WEBVTT 00:00:00.000 --> 00:00:02.940 Have you ever typed a question into an AI tool, 00:00:03.399 --> 00:00:06.139 hoping for something brilliant, only to get back 00:00:06.139 --> 00:00:09.140 a response that felt, well, a bit dull, maybe 00:00:09.140 --> 00:00:11.960 generic? It's a common frustration. And if you 00:00:11.960 --> 00:00:14.789 felt it, you are definitely not alone. Welcome 00:00:14.789 --> 00:00:17.230 to the deep dive. Today, yeah, we're really going 00:00:17.230 --> 00:00:19.929 to unpack why those AI conversations so often 00:00:19.929 --> 00:00:23.149 fall flat. This deep dive into mastering AI, 00:00:23.449 --> 00:00:25.429 prompt engineering for better results, is basically 00:00:25.429 --> 00:00:28.429 going to show us that the AI itself, it usually 00:00:28.429 --> 00:00:30.489 isn't the problem. It's almost entirely about 00:00:30.489 --> 00:00:32.890 how we ask. So we'll tackle the five biggest 00:00:32.890 --> 00:00:34.649 mistakes most people make, then we'll reveal 00:00:34.649 --> 00:00:37.920 five... really game -changing techniques to get 00:00:37.920 --> 00:00:40.299 exactly what you want from AI. We'll even dive 00:00:40.299 --> 00:00:42.320 into some real -world examples, see how they 00:00:42.320 --> 00:00:44.299 work, and learn how to kind of course correct 00:00:44.299 --> 00:00:46.340 when a prompt goes wrong. You should think of 00:00:46.340 --> 00:00:48.939 AI as an incredibly smart assistant, yeah, but 00:00:48.939 --> 00:00:51.079 one that needs truly clear, structured guidance 00:00:51.079 --> 00:00:52.899 from you. It's an interesting dance, isn't it? 00:00:53.119 --> 00:00:56.640 We sort of expect magic from tools like ChatGPT, 00:00:56.719 --> 00:00:58.259 and then when we get something unhelpful, it's 00:00:58.259 --> 00:01:00.770 easy to blame the tool. But the core insight, 00:01:01.009 --> 00:01:03.649 I think, what truly resonates here is that AI 00:01:03.649 --> 00:01:06.670 isn't a mind reader. It just can't be. It creates 00:01:06.670 --> 00:01:08.810 based on patterns on these vast amounts of data 00:01:08.810 --> 00:01:12.349 it's been trained on. It doesn't... Intuit your 00:01:12.349 --> 00:01:15.549 unexpressed thoughts. Exactly. If you throw a 00:01:15.549 --> 00:01:18.269 vague request its way, you're almost guaranteed 00:01:18.269 --> 00:01:20.310 to get a vague answer back. It's kind of garbage 00:01:20.310 --> 00:01:23.209 in, garbage out. But when you switch to clear, 00:01:23.650 --> 00:01:25.930 detailed, and really structured instructions, 00:01:26.530 --> 00:01:28.989 that's when the surprising, sometimes even brilliant 00:01:28.989 --> 00:01:31.730 results start to emerge. So this whole deep dive 00:01:31.730 --> 00:01:34.150 is designed as your shortcut, really, to mastering 00:01:34.150 --> 00:01:36.930 that fundamental communication. So, okay, given 00:01:36.930 --> 00:01:39.829 that... How do we begin to bridge that gap? How 00:01:39.829 --> 00:01:42.549 can we make sure AI truly understands our complex 00:01:42.549 --> 00:01:44.510 human thoughts or intentions? I think it really 00:01:44.510 --> 00:01:46.629 comes down to being incredibly deliberate. We 00:01:46.629 --> 00:01:48.890 need to give it structured, detailed instructions. 00:01:49.230 --> 00:01:52.469 Treat it like a smart, but very literal assistant. 00:01:52.730 --> 00:01:54.650 Okay, right. Let's unpack this further then. 00:01:54.650 --> 00:01:57.109 Before we even get to the cool techniques, the 00:01:57.109 --> 00:01:59.829 powerful stuff, we need to talk about five things 00:01:59.829 --> 00:02:01.950 most people are doing wrong. These are the most 00:02:01.950 --> 00:02:05.069 common errors that just, well, tank your AI results. 00:02:05.189 --> 00:02:08.490 Okay, mistake number one. writing vague or just 00:02:08.490 --> 00:02:11.370 very short prompts. This is like going to a top 00:02:11.370 --> 00:02:13.770 chef, handing them maybe two random ingredients 00:02:13.770 --> 00:02:16.469 and expecting a gourmet meal. It just won't happen. 