WEBVTT 00:00:00.000 --> 00:00:03.600 OK, so for anyone who's already using AI tools 00:00:03.600 --> 00:00:08.500 like ChatGPT, Claw, Gemini, they can do amazing 00:00:08.500 --> 00:00:11.960 things. But be honest, how often do the results 00:00:11.960 --> 00:00:15.880 really feel expert level? Right. You ask a question, 00:00:15.919 --> 00:00:19.399 you get a response, and it feels, well, maybe 00:00:19.399 --> 00:00:22.940 a bit boring, a bit generic. Exactly. Bland even. 00:00:23.359 --> 00:00:25.339 And look, you're definitely not alone if you 00:00:25.339 --> 00:00:27.000 feel that way. Most people are getting pretty 00:00:27.000 --> 00:00:29.019 average results. And the thing is, it's usually 00:00:29.019 --> 00:00:31.399 not the AI's fault. It really comes down to the 00:00:31.399 --> 00:00:33.460 input. The input, yeah. Think about it like this. 00:00:33.820 --> 00:00:38.240 Imagine an AI is a master chef. Unbelievable 00:00:38.240 --> 00:00:41.140 skills. But if you only give them basic ingredients, 00:00:41.780 --> 00:00:43.500 like just stuff from the back of the cupboard. 00:00:43.579 --> 00:00:46.079 They can only do so much. Precisely. They'll 00:00:46.079 --> 00:00:48.259 make something edible, sure, but probably forgettable. 00:00:48.460 --> 00:00:50.859 Now what if you gave that same chef premium, 00:00:51.179 --> 00:00:53.399 carefully chosen, specific ingredients? Then 00:00:53.399 --> 00:00:55.579 they can create something incredible. A masterpiece. 00:00:55.820 --> 00:00:58.520 That's the principle with AI. The quality, the 00:00:58.520 --> 00:01:00.320 distinctiveness of the output. It's directly 00:01:00.320 --> 00:01:02.399 tied to the quality of the information you feed 00:01:02.399 --> 00:01:04.900 it. And that's the big misconception we really 00:01:04.900 --> 00:01:06.579 want to tackle in this deep dive. People think, 00:01:06.599 --> 00:01:09.500 oh, I need the newest AI model or I need to learn 00:01:09.500 --> 00:01:12.700 some super complex prompting technique. But the 00:01:12.700 --> 00:01:16.989 truth we're exploring today is... Well, maybe 00:01:16.989 --> 00:01:19.689 a bit counterintuitive. It's less about the chef 00:01:19.689 --> 00:01:21.750 and more about the quality of the ingredients 00:01:21.750 --> 00:01:24.349 you bring to the kitchen. Well said. So we're 00:01:24.349 --> 00:01:27.670 going to unpack this really effective 15 -minute 00:01:27.670 --> 00:01:30.530 workflow. It's designed to seriously upgrade 00:01:30.530 --> 00:01:33.549 your AI interactions. It's not about fancy tech. 00:01:33.950 --> 00:01:37.489 It's about feeding the AI high -quality research 00:01:37.489 --> 00:01:40.430 -backed info so the results genuinely stand out. 00:01:40.709 --> 00:01:42.930 OK, so let's dig into why those results often 00:01:42.930 --> 00:01:46.040 feel generic in the first place. The root cause 00:01:46.040 --> 00:01:49.120 is actually quite simple. These AI models, they 00:01:49.120 --> 00:01:51.200 learn from a huge, huge amount of internet data. 00:01:51.280 --> 00:01:53.200 So when you ask a simple kind of open -ended 00:01:53.200 --> 00:01:55.480 question. It gives you the average answer. Exactly. 00:01:55.540 --> 00:01:57.700 It defaults to the average answer from the internet. 00:01:57.939 --> 00:02:00.540 It's designed to be broadly helpful, but broad 00:02:00.540 --> 00:02:04.480 often translates to vague, forgettable, generic. 00:02:04.760 --> 00:02:06.680 That makes total sense, like asking for a good 00:02:06.680 --> 00:02:09.060 movie and getting something popular. Technically 00:02:09.060 --> 00:02:12.460 right, but not helpful. Right. So when you say 00:02:12.460 --> 00:02:15.639 average answers, is that because the AI is limited 00:02:15.639 --> 00:02:18.199 or is it more about how we're asking? Oh, it's 00:02:18.199 --> 00:02:20.300 much more about how we're asking. The AI isn't 00:02:20.300 --> 00:02:23.139 the limitation here. It's just trying its best 00:02:23.139 --> 00:02:25.960 with what it's given. The real limitation comes 00:02:25.960 --> 00:02:29.280 when we don't provide that