.320 from insiders. This is about watching what the 00:09:20.320 --> 00:09:22.779 execs are doing with our own money. Right. The 00:09:22.779 --> 00:09:25.820 prompt tells the AI to scan recent insider training 00:09:25.820 --> 00:09:29.480 data, say, the last 90 days. But we're only interested 00:09:29.480 --> 00:09:32.200 in large open market purchases. That means the 00:09:32.200 --> 00:09:34.259 executive actively decided to go out and buy 00:09:34.259 --> 00:09:36.509 stock on the market. not just getting options 00:09:36.509 --> 00:09:38.730 or grants. And the amount matters? Yeah, we set 00:09:38.730 --> 00:09:41.409 a threshold, like over $100 ,000, and focused 00:09:41.409 --> 00:09:45.090 on key roles, CEO, CFO, the top people. We want 00:09:45.090 --> 00:09:47.450 significant buys by people who know the business 00:09:47.450 --> 00:09:50.649 best. We tell the AI to ignore planned sales 00:09:50.649 --> 00:09:53.269 or stock grants. OK, and there's another critical 00:09:53.269 --> 00:09:55.950 filter here. You only want to see this buying 00:09:55.950 --> 00:09:57.889 if the stock price has actually gone down over 00:09:57.889 --> 00:10:00.350 the last six months. Why is that combination 00:10:00.350 --> 00:10:04.190 so potent? Because think about it. The boss, 00:10:04.470 --> 00:10:06.570 the person who arguably knows the company's real 00:10:06.570 --> 00:10:09.110 future prospects better than anyone, is putting 00:10:09.110 --> 00:10:12.509 down serious cash to buy more stock precisely 00:10:12.509 --> 00:10:14.330 when the market sentiment is negative and the 00:10:14.330 --> 00:10:16.850 price has fallen. It's often a strong signal 00:10:16.850 --> 00:10:18.889 that they believe the current low price is a 00:10:18.889 --> 00:10:21.049 mistake, that there's value the market isn't 00:10:21.049 --> 00:10:23.629 seeing. Interesting. Okay, finally, prompt 10. 00:10:24.090 --> 00:10:27.529 Finding boring but safe income. This one's tailored 00:10:27.529 --> 00:10:29.769 for dividend investors looking for reliability. 00:10:30.090 --> 00:10:32.049 Yeah, this is all about safety and consistency. 00:10:32.669 --> 00:10:35.389 We focus to search on typically defensive sectors, 00:10:35.830 --> 00:10:38.830 utilities, consumer staples, industrials. Then 00:10:38.830 --> 00:10:41.649 we layer the conditions. They need a solid dividend 00:10:41.649 --> 00:10:45.269 yield, say, above 3 .5 percent. Okay. And crucially, 00:10:45.429 --> 00:10:48.649 a safe payout ratio. below 70%. And the payout 00:10:48.649 --> 00:10:50.929 ratio, just quickly, that's the percentage of 00:10:50.929 --> 00:10:53.250 the company's profits that it uses to pay the 00:10:53.250 --> 00:10:55.769 dividends. So lower means more cushion, more 00:10:55.769 --> 00:10:58.590 safety. Exactly. A lower payout ratio means they 00:10:58.590 --> 00:11:00.950 can likely sustain and even grow the dividend, 00:11:01.370 --> 00:11:04.649 even if earnings dip a bit. And then the final 00:11:04.649 --> 00:11:07.620 safety net. We demand a history. they must have 00:11:07.620 --> 00:11:09.460 increased their dividend for at least five straight 00:11:09.460 --> 00:11:12.399 years. Companies like Duke Energy or Coca -Cola 00:11:12.399 --> 00:11:14.820 often pop up here. It's a very disciplined filter 00:11:14.820 --> 00:11:16.860 designed to find steady companies that reliably 00:11:16.860 --> 00:11:19.559 pay and grow their dividends without taking excessive 00:11:19.559 --> 00:11:22.279 risks. You know, that kind of systematic approach, 00:11:22.500 --> 00:11:25.179 looking for specific data points, it also helps 00:11:25.179 --> 00:11:28.620 find these weird market moments, these anomalies, 00:11:28.740 --> 00:11:30.440 like that Coinbase trade example you mentioned. 00:11:30.580 --> 00:11:32.360 Oh yeah, that was a classic case of the hidden 00:11:32.360 --> 00:11:35.820 asset strategy playing out. For a brief window 00:11:35.820 --> 00:11:39.639 in, I think it was early 2023, Coinbase's total 00:11:39.639 --> 00:11:42.320 market value, its stock price times shares, was 00:11:42.320 --> 00:11:44.500 actually lower than the amount of cash and equivalents 00:11:44.500 --> 00:11:46.080 they reported holding on their balance sheet. 00:11:46.700 --> 00:11:48.299 Wait, how does that even happen? How does the 00:11:48.299 --> 00:11:51.100 market make such a seemingly obvious mistake? 