00:02:16.770 --> 00:02:18.810 The AI simply has too little to work with. So 00:02:18.810 --> 00:02:21.090 it leads to outputs that are often random or 00:02:21.090 --> 00:02:24.280 just... Not helpful at all. Right a classic bad 00:02:24.280 --> 00:02:26.879 example would be just typing write some advertising 00:02:26.879 --> 00:02:28.780 slogans. What do you get? You'll often get cliches 00:02:28.780 --> 00:02:31.400 like connecting passions or maybe innovation 00:02:31.400 --> 00:02:34.000 for tomorrow, you know stuff you could find anywhere 00:02:34.000 --> 00:02:36.199 totally generic But consider the difference with 00:02:36.199 --> 00:02:38.599 a good example something like write five slogans 00:02:38.599 --> 00:02:41.430 for tin to brew coffee Okay, good start. Target 00:02:41.430 --> 00:02:44.629 25 to 35 year old office workers. Focus on flavor 00:02:44.629 --> 00:02:47.610 and concentration. Use a modern short tone. See 00:02:47.610 --> 00:02:49.550 the detail. Then you get something like tinty 00:02:49.550 --> 00:02:53.189 brew, rich coffee, bright ideas. Ooh, nice. Or 00:02:53.189 --> 00:02:56.710 maybe fuel your focus. Tinty brew. Much, much 00:02:56.710 --> 00:02:59.129 better, right? So the fix is basically always, 00:02:59.229 --> 00:03:01.889 always add details. Think about the industry, 00:03:02.069 --> 00:03:04.110 your target audience, the purpose of the output, 00:03:04.310 --> 00:03:06.629 any length limits, and what you want the final 00:03:06.629 --> 00:03:09.419 look and feel to be. Mistake number two. And 00:03:09.419 --> 00:03:11.340 this is a big one for many of us, myself included 00:03:11.340 --> 00:03:14.300 sometimes, treating AI like a Google search engine. 00:03:14.860 --> 00:03:16.900 We're just so deeply conditioned by years of 00:03:16.900 --> 00:03:20.280 using search. But AI doesn't search the web in 00:03:20.280 --> 00:03:22.780 real time like Google does. It's a generative 00:03:22.780 --> 00:03:25.479 tool. It creates based on its training data. 00:03:25.960 --> 00:03:28.060 So the fundamental shift in thinking is it's 00:03:28.060 --> 00:03:30.389 a creation machine. Not a search machine. That's 00:03:30.389 --> 00:03:33.389 such a crucial distinction. So if you ask famous 00:03:33.389 --> 00:03:35.569 tourist spots in Delat, yeah, you'll get a boring 00:03:35.569 --> 00:03:37.449 bulleted list. It's like reading a textbook entry. 00:03:37.469 --> 00:03:40.050 There's no smark, no depth. Exactly. But what 00:03:40.050 --> 00:03:44.250 if instead you say act as a local expert on Delat? 00:03:44.449 --> 00:03:47.009 OK, setting a role, write a three day, two night 00:03:47.009 --> 00:03:49.490 travel plan for a couple, mixing famous spots 00:03:49.490 --> 00:03:52.090 with less known coffee shops. Getting specific. 00:03:52.569 --> 00:03:55.669 Focus on romantic, non crowded activities presented 00:03:55.669 --> 00:03:58.120 in a table. Now, that gives a direction. What 00:03:58.120 --> 00:04:00.580 you get back is an entirely different experience. 00:04:00.639 --> 00:04:02.879 Well, that's beautiful. Yeah. The result is a 00:04:02.879 --> 00:04:05.360 detailed personal itinerary. Suddenly, you're 00:04:05.360 --> 00:04:08.240 getting suggestions like, sunrise, drive up Langvang 00:04:08.240 --> 00:04:11.219 Mountain, maybe, or a quiet dinner at a nearby 00:04:11.219 --> 00:04:14.039 bistro overlooking the city. See, it feels curated. 