specific nuanced context 00:02:29.280 --> 00:02:32.400 it needs to really shine. The shift in approach 00:02:32.400 --> 00:02:34.599 is key. Absolutely critical. Instead of just 00:02:34.599 --> 00:02:37.699 a basic question, imagine first equipping the 00:02:37.699 --> 00:02:41.030 AI with, say, carefully selected academic research, 00:02:41.330 --> 00:02:43.490 expert insights, proven strategies. Then it can 00:02:43.490 --> 00:02:45.849 go beyond the generics. Way beyond. It can become 00:02:45.849 --> 00:02:48.189 genuinely authoritative. And that's the core 00:02:48.189 --> 00:02:51.830 idea we're unlocking today. The secret isn't 00:02:51.830 --> 00:02:55.110 necessarily a better AI tool or waiting for version 00:02:55.110 --> 00:02:57.990 5 .0. It's about getting better at preparing 00:02:57.990 --> 00:03:00.729 the information. Preparing the ingredients. Exactly. 00:03:00.810 --> 00:03:04.500 When you give an AI that specific credible research 00:03:04.500 --> 00:03:07.039 back context, the output can sound like it came 00:03:07.039 --> 00:03:09.680 from a real subject matter expert, not just a 00:03:09.680 --> 00:03:12.199 quick web search. Absolutely. And the how is 00:03:12.199 --> 00:03:13.819 what we've really focused on. We've boiled it 00:03:13.819 --> 00:03:17.080 down to a pretty efficient three tool strategy. 00:03:17.340 --> 00:03:19.599 Free tools. Yeah, three completely free tools. 00:03:19.719 --> 00:03:21.740 And when you use them in sequence, they work 00:03:21.740 --> 00:03:24.879 together like a well -oiled machine for this 00:03:24.879 --> 00:03:27.939 workflow. OK. Intrigued. Let's bring down these 00:03:27.939 --> 00:03:31.419 tools. All right. First up, Google Scholar. That's 00:03:31.419 --> 00:03:33.939 our source for finding credible academic research. 00:03:34.080 --> 00:03:36.800 Got it. Academic backbone. Then there's Notebook 00:03:36.800 --> 00:03:40.000 LM. This tool is fantastic for organizing and, 00:03:40.000 --> 00:03:42.060 importantly, analyzing that research. We can 00:03:42.060 --> 00:03:44.699 analyze the research. And the third. And finally, 00:03:44.919 --> 00:03:47.520 Gemini Advanced. We use this for actually generating 00:03:47.520 --> 00:03:50.639 those expert -level outputs based on everything 00:03:50.639 --> 00:03:53.360 we've gathered. Google Scholar, Notebook LM, 00:03:53.520 --> 00:03:55.819 Gemini Advanced. Each one has a very specific 00:03:55.819 --> 00:03:58.620 job, and they really complement each other. together 00:03:58.620 --> 00:04:00.840 in this way and the results, they're genuinely 00:04:00.840 --> 00:04:04.479 impressive. It's a system for moving beyond those 00:04:04.479 --> 00:04:07.039 average AI interactions. Okay, let's dive into 00:04:07.039 --> 00:04:10.419 step one then. Mastering Google Scholar. You 00:04:10.419 --> 00:04:13.340 mentioned a 3 -2 -1 method because Google Scholar 00:04:13.530 --> 00:04:16.310 I mean, it's this incredible treasure chest of 00:04:16.310 --> 00:04:19.589 credible info, right? Peer -reviewed stuff. But 00:04:19.589 --> 00:04:22.029 honestly, most people probably don't know how 00:04:22.029 --> 00:04:25.209 to really use it effectively for practical tasks. 00:04:25.470 --> 00:04:27.589 That's very true. It can feel a bit academic, 00:04:27.730 --> 00:04:31.170 maybe intimidating. Yeah. So this 3 -2 -1 method, 00:04:31.389 --> 00:04:34.149 it's something refined over years in research 00:04:34.149 --> 00:04:36.449 settings. It's probably the most efficient way 00:04:36.449 --> 00:04:38.490 I've found to get the highest quality sources 00:04:38.490 --> 00:04:40.709 on a topic in about five minutes. Five minutes. 00:04:40.829 --> 00:04:43.319 OK, I'm listening. What's the method? It's elegant 00:04:43.319 --> 00:04:45.899 because it's simple but powerful. First, the 00:04:45.899 --> 00:04:47.899 three stands for finding the three most cited 00:04:47.899 --> 00:04:50.699 papers on your topic. Most cited. Right. Go to 00:04:50.699 --> 00:04:52.699 Google Scholar. Type in your main topic. Let's 00:04:52.699 --> 00:04:55.319 stick with that example. Effective online learning 00:04:55.319 --> 00:04:59.040 strategies. Scholar automatically ranks by relevance 00:04:59.040 --> 00:05:02.240 and citation count, those top results. They're 00:05:02.240 --> 00:05:04.579 usually the foundational most respected papers. 