00:11:51.259 --> 00:11:54.159 Pure fear, basically. It was the depths of the 00:11:54.159 --> 00:11:56.580 crypto winter, right? Investors were panicking 00:11:56.580 --> 00:11:59.179 about everything crypto related, and that fear 00:11:59.179 --> 00:12:01.320 just overwhelmed the basic balance sheet math 00:12:01.320 --> 00:12:03.759 for a while. But systematic research, the kind 00:12:03.759 --> 00:12:05.919 that specifically flagged situations where cash 00:12:05.919 --> 00:12:08.440 holdings exceed market cap, would have instantly 00:12:08.440 --> 00:12:11.120 highlighted that. It screamed limited downside 00:12:11.120 --> 00:12:14.059 risk with potentially huge upside if sentiment 00:12:14.059 --> 00:12:16.940 just returned to normal. So if we kind of zoom 00:12:16.940 --> 00:12:19.580 out and tie this all together, this AI -driven 00:12:19.580 --> 00:12:22.399 approach, it's systematic, it's active, it's 00:12:22.399 --> 00:12:24.940 definitely not passive index investing, and it's 00:12:24.940 --> 00:12:27.500 not random guesswork either. The real power is 00:12:27.500 --> 00:12:29.980 in how you structure your research process. Right. 00:12:30.019 --> 00:12:32.980 It starts with having a clear, testable idea. 00:12:33.149 --> 00:12:36.269 your own conviction. Maybe it's NEFL, maybe it's 00:12:36.269 --> 00:12:38.529 something else. Like, I believe companies investing 00:12:38.529 --> 00:12:41.269 heavily in R &D during downturns will win. Whatever 00:12:41.269 --> 00:12:43.809 it is, make it specific. Then use the AI tools, 00:12:44.309 --> 00:12:46.330 like stacking Lego blocks of data, you know? 00:12:46.570 --> 00:12:48.509 Let the AI do the heavy lifting of gathering 00:12:48.509 --> 00:12:50.909 the info quickly based on your rules. Don't get 00:12:50.909 --> 00:12:54.070 bogged down in manual data entry. And crucially, 00:12:54.350 --> 00:12:57.720 go beyond just yes -no. Use those scoring systems, 00:12:57.720 --> 00:13:00.500 like 1 to 10, to capture the nuance. That detail 00:13:00.500 --> 00:13:02.779 is often where the real edge lies. And finally, 00:13:02.919 --> 00:13:06.220 test your ideas. Use historical data to see if 00:13:06.220 --> 00:13:08.840 your hypothesis actually held water in the past. 00:13:09.240 --> 00:13:11.539 It doesn't guarantee future results, but it builds 00:13:11.539 --> 00:13:13.759 confidence. And it's important to remember, the 00:13:13.759 --> 00:13:16.539 goal here isn't some magic formula, some get 00:13:16.539 --> 00:13:18.940 -rich -quick secret that works forever. It can't 00:13:18.940 --> 00:13:21.679 be. The real goal is making slightly better, 00:13:21.779 --> 00:13:25.019 more informed decisions consistently over a long 00:13:25.019 --> 00:13:27.450 period of time. That's how genuine wealth is 00:13:27.450 --> 00:13:29.580 typically built. Absolutely. And of course, we 00:13:29.580 --> 00:13:32.899 have to say this deep dive is purely for educational 00:13:32.899 --> 00:13:35.539 purposes. It is definitely not financial advice. 00:13:35.860 --> 00:13:38.720 And past performance, as we saw even with NEFL, 00:13:39.179 --> 00:13:41.539 never guarantees future results. Yeah. And it's 00:13:41.539 --> 00:13:43.460 worth stressing, if you do use a quality filter 00:13:43.460 --> 00:13:46.100 like NEFL, that strategy inherently requires 00:13:46.100 --> 00:13:48.200 patience. You likely need a long time horizon, 00:13:48.840 --> 00:13:51.100 think 10 years or more, for those kinds of quality 00:13:51.100 --> 00:13:53.440 factors to really play out. And always, always 00:13:53.440 --> 00:13:55.580 manage your risk and make sure your overall portfolio 00:13:55.580 --> 00:13:58.220 is diversified. Don't bet the farm on any single 00:13:58.220 --> 00:14:01.440 strategy. For sure. But now, you've got the tools. 00:14:01.539 --> 00:14:03.519 You've seen the framework. You have some powerful 00:14:03.519 --> 00:14:06.299 prompt examples. We really hope you feel encouraged 00:14:06.299 --> 00:14:09.279 to open up your favorite AI tool right now and 00:14:09.279 --> 00:14:11.440 just test one of these strategies. Start building 00:14:11.440 --> 00:14:14.000 your own systematic approach today. Thank you 00:14:14.000 --> 00:14:15.679 for taking this deep dive with us.