00:04:14.539 --> 00:04:17.579 So the fix is, ask the AI to create a specific 00:04:17.579 --> 00:04:20.100 product, an article, a plan, a piece of code, 00:04:20.399 --> 00:04:22.759 and, critically, give it a roll. and tell it 00:04:22.759 --> 00:04:24.560 how to present that information. OK, mistake 00:04:24.560 --> 00:04:28.279 number three, adding too much fluff and unnecessary 00:04:28.279 --> 00:04:30.660 politeness. This one surprises a lot of people, 00:04:30.660 --> 00:04:32.720 but it's surprisingly common. We find ourselves 00:04:32.720 --> 00:04:34.560 saying please and thank you so much. We're just 00:04:34.560 --> 00:04:37.000 polite people. But the AI, well, it doesn't have 00:04:37.000 --> 00:04:39.220 feelings. And these extra words can actually 00:04:39.220 --> 00:04:41.259 clutter your instructions, making them less clear, 00:04:41.579 --> 00:04:43.720 maybe even less efficient. Totally. You're wasting 00:04:43.720 --> 00:04:47.259 valuable prompt real estate. A bad example. Hello, 00:04:47.300 --> 00:04:49.519 could you please do me a favor? If you can, please 00:04:49.519 --> 00:04:51.100 summarize this article for me. Thank you very 00:04:51.100 --> 00:04:54.490 much. Right. The AI will still summarize. Of 00:04:54.490 --> 00:04:56.410 course it will, but you've added this inefficient 00:04:56.410 --> 00:04:59.250 layer of communication. A good example is simply, 00:04:59.850 --> 00:05:01.730 summarize the following article into three main 00:05:01.730 --> 00:05:04.269 bullet points. Each bullet point, no more than 00:05:04.269 --> 00:05:08.490 20 words. Clear, concise, direct. Exactly. The 00:05:08.490 --> 00:05:12.550 fix. Be direct. AI thrives on clear, unambiguous 00:05:12.550 --> 00:05:15.029 commands. Get straight to the point. Mistake 00:05:15.029 --> 00:05:18.009 number four. Trying to do absolutely everything 00:05:18.009 --> 00:05:21.259 in one giant prompt. This is a trap I still wrestle 00:05:21.259 --> 00:05:23.480 with myself sometimes, I have to admit. It's 00:05:23.480 --> 00:05:25.600 like trying to solve a huge complex puzzle all 00:05:25.600 --> 00:05:28.120 in one go. The AI often gets confused, it might 00:05:28.120 --> 00:05:30.740 miss important details, or the output just drifts. 00:05:31.319 --> 00:05:33.120 I still find myself getting prompt drift when 00:05:33.120 --> 00:05:35.339 I try to cram too much in. Yeah, you're basically 00:05:35.339 --> 00:05:37.980 trying to give it a whole novel to write. Like, 00:05:38.079 --> 00:05:40.930 asking for, um... A complete marketing plan for 00:05:40.930 --> 00:05:43.110 a new sunscreen, including competitor analysis, 00:05:43.449 --> 00:05:46.009 social media, paid ads, email scripts, and influencer 00:05:46.009 --> 00:05:49.009 ideas. Oh, that's not just a lot. It's an entire 00:05:49.009 --> 00:05:51.889 project. It rarely works well in a single shot. 00:05:52.350 --> 00:05:54.689 The fix here is pretty straightforward. Break 00:05:54.689 --> 00:05:57.490 down complex tasks into small, step -by -step 00:05:57.490 --> 00:06:02.089 prompts. And finally, mistake number five. Accepting 00:06:02.089 --> 00:06:04.689 the very first answer you get, I'd argue this 00:06:04.689 --> 00:06:06.870 might be the biggest missed opportunity of all. 00:06:06.990 --> 00:06:09.790 So many people treat that initial response as 00:06:09.790 --> 00:06:11.990 final, even if it's not quite right. They just 00:06:11.990 --> 00:06:14.370 walk away, completely missing the real power 00:06:14.370 --> 00:06:17.589 of AI, which is iteration. Exactly. The true 00:06:17.589 --> 00:06:20.329 power of AI lies in its ability to iterate and 00:06:20.329 --> 00:06:22.769 improve based on your feedback. Think of that 00:06:22.769 --> 00:06:25.029 first answer as a first draft, nothing more. 00:06:25.529 --> 00:06:27.990 Your job, then, is to be its editor. Give it 00:06:27.990 --> 00:06:29.990 direction. So you provide feedback. You might 00:06:29.990 --> 00:06:32.589 say, make the tone funnier, or maybe add three 00:06:32.589 --> 00:06:35.759 specific actionable examples. The AI isn't just 00:06:35.759 --> 00:06:37.819 a generator. It's a highly capable editor, too, 00:06:38.040 --> 00:06:40.079 if you guide it. OK, so if we pull back for a 00:06:40.079 --> 00:06:43.800 second, it sounds like AI isn't just a fact finder, 00:06:44.079 --> 00:06:46.000 definitely not a mind reader, or some kind of 