00:05:05.079 --> 00:05:07.279 Your job is just to download the top three that 00:05:07.279 --> 00:05:10.040 have a direct PDF link available. And why focus 00:05:10.040 --> 00:05:12.930 on most cited specifically? What does that tell 00:05:12.930 --> 00:05:16.410 us? Think of citations as votes of confidence 00:05:16.410 --> 00:05:19.189 from other researchers. Ah, like peer validation. 00:05:19.470 --> 00:05:22.430 Exactly. If hundreds, maybe thousands of other 00:05:22.430 --> 00:05:24.810 experts have referenced a paper in their own 00:05:24.810 --> 00:05:27.670 published work, that's a really strong signal. 00:05:27.930 --> 00:05:30.230 That means the research is solid, it stood up 00:05:30.230 --> 00:05:32.910 to scrutiny, and it's likely had a significant 00:05:32.910 --> 00:05:35.829 impact. Makes sense. And you mentioned a tip 00:05:35.829 --> 00:05:39.139 for finding PDFs. Oh, yeah. Quick pro tip. If 00:05:39.139 --> 00:05:41.779 you're not seeing many direct PDF links, just 00:05:41.779 --> 00:05:44.759 add file type dot PDF to your search query. File 00:05:44.759 --> 00:05:47.360 type dot PDF. Yep. That tells Google Scholar 00:05:47.360 --> 00:05:50.160 to only show results where a PDF is directly 00:05:50.160 --> 00:05:52.079 downloadable. Saves a lot of clicking around. 00:05:52.360 --> 00:05:54.120 Smart. OK, so that's the three. What's the two? 00:05:54.259 --> 00:05:57.079 The two is for getting the two most recent papers. 00:05:57.519 --> 00:06:00.620 This is crucial. It balances those classic, highly 00:06:00.620 --> 00:06:03.279 cited papers with the absolute latest thinking. 00:06:03.459 --> 00:06:05.959 Right, because fields evolve. Constantly. So 00:06:05.959 --> 00:06:07.639 on the left side of the Google Scholar results 00:06:07.639 --> 00:06:10.199 page, you'll see a filter for time. Click anytime 00:06:10.199 --> 00:06:12.779 and change it to show papers published in, say, 00:06:12.839 --> 00:06:15.879 the last two or three years. Download two solid 00:06:15.879 --> 00:06:18.160 -looking recent papers. This gives you the cutting 00:06:18.160 --> 00:06:20.800 edge, the newest discoveries, and sometimes you'll 00:06:20.800 --> 00:06:23.199 see overlap. A paper might be both highly cited 00:06:23.199 --> 00:06:26.120 and very recent. Is that good? That's a fantastic 00:06:26.120 --> 00:06:27.879 sign. It means you've likely found something 00:06:27.879 --> 00:06:31.540 that's both influential and super current. Gold 00:06:31.540 --> 00:06:36.180 dust. Nice. OK, three most cited, two most recent. 00:06:36.240 --> 00:06:38.699 What's the one? The one is, I think, the secret 00:06:38.699 --> 00:06:42.860 weapon here. Find one review paper. A review 00:06:42.860 --> 00:06:45.720 paper? Yeah. Just add the word review to your 00:06:45.720 --> 00:06:48.319 original search query. So effective online learning 00:06:48.319 --> 00:06:51.579 strategies review. OK. Why are view papers so 00:06:51.579 --> 00:06:53.779 valuable? Ah, they're gold mines. Seriously. 00:06:54.400 --> 00:06:56.779 Review papers are usually written by established 00:06:56.779 --> 00:06:59.199 researchers who have spent months, maybe even 00:06:59.199 --> 00:07:02.259 years, reading and analyzing hundreds of individual 00:07:02.259 --> 00:07:04.899 studies on a specific topic. Wow. They synthesize 00:07:04.899 --> 00:07:07.060 all the key findings. They identify the major 00:07:07.060 --> 00:07:09.300 trends, point out where there are still gaps 00:07:09.300 --> 00:07:12.699 in the research. It's all summarized in one comprehensive 00:07:12.699 --> 00:07:14.860 document. So it's like having a research assistant 00:07:14.860 --> 00:07:17.439 do a massive literature review for you. Exactly 00:07:17.439 --> 00:07:20.060 that. It saves an incredible amount of time and 00:07:20.060 --> 00:07:22.339 gives you a high -level overview you just can't 00:07:22.339 --> 00:07:25.699 get otherwise. So the 321 method combined. It 00:07:25.699 --> 00:07:27.899 gives you this really well -rounded perspective. 00:07:28.100 --> 00:07:30.480 You get the historical foundation from the most 00:07:30.480 --> 00:07:32.959 cited. You get the cutting edge insights from 00:07:32.959 --> 00:07:35.959 the recent papers. And you get that integrated 00:07:35.959 --> 00:07:38.980 holistic overview from the review paper. It ensures 00:07:38.980 --> 00:07:41.459 you're not missing critical angles. Precisely. 00:07:41.839 --> 00:07:44.139 And importantly, you're building your knowledge 00:07:44.139 --> 00:07:46.899 on sources that actual experts in the field trust 00:07:46.899 --> 00:07:50.300 and use themselves. It's robust. It's efficient 00:07:50.300 --> 00:07:53.100 and it sets you up perfectly for the next step. 00:07:53.459 --> 00:07:56.339 Okay, great. So we've used the 321 method and 00:07:56.339 --> 00:07:58.959 we've got our six high quality PDFs from Google 00:07:58.959 --> 00:08:02.680 Scholar. Powerful first step. But now comes the 00:08:02.680 --> 00:08:05.040 maybe slightly daunting part. You're looking 00:08:05.040 --> 00:08:08.339 at these dense academic papers. The wall of text. 00:08:08.420 --> 00:08:10.209 Yeah. How do you actually make sense of it all 00:08:10.209 --> 00:08:12.029 efficiently? This feels like where overwhelm 00:08:12.029 --> 00:08:14.290 could set in. And that's exactly where our second 00:08:14.290 --> 00:08:17.410 tool, Notebook LM, comes into play. It's designed 00:08:17.410 --> 00:08:20.009 to turn that potential overwhelm into clear, 00:08:20.310 --> 00:08:22.529 actionable insights. Notebook LM, that's the 00:08:22.529 --> 00:08:24.509 Google AI research assistant you mentioned. That's 00:08:24.509 --> 00:08:27.129 the one. And its key strength is that it's specifically 00:08:27.129 --> 00:08:29.629 designed to work with your uploaded documents. 00:08:29.810 --> 00:08:32.389 This is crucial. Why is that so important? Because 00:08:32.389 --> 00:08:35.590 unlike general AI chatbots that pull from the 00:08:35.590 --> 00:08:38.080 whole internet and might... you know, occasionally 00:08:38.080 --> 00:08:39.980 hallucinate or make things up. Right, we've all 00:08:39.980 --> 00:08:42.899 seen that happen. Notebook LM stays rigorously 00:08:42.899 --> 00:08:45.139 grounded in the source material you provide. 00:08:45.340 --> 00:08:48.519 It bases its answers and analysis directly on 00:08:48.519 --> 00:08:51.240 the content of those PDFs you uploaded. Okay, 00:08:51.340 --> 00:08:53.740 that's a big deal. It keeps it focused and factual 00:08:53.740 --> 00:08:56.779 based on my research. Exactly. So setting it 00:08:56.779 --> 00:08:59.000 up is actually really easy. You go to Notebook 00:08:59.000 --> 00:09:02.000 LM, create a new notebook, give it a clear, specific 00:09:02.000 --> 00:09:04.899 name related to your topic like online learning 00:09:04.899 --> 00:09:07.259 strategies research. That helps the AI understand 00:09:07.259 --> 00:09:10.159 the focus. Yep, gives it context. Then you just 00:09:10.159 --> 00:09:12.259 upload those six PDFs you downloaded from Google 00:09:12.259 --> 00:09:15.330 Scholar. drag and drop, or use the upload button. 00:09:15.370 --> 00:09:17.330 And it just digests them? It processes them, 00:09:17.529 --> 00:09:19.789 yeah. Reads them, understands the content. Usually 00:09:19.789 --> 00:09:21.450 only takes a minute or two, depending on the 00:09:21.450 --> 00:09:25.230 size. Okay, PDFs uploaded. Now what? You mentioned 00:09:25.230 --> 00:09:27.669 something about using the discover function strategically. 00:09:28.110 --> 00:09:29.870 My first thought would be, okay, find me more 00:09:29.870 --> 00:09:32.690 papers. That's the common instinct. And Discover 00:09:32.690 --> 00:09:35.169 can find more academic papers, but the results 00:09:35.169 --> 00:09:38.210 can be a bit mixed sometimes. For this specific 00:09:38.210 --> 00:09:40.370 workflow, we use it a bit differently. How so? 00:09:40.549 --> 00:09:42.929 Instead of asking for more papers, ask it to 00:09:42.929 --> 00:09:45.570 find related practical content, like YouTube 00:09:45.570 --> 00:09:47.899 videos. Ah, interesting. Give me an example. 