00:06:46.000 --> 00:06:47.699 one -shot genie, right? No, definitely not a 00:06:47.699 --> 00:06:49.819 genie. That's a good way to put it. It's fundamentally 00:06:49.819 --> 00:06:53.100 a creation tool. And what I'm learning, what 00:06:53.100 --> 00:06:56.699 we're seeing, is it needs that precise. iterative 00:06:56.699 --> 00:06:59.319 guidance from us. It's almost like we're constantly 00:06:59.319 --> 00:07:02.139 refining its understanding, step by step, rather 00:07:02.139 --> 00:07:04.439 than just hitting a button and expecting perfect 00:07:04.439 --> 00:07:07.300 magic. Okay, now that we've cleared those common 00:07:07.300 --> 00:07:09.779 hurdles, those mistakes, let's really dive into 00:07:09.779 --> 00:07:12.120 the techniques that will truly elevate you, turn 00:07:12.120 --> 00:07:15.740 you from an average user into a skilled AI collaborator. 00:07:16.139 --> 00:07:18.220 Alright, technique number one, first principles 00:07:18.220 --> 00:07:20.920 thinking. Now, this isn't about just copying 00:07:20.920 --> 00:07:23.300 prompts you find online somewhere, no. This is 00:07:23.300 --> 00:07:25.879 about building your own perfect prompt from scratch. 00:07:26.319 --> 00:07:28.279 Think of it as stripping away all the assumptions 00:07:28.279 --> 00:07:30.139 to get to the fundamental building blocks of 00:07:30.139 --> 00:07:32.779 what you truly need. It's a concept borrowed 00:07:32.779 --> 00:07:34.980 from physics, actually. You break down your request 00:07:34.980 --> 00:07:37.920 into its absolute atoms. It requires a moment 00:07:37.920 --> 00:07:40.060 of deep reflection first, doesn't it? Yeah. You 00:07:40.060 --> 00:07:42.279 have to really ask yourself, What exactly do 00:07:42.279 --> 00:07:44.560 I want the AI to produce? Who is the audience 00:07:44.560 --> 00:07:47.420 for this? What's the desired format? What background 00:07:47.420 --> 00:07:50.120 information does the AI absolutely need to know 00:07:50.120 --> 00:07:53.079 to succeed? And crucially, what are the specific 00:07:53.079 --> 00:07:55.699 constraints or boundaries? Right. So for a job 00:07:55.699 --> 00:07:58.240 description, don't just type write a job description. 00:07:58.300 --> 00:08:00.639 That's way too vague. Apply first principles. 00:08:01.100 --> 00:08:03.819 You build one specifically for, say, a content 00:08:03.819 --> 00:08:07.199 marketing specialist at an ed tech company requiring 00:08:07.199 --> 00:08:10.259 SEO expertise, maybe targeting university students. 00:08:10.620 --> 00:08:13.639 in a very specific, maybe encouraging tone. That's 00:08:13.639 --> 00:08:15.439 first principles thinking in action. Technique 00:08:15.439 --> 00:08:18.139 number two, the five boxes method. This is such 00:08:18.139 --> 00:08:20.879 a simple yet incredibly powerful framework. It 00:08:20.879 --> 00:08:22.879 really works. You just fill in these five elements, 00:08:23.439 --> 00:08:26.040 role, task, context, constraints, and output 00:08:26.040 --> 00:08:28.480 format. Get these five things down and your prompt 00:08:28.480 --> 00:08:31.339 becomes nearly perfect almost instantly. It brings 00:08:31.339 --> 00:08:33.480 so much clarity. Let's run a quick example, make 00:08:33.480 --> 00:08:36.480 it concrete. Imagine you're an experienced hiring 00:08:36.480 --> 00:08:39.639 manager. OK, that's your role. Your task is to 00:08:39.639 --> 00:08:42.019 draft an email to reject a candidate named John 00:08:42.019 --> 00:08:45.840 Smith. The context is important. John has good 00:08:45.840 --> 00:08:48.220 skills, but maybe lacks practical experience 00:08:48.220 --> 00:08:49.799 for this role. And you want to stay in touch 00:08:49.799 --> 00:08:52.500 for future possibilities, constraints. It needs 00:08:52.500 --> 00:08:55.220 to be short, professional, but also empathetic. 00:08:55.360 --> 00:08:57.799 You don't want to burn bridges. And the output 00:08:57.799 --> 00:09:00.690 format. Clearly structured with maybe a greeting, 00:09:00.870 --> 00:09:03.269 the result, brief feedback, and then well wishes. 