00:09:48.240 --> 00:09:50.460 Sure. You could ask something like, please find 00:09:50.460 --> 00:09:53.000 YouTube videos offering useful tips on how to 00:09:53.000 --> 00:09:55.500 improve time management skills for online students. 00:09:55.659 --> 00:09:58.240 I see. So you're deliberately bridging the gap 00:09:58.240 --> 00:10:01.879 between the deep academic research in your PDFs 00:10:01.879 --> 00:10:04.700 and more practical, maybe real world applications 00:10:04.700 --> 00:10:07.960 or advice. Precisely. You get that solid academic 00:10:07.960 --> 00:10:10.600 foundation from the papers and then you layer 00:10:10.600 --> 00:10:13.360 on practical tips or demonstrations from, say, 00:10:13.559 --> 00:10:16.000 video content. It creates a much more holistic 00:10:16.000 --> 00:10:18.980 understanding theory meets practice. That's really 00:10:18.980 --> 00:10:21.379 clever. Okay, so we've got our sources loaded, 00:10:21.480 --> 00:10:23.580 maybe added some practical videos. What next 00:10:23.580 --> 00:10:26.960 in Notebook LM? Generate a mind map. Notebook 00:10:26.960 --> 00:10:29.600 LM can create a visual overview of your research 00:10:29.600 --> 00:10:32.220 sources. A mind map, how's that helpful? It's 00:10:32.220 --> 00:10:34.799 incredibly useful for quickly seeing the main 00:10:34.799 --> 00:10:37.159 themes and subtopics that emerge across all your 00:10:37.159 --> 00:10:39.240 documents. You can see how concepts connect, 00:10:39.379 --> 00:10:41.240 which ideas are prominent. Like a bird's -eye 00:10:41.240 --> 00:10:44.659 view. Exactly, and you can click to expand different 00:10:44.659 --> 00:10:46.500 branches of the map to quickly get a summary 00:10:46.500 --> 00:10:48.960 of the information related to that specific theme 00:10:48.960 --> 00:10:51.220 within your sources. It helps you orient yourself 00:10:51.220 --> 00:10:54.759 really fast. Okay, mind map done. Now the crucial 00:10:54.759 --> 00:10:59.059 part, actually extracting the killer insights. 00:10:59.419 --> 00:11:01.539 I know you have two specific questions you recommend 00:11:01.539 --> 00:11:04.419 asking Notebook LM. Yes, these two questions 00:11:04.419 --> 00:11:06.559 consistently deliver high -value insights. They 00:11:06.559 --> 00:11:08.860 cut through the noise. What are they? First question. 00:11:09.019 --> 00:11:12.279 Following the 80 -20 rule, list the 10 most impactful 00:11:12.279 --> 00:11:14.860 aspects related to your topic. The 80 -20 rule. 00:11:15.100 --> 00:11:17.120 Yeah. So focusing on a vital few things that 00:11:17.120 --> 00:11:19.399 make the biggest difference. Exactly. It forces 00:11:19.399 --> 00:11:22.480 the AI to identify what truly matters most according 00:11:22.480 --> 00:11:24.700 to the research you provide. OK, that's powerful. 00:11:25.000 --> 00:11:27.480 And the second question. This one often uncovers 00:11:27.480 --> 00:11:30.899 the real gems, the surprising stuff. Ask, what 00:11:30.899 --> 00:11:33.159 are the 10 most counterintuitive elements about 00:11:33.159 --> 00:11:36.610 your topic? Counterintuitive. Things that go 00:11:36.610 --> 00:11:38.850 against common assumptions. Precisely. These 00:11:38.850 --> 00:11:40.610 surprising findings, the things that make you 00:11:40.610 --> 00:11:44.049 go, huh, really, huh, often become the most interesting, 00:11:44.470 --> 00:11:47.110 memorable, and valuable parts of whatever content 00:11:47.110 --> 00:11:49.730 you create later. They make your output stand 00:11:49.730 --> 00:11:52.289 out. Find the most impactful. Find the most surprising. 00:11:53.129 --> 00:11:55.230 Got it. And you mentioned something important 00:11:55.230 --> 00:11:58.250 about saving these insights. Yes. Crucial point. 00:11:58.769 --> 00:12:00.809 Notebook LM doesn't currently save your chat 00:12:00.809 --> 00:12:03.730 history between sessions. So when it generates 00:12:03.730 --> 00:12:06.690 those impactful or counterintuitive lists or 00:12:06.690 --> 00:12:08.850 any other insight you find valuable, copy and 00:12:08.850 --> 00:12:11.070 paste them somewhere immediately. Right away. 00:12:11.190 --> 00:12:13.570 Put them in the built -in notes section within 00:12:13.570 --> 00:12:15.690 Notebook LM itself or in a separate document. 00:12:15.789 --> 00:12:17.529 Just make sure you capture them before you close 00:12:17.529 --> 00:12:19.879 the notebook. Good tip. Don't want to lose that 00:12:19.879 --> 00:12:24.279 gold. Okay, so we've gathered research, analyzed 00:12:24.279 --> 00:12:27.259 it with New Book LM, extracted the key impactful 00:12:27.259 --> 00:12:29.460 and counterintuitive points. Which brings us 00:12:29.460 --> 00:12:32.049 neatly to step three. This is the moment where 00:12:32.049 --> 00:12:34.990 the magic really happens, creating that truly 00:12:34.990 --> 00:12:38.250 expert level content using Gemini Advanced. Right. 00:12:38.269 --> 00:12:39.970 We've done the prep work, gathered the premium 00:12:39.970 --> 00:12:42.450 ingredients. Exactly. We've done the heavy lifting 00:12:42.450 --> 00:12:44.690 of finding quality research and pulling out the 00:12:44.690 --> 00:12:47.669 core insights. Now, it's time to leverage all 00:12:47.669 --> 00:12:52.389 of that to generate something. The key difference 00:12:52.389 --> 00:12:55.450 here, the real game changer compared to just 00:12:55.450 --> 00:12:58.889 asking a simple question, is how we prompt Gemini, 00:12:58.990 --> 00:13:01.259 right? Absolutely. It's all about the prompt. 00:13:01.860 --> 00:13:04.000 Instead of a basic query, you're going to construct 00:13:04.000 --> 00:13:06.940 a rich, detailed prompt that feeds Gemini all 00:13:06.940 --> 00:13:09.100 the valuable information you just gathered from 00:13:09.100 --> 00:13:11.240 Notebook LM. Okay, walk me through the structure 00:13:11.240 --> 00:13:13.639 of this super prompt. Sure. It generally follows 00:13:13.639 --> 00:13:16.440 a clear structure. First, you define a specific 00:13:16.440 --> 00:13:19.960 role for the AI. Like telling it who to be. Precisely. 00:13:20.100 --> 00:13:22.840 Start with something like, as an expert in your 00:13:22.840 --> 00:13:26.240 specific role, create. So maybe as an expert 00:13:26.240 --> 00:13:28.639 in educational program design. Got it. So that's 00:13:28.639 --> 00:13:31.480 the context. What's next? Then state the specific 00:13:31.480 --> 00:13:33.879 deliverable you want. What exactly should it 00:13:33.879 --> 00:13:36.799 create? For instance, a detailed outline for 00:13:36.799 --> 00:13:39.820 an online course. Roll, then deliverable. Clear. 00:13:40.100 --> 00:13:42.019 Now, here comes the crucial part, feeding it 00:13:42.019 --> 00:13:44.799 the insights. You literally paste in the lists 00:13:44.799 --> 00:13:47.419 you got from Notebook LM. Start with the most 00:13:47.419 --> 00:13:50.259 impactful factors. Just paste that 80 -20 list. 00:13:50.360 --> 00:13:53.289 Yep. Then, paste in the counterintuitive findings 00:13:53.289 --> 00:13:55.789 that list of surprising elements. Okay, impactful 00:13:55.789 --> 00:13:58.110 and counterintuitive points included. You might 00:13:58.110 --> 00:14:00.350 also want to add a few other key research themes 00:14:00.350 --> 00:14:02.309 that you identified, perhaps from looking at 00:14:02.309 --> 00:14:05.090 your mind map in Notebook LM. Just bullet points 00:14:05.090 --> 00:14:08.149 of other important concepts. So, layering in 00:14:08.149 --> 00:14:10.610 more context from the research. Exactly. And 00:14:10.610 --> 00:14:13.149 finally, outline the deliverable requirements. 00:14:13.570 --> 00:14:16.230 Be specific about what the final output should 00:14:16.230 --> 00:14:17.769 look like. What do you mean by requirements? 