00:09:03.509 --> 00:09:06.110 Boom, five boxes filled. Technique number three, 00:09:06.429 --> 00:09:09.190 prompt chaining. Okay, for truly complex projects, 00:09:09.529 --> 00:09:11.590 this is your secret weapon. Think of it like 00:09:11.590 --> 00:09:14.070 stacking Lego blocks of data or information. 00:09:14.549 --> 00:09:17.309 You use multiple smaller focused prompts, with 00:09:17.309 --> 00:09:19.090 each one building directly on the information 00:09:19.090 --> 00:09:20.970 and the output from the prompt just before it. 00:09:21.070 --> 00:09:23.789 Yeah, this really helps the AI maintain focus 00:09:23.789 --> 00:09:26.429 and depth on each individual step. It prevents 00:09:26.429 --> 00:09:28.929 that drift we talked about. So imagine planning 00:09:28.929 --> 00:09:31.669 a five -day solo trip to Hanoi. Your first prompt 00:09:31.669 --> 00:09:34.929 might be, give me 10 must -do activities in Hanoi 00:09:34.929 --> 00:09:37.809 for a solo traveler. Then your second prompt, 00:09:37.970 --> 00:09:40.789 building on that, OK, create a detailed five 00:09:40.789 --> 00:09:42.629 -day schedule based on those activities, grouping 00:09:42.629 --> 00:09:45.990 them geographically. Third, now suggest some 00:09:45.990 --> 00:09:48.090 specific street foods to try near the locations 00:09:48.090 --> 00:09:51.809 for each day. getting tasty fourth generate a 00:09:51.809 --> 00:09:55.350 practical packing list for hanoi in say october 00:09:55.350 --> 00:09:58.929 and maybe finally fifth Write a short, inspiring 00:09:58.929 --> 00:10:01.110 paragraph to get me excited about my trip based 00:10:01.110 --> 00:10:03.570 on this itinerary. See how that step -by -step 00:10:03.570 --> 00:10:06.470 chained approach yields far richer, more detailed, 00:10:06.629 --> 00:10:08.809 and ultimately much more useful results? It's 00:10:08.809 --> 00:10:11.190 so much better than one huge overwhelming request. 00:10:11.570 --> 00:10:13.570 Okay, technique number four, meta -prompting. 00:10:13.690 --> 00:10:15.149 Now here's where it gets really interesting and 00:10:15.149 --> 00:10:17.929 feels a little bit like Inception, maybe? This 00:10:17.929 --> 00:10:20.809 technique uses the AI itself as your prompt writing 00:10:20.809 --> 00:10:23.159 coach. It's almost like having a prompt architect 00:10:23.159 --> 00:10:24.679 right there in your chat window helping you build 00:10:24.679 --> 00:10:27.840 the prompt. Yes. Metaprompting means you literally 00:10:27.840 --> 00:10:31.000 ask the AI to help you create the best possible 00:10:31.000 --> 00:10:33.440 prompt for your task. You might ask it something 00:10:33.440 --> 00:10:35.779 like, what information do you need from me to 00:10:35.779 --> 00:10:37.960 help me create an effective learning plan to 00:10:37.960 --> 00:10:40.360 reach a specific language proficiency goal, say 00:10:40.360 --> 00:10:43.539 a C1 level in English, within the next six months? 00:10:43.720 --> 00:10:46.659 The AI will then likely ask you clarifying questions, 00:10:46.700 --> 00:10:48.960 like about your current level, how much time 00:10:48.960 --> 00:10:50.580 you can study, what your weaknesses are, stuff 00:10:50.580 --> 00:10:53.009 like that. Once you've answered its questions, 00:10:53.250 --> 00:10:55.929 you then ask it to draft the optimal prompt for 00:10:55.929 --> 00:10:59.350 you to use. It's a fantastic way to ensure clarity 00:10:59.350 --> 00:11:02.370 and completeness. And that leads perfectly into 00:11:02.370 --> 00:11:04.809 technique number five, intelligent workflow. 00:11:05.250 --> 00:11:07.289 This is where you combine these powerful tools. 00:11:07.330 --> 00:11:09.429 You use meta prompting first to get an overall 00:11:09.429 --> 00:11:11.929 strategic roadmap, that big picture. And then 00:11:11.929 --> 00:11:14.450 you use prompt chaining to execute each individual 00:11:14.450 --> 00:11:17.049 step of that roadmap methodically. Right. So 00:11:17.049 --> 00:11:19.669 you could start with a broad meta prompt like. 00:11:19.960 --> 00:11:22.779 Act as a strategic marketing expert. Outline 00:11:22.779 --> 00:11:24.960 a comprehensive roadmap for launching a new online 00:11:24.960 --> 00:11:27.159 course covering everything from initial messaging 00:11:27.159 --> 00:11:30.600 to post -launch content strategy. The AI will 00:11:30.600 --> 00:11:33.500 then give you, probably, a numbered list of strategic 00:11:33.500 --> 00:11:36.600 steps, like a project plan. And then you execute 00:11:36.600 --> 00:11:39.809 each step with specific chained prompts. So you'd 00:11:39.809 --> 00:11:42.389 say, OK, let's start with step one. Identifying 00:11:42.389 --> 00:11:45.409 the target audience. Help me create three detailed 00:11:45.409 --> 00:11:47.529 potential customer personas for this course, 00:11:47.990 --> 00:11:50.269 including their primary pain points and aspirations. 00:11:50.509 --> 00:11:52.889 And you work through the list. Whoa. OK, just 00:11:52.889 --> 00:11:55.289 imagine scaling this kind of intelligent workflow 00:11:55.289 --> 00:11:59.429 for something huge, like a massive content strategy 00:11:59.429 --> 00:12:01.909 for an entire global organization. The implications 00:12:01.909 --> 00:12:04.950 here are enormous. The possibilities are truly 00:12:04.950 --> 00:12:08.370 huge. So to kind of synthesize these techniques, 00:12:08.990 --> 00:12:11.450 They're fundamentally about breaking down complexity, 00:12:11.669 --> 00:12:14.350 right? And engaging in a much deeper collaboration 00:12:14.350 --> 00:12:16.370 with the AI, wouldn't you say? Absolutely. It 00:12:16.370 --> 00:12:18.769 shifts the entire dynamic. You move from simply 00:12:18.769 --> 00:12:21.250 commanding a tool to strategizing with an intelligent 00:12:21.250 --> 00:12:23.950 partner. And interestingly, that often forces 00:12:23.950 --> 00:12:26.049 us to clarify our own thinking in the process, 00:12:26.149 --> 00:12:27.730 doesn't it? You can't be vague with your partner. 00:12:27.950 --> 00:12:29.769 OK, let's see. These powerful techniques really 00:12:29.769 --> 00:12:32.169 shine now with two practical examples, things 00:12:32.169 --> 00:12:35.649 you can try, like right now. First example, drafting 00:12:35.649 --> 00:12:37.950 a LinkedIn post for personal branding. We'll 00:12:37.950 --> 00:12:40.830 use the five boxes method here. The bad way, 00:12:40.830 --> 00:12:43.230 as we know, is just write a LinkedIn post about 00:12:43.230 --> 00:12:45.470 continuous learning. What you get is probably 00:12:45.470 --> 00:12:49.710 generic, maybe cliche. Forgettable. The good 00:12:49.710 --> 00:12:52.690 way meticulously uses the five boxes. So, your 00:12:52.690 --> 00:12:54.509 role. Maybe you're a career development expert. 00:12:54.809 --> 00:12:57.070 Your task. Create a LinkedIn post about the importance 00:12:57.070 --> 00:12:59.779 of learning a new skill each quarter. The context. 00:13:00.200 --> 00:13:02.299 Share your personal story about, say, learning 00:13:02.299 --> 00:13:05.700 data analysis recently. Make it real. Constraints. 00:13:05.860 --> 00:13:08.799 Maintain a positive, encouraging tone. Include 00:13:08.799 --> 00:13:11.740 maybe two, three relevant emojis. Ask an open 00:13:11.740 --> 00:13:13.860 -ended question at the end for engagement. And 00:13:13.860 --> 00:13:16.600 use three to four relevant hashtags. Format. 00:13:16.879 --> 00:13:18.799 Structure it with a clear hook, your story, and 00:13:18.799 --> 00:13:20.879 a compelling conclusion or call to action. And 00:13:20.879 --> 00:13:22.720 the result isn't just a post anymore. It's an 00:13:22.720 --> 00:13:25.799 authentic, valuable, and genuinely engaging piece 00:13:25.799 --> 00:13:28.259 of personal branding content. It connects. Okay. 00:13:28.320 --> 00:13:30.740 Second example, creating ideas for a short video 00:13:30.740 --> 00:13:32.679 series on TikTok. And this time we'll use prompt 00:13:32.679 --> 00:13:35.440 chaining the bad way. Just TikTok video ideas 00:13:35.440 --> 00:13:37.580 for a cooking channel. Super general, super vague. 