00:14:18.090 --> 00:14:21.100 Describe the format. the sections, the tone, 00:14:21.559 --> 00:14:24.700 the target audience, for example. The outline 00:14:24.700 --> 00:14:27.840 should include five modules. Each module needs 00:14:27.840 --> 00:14:30.539 learning objectives, key activities, and a mini 00:14:30.539 --> 00:14:33.960 quiz. The tone should be encouraging and accessible 00:14:33.960 --> 00:14:36.559 for beginners. Wow, okay, so you're giving it 00:14:36.559 --> 00:14:39.519 the role, the task, the core insights from your 00:14:39.519 --> 00:14:42.080 research, and detailed instructions on the final 00:14:42.080 --> 00:14:44.320 product. Precisely. You're giving it everything 00:14:44.320 --> 00:14:47.000 it needs to act like an expert using the specific 00:14:47.000 --> 00:14:49.220 high -quality information you provided. And the 00:14:49.220 --> 00:14:51.259 difference in the output. I assume it's noticeable. 00:14:51.470 --> 00:14:54.370 Oh, it's night and day. It's immediately strikingly 00:14:54.370 --> 00:14:56.889 obvious. Instead of generic fluff you could find 00:14:56.889 --> 00:14:59.049 anywhere... It actually uses the research. Yes. 00:14:59.289 --> 00:15:02.149 It will reference specific findings. It will 00:15:02.149 --> 00:15:04.809 incorporate those counterintuitive insights that 00:15:04.809 --> 00:15:07.289 make the content so much more interesting and 00:15:07.289 --> 00:15:10.730 informative. The whole thing feels authoritative, 00:15:11.389 --> 00:15:14.549 deeply knowledgeable. It goes way beyond surface 00:15:14.549 --> 00:15:17.289 -level stuff. It sounds like it came from someone 00:15:17.289 --> 00:15:20.470 who... actually knows the topic deeply because 00:15:20.470 --> 00:15:22.490 you fed it that deep knowledge. It's exactly 00:15:22.490 --> 00:15:24.710 it. Can you give a concrete example, like back 00:15:24.710 --> 00:15:26.789 to the online learning strategies topic? Sure. 00:15:27.129 --> 00:15:29.210 Instead of Gemini just giving you a generic list 00:15:29.210 --> 00:15:32.129 like make a schedule, take breaks, which is okay, 00:15:32.450 --> 00:15:35.409 but basic. Right. Using this detailed prompt 00:15:35.409 --> 00:15:38.629 method, drawing on research about, say, cognitive 00:15:38.629 --> 00:15:41.549 load theory or spaced repetition you found, you 00:15:41.549 --> 00:15:43.610 could ask it to create an interactive course 00:15:43.610 --> 00:15:46.669 outline for developing 21st century online learning 00:15:46.669 --> 00:15:49.450 spills. And it might come back with specific 00:15:49.450 --> 00:15:52.110 modules on metacognition, digital collaboration 00:15:52.110 --> 00:15:55.889 tools, and information literacy, suggesting specific 00:15:55.889 --> 00:15:58.230 activities like peer review assignments based 00:15:58.230 --> 00:16:01.649 on research findings or mini quizzes using principles 00:16:01.649 --> 00:16:05.259 of active recall you highlight. Wow. That's a 00:16:05.259 --> 00:16:07.620 huge difference. Much more specific, actionable, 00:16:08.039 --> 00:16:09.820 and evidence -based. Completely different level. 00:16:10.179 --> 00:16:12.460 And once you get comfortable with this basic 00:16:12.460 --> 00:16:15.100 text -based workflow, you can really start thinking 00:16:15.100 --> 00:16:18.059 bigger. How so? What are some advanced applications? 00:16:18.379 --> 00:16:20.899 Well, don't just stop at outlines or articles. 00:16:21.320 --> 00:16:24.080 Use those insights to create multimedia content. 00:16:24.519 --> 00:16:27.679 Ask Gemini. Based on these key points and counterintuitive 00:16:27.679 --> 00:16:31.039 findings, outline a 10 -slide presentation. Oh, 00:16:31.080 --> 00:16:33.639 interesting. Like for a work presentation. Exactly. 00:16:34.059 --> 00:16:36.370 Or... Draft a script for a five -minute YouTube 00:16:36.370 --> 00:16:39.129 video explaining these surprising research findings 00:16:39.129 --> 00:16:42.330 about your topic. Or even design the layout concept 00:16:42.330 --> 00:16:44.769 for an infographic summarizing the most effective 00:16:44.769 --> 00:16:47.389 strategies discussed in this research. So repurposing 00:16:47.389 --> 00:16:50.919 the core insights into different formats. Think 00:16:50.919 --> 00:16:53.100 about business strategy, too. If your research 00:16:53.100 --> 00:16:56.519 was on, say, the effectiveness of different customer 00:16:56.519 --> 00:16:59.860 loyalty programs, you could use this method to 00:16:59.860 --> 00:17:03.159 generate a detailed plan for a new loyalty program, 00:17:03.440 --> 00:17:07.000 specifying reward types, tier structures, KPIs 00:17:07.000 --> 00:17:09.819 to track, all grounded in the academic evidence 00:17:09.819 --> 00:17:13.000 you gathered, not just industry fads or guesswork. 