00:13:37.799 --> 00:13:40.059 You'll get obvious stuff. Right. The good way 00:13:40.059 --> 00:13:43.019 starts with prompt one. Okay. I run a TikTok 00:13:43.019 --> 00:13:44.919 channel targeting busy people, think students, 00:13:45.080 --> 00:13:47.419 young professionals. Give me five main concepts 00:13:47.419 --> 00:13:51.500 for video series. The AI might respond with interesting 00:13:51.500 --> 00:13:54.139 ideas like maybe upgrade your instant noodles. 00:13:54.759 --> 00:13:57.919 Okay. That has potential. Then, prompt 2, you 00:13:57.919 --> 00:14:00.759 chain it. OK, I love the upgrade your instant 00:14:00.759 --> 00:14:03.759 noodles concept. Develop this into five specific 00:14:03.759 --> 00:14:06.440 video script ideas. Give me the dish name, any 00:14:06.440 --> 00:14:08.659 special ingredients needed, and a surprising 00:14:08.659 --> 00:14:11.759 twist for each one. And maybe finally, prompt 00:14:11.759 --> 00:14:14.620 3. Let's choose the soy -marinated egg ramen 00:14:14.620 --> 00:14:17.399 idea. Write a detailed 60 -second script for 00:14:17.399 --> 00:14:19.940 that video, including suggested visuals, audio 00:14:19.940 --> 00:14:22.460 cues, and duration notes for each segment. See, 00:14:22.480 --> 00:14:24.379 this isn't just about brainstorming ideas anymore. 00:14:24.740 --> 00:14:27.679 This transforms those vague concepts into a detailed, 00:14:27.899 --> 00:14:30.570 almost production -ready content plan. It's a 00:14:30.570 --> 00:14:32.629 massive leap forward in utility. These examples 00:14:32.629 --> 00:14:34.669 really hammer home how structure makes all the 00:14:34.669 --> 00:14:36.370 difference, doesn't it? It takes those abstract 00:14:36.370 --> 00:14:39.370 ideas and makes them tangible, actionable. It 00:14:39.370 --> 00:14:41.610 absolutely does. Structure really is the bridge. 00:14:41.970 --> 00:14:45.009 It transforms those vague intentions into actionable, 00:14:45.529 --> 00:14:47.669 high -quality content that actually serve your 00:14:47.669 --> 00:14:49.809 purpose. Now look, even with all these amazing 00:14:49.809 --> 00:14:51.970 techniques, sometimes things still don't work 00:14:51.970 --> 00:14:54.570 out perfectly. It happens. Don't panic. Here's 00:14:54.570 --> 00:14:56.429 how to troubleshoot your prompts when they go 00:14:56.429 --> 00:14:59.840 a little off -kilter. Step one. Just reread exactly 00:14:59.840 --> 00:15:02.200 what you wrote. This might sound obvious, but 00:15:02.200 --> 00:15:05.039 honestly, it's incredibly crucial. Look carefully 00:15:05.039 --> 00:15:07.419 for missing details. Any vague words like it, 00:15:07.720 --> 00:15:11.379 or they, or even simple typos. Often, a tiny 00:15:11.379 --> 00:15:14.240 wording change can fix a surprising amount. Step 00:15:14.240 --> 00:15:16.980 two. Adjust your constraints or the requested 00:15:16.980 --> 00:15:19.700 format. If the outfit is way too long or too 00:15:19.700 --> 00:15:22.080 short, or the tone is wrong, maybe too formal, 00:15:22.279 --> 00:15:24.759 too casual, just tell it. You can simply say, 00:15:24.779 --> 00:15:26.639 rewrite the answer above, but limit it to five 00:15:26.639 --> 00:15:28.759 sentences. Or maybe, could you adjust the tone 00:15:28.759 --> 00:15:32.460 to be funny and sarcastic? Step three. Provide 00:15:32.460 --> 00:15:34.360 an example. This is a really powerful technique 00:15:34.360 --> 00:15:37.179 called few -shot prompting. It's called FewShot 00:15:37.179 --> 00:15:40.139 because you're giving the AI just a few examples, 00:15:40.220 --> 00:15:42.580 maybe even one, of the desired output format. 