00:17:13.210 --> 00:17:15.690 That's incredibly practical for business applications. 00:17:15.829 --> 00:17:18.029 Definitely. And even for personal development. 00:17:18.269 --> 00:17:20.150 Say you researched mindfulness techniques for 00:17:20.150 --> 00:17:22.789 stress reduction. Okay. You could ask Gemini 00:17:22.789 --> 00:17:26.150 to create a personalized 30 -day mindfulness 00:17:26.150 --> 00:17:28.670 practice program, complete with specific daily 00:17:28.670 --> 00:17:31.630 exercises, maybe even short scientific explanations 00:17:31.630 --> 00:17:34.089 for why each practice works based on the papers 00:17:34.089 --> 00:17:36.569 you fed it. Tailored evidence -based personal 00:17:36.569 --> 00:17:39.210 plans. The possibilities really do seem limitless 00:17:39.210 --> 00:17:41.539 once you nail this preparation workflow. They 00:17:41.539 --> 00:17:43.339 really are. And it brings us back to the core 00:17:43.339 --> 00:17:45.240 message, really, the huge difference between 00:17:45.240 --> 00:17:47.619 getting, you know, kind of average AI results 00:17:47.619 --> 00:17:50.660 and consistently producing extraordinary expert 00:17:50.660 --> 00:17:52.799 level outputs. It isn't about chasing the newest 00:17:52.799 --> 00:17:54.980 software update or learning some secret prompt 00:17:54.980 --> 00:17:57.880 trick. No, it's fundamentally about changing 00:17:57.880 --> 00:18:00.380 your approach. It's about the preparation. That 00:18:00.380 --> 00:18:03.099 15 minutes upfront. That 15 minute investment 00:18:03.099 --> 00:18:06.539 in gathering high quality targeted research before 00:18:06.539 --> 00:18:09.980 you even open Gemini or your AI tool of choice. 00:18:10.420 --> 00:18:12.500 That's what transforms you. You go from someone 00:18:12.500 --> 00:18:15.319 getting generic stuff back to someone who can 00:18:15.319 --> 00:18:18.099 consistently generate content that feels genuinely 00:18:18.099 --> 00:18:20.819 experts. Instantly. The return on that small 00:18:20.819 --> 00:18:23.460 time investment is massive. So the encouragement 00:18:23.460 --> 00:18:26.180 for everyone listening is Try it. Absolutely. 00:18:26.500 --> 00:18:29.059 Pick a topic that genuinely matters to you, something 00:18:29.059 --> 00:18:31.539 for work or a personal interest. Follow this 00:18:31.539 --> 00:18:35.279 workflow exactly. Google Scholar 321, Notebook 00:18:35.279 --> 00:18:38.059 LM for analysis and insights, then the structured 00:18:38.059 --> 00:18:40.339 prompt into Gemini. Just try it once, following 00:18:40.339 --> 00:18:43.079 the steps. And just prepare to be, well, pretty 00:18:43.079 --> 00:18:45.319 amazed by the difference in what comes out. It 00:18:45.319 --> 00:18:48.079 really shifts the paradigm from just asking AI 00:18:48.079 --> 00:18:50.940 questions to collaborating with AI using high 00:18:50.940 --> 00:18:52.680 quality information. That's a great way to put 00:18:52.680 --> 00:18:55.539 it. Maybe the final thought to leave everyone 00:18:55.539 --> 00:18:59.400 with is this. Once you experience what's possible, 00:18:59.779 --> 00:19:01.660 once you see the kind of expert level output 00:19:01.660 --> 00:19:03.839 you can get when you combine quality research 00:19:03.839 --> 00:19:07.079 with these powerful AI tools, how will that change 00:19:07.079 --> 00:19:09.579 things for you? How will this new approach change 00:19:09.579 --> 00:19:11.779 the way you tackle everyday information challenges, 00:19:12.200 --> 00:19:14.720 the way you approach creative projects, or even 00:19:14.720 --> 00:19:16.140 how you build your own expertise?