00:15:42.779 --> 00:15:45.340 You're teaching it on the fly. So if the AI isn't 00:15:45.340 --> 00:15:47.379 quite following the structure you want, show 00:15:47.379 --> 00:15:49.940 it. Give it an example, like, please list the 00:15:49.940 --> 00:15:52.019 ideas as bullet points using the following structure. 00:15:52.360 --> 00:15:55.519 Idea, name, name, short description, one sentence 00:15:55.519 --> 00:15:59.019 description, target audience, who it's for. And 00:15:59.019 --> 00:16:01.970 step four. Sometimes you just need to try a different 00:16:01.970 --> 00:16:04.649 AI model, seriously. Different AI models like 00:16:04.649 --> 00:16:07.529 ChachiPT, Gemini, Claude, they all have different 00:16:07.529 --> 00:16:09.389 strengths, different knowledge cutoffs, sometimes 00:16:09.389 --> 00:16:12.070 even subtly different personalities. So copy 00:16:12.070 --> 00:16:13.789 the exact same prompt you're using into another 00:16:13.789 --> 00:16:15.769 model. You might be genuinely surprised by the 00:16:15.769 --> 00:16:17.610 difference in the results. So when we're troubleshooting, 00:16:17.750 --> 00:16:19.350 it often seems like it's more about refining 00:16:19.350 --> 00:16:21.309 our own communication, our own instructions, 00:16:21.789 --> 00:16:24.629 rather than the AI fundamentally failing, right? 00:16:25.029 --> 00:16:28.129 Precisely. It really boils down to becoming a 00:16:28.129 --> 00:16:30.730 better, more thoughtful editor of your own instructions. 00:16:31.289 --> 00:16:33.950 That iterative process, that refinement, that's 00:16:33.950 --> 00:16:36.370 where the magic really happens. For those of 00:16:36.370 --> 00:16:38.450 you who want to level up even further, really 00:16:38.450 --> 00:16:41.070 become power users. These advanced tips build 00:16:41.070 --> 00:16:43.470 on everything we've covered and can truly transform 00:16:43.470 --> 00:16:46.570 your workflow. First up, the quick prompt outline 00:16:46.570 --> 00:16:49.250 method. This is a great habit. Before writing 00:16:49.250 --> 00:16:51.470 any prompt, just take 15 seconds. Seriously, 00:16:51.669 --> 00:16:53.549 that's it. Quickly answer these four questions 00:16:53.549 --> 00:16:55.960 in your head. One, what's the final result I'm 00:16:55.960 --> 00:16:58.159 actually looking for? Two, who's the audience 00:16:58.159 --> 00:17:01.019 for this? Three, what's the specific format and 00:17:01.019 --> 00:17:04.119 tone needed? And four, what's the absolutely 00:17:04.119 --> 00:17:07.119 essential context the AI needs to know? Building 00:17:07.119 --> 00:17:09.619 this simple habit will help you create high quality 00:17:09.619 --> 00:17:12.480 prompts much, much faster and with far greater 00:17:12.480 --> 00:17:15.440 consistency. It's a small mental investment with 00:17:15.440 --> 00:17:18.420 really huge returns down the line. Next, really 00:17:18.420 --> 00:17:20.759 cultivate the iteration mindset. We touched on 00:17:20.759 --> 00:17:24.259 this, but it's key. Remember, Even the experts 00:17:24.259 --> 00:17:27.140 rarely get absolutely perfect results in the 00:17:27.140 --> 00:17:29.079 very first try. It just doesn't happen often. 00:17:29.599 --> 00:17:32.059 So always, always treat that first response you 00:17:32.059 --> 00:17:34.480 get as nothing more than a draft. a starting 00:17:34.480 --> 00:17:38.000 point. Then iterate. Give a clear, specific feedback. 00:17:38.319 --> 00:17:40.740 Make this section more specific by adding relevant 00:17:40.740 --> 00:17:43.319 statistics. Or could you adjust the tone here 00:17:43.319 --> 00:17:45.339 to be more friendly and welcoming for beginners? 00:17:45.700 --> 00:17:48.019 Maybe add a paragraph that addresses potential 00:17:48.019 --> 00:17:50.720 common objections. Keep refining. Keep pushing 00:17:50.720 --> 00:17:53.980 the AI for better, more nuanced output. And finally, 00:17:54.039 --> 00:17:56.319 a really practical tip. Build your own prompt 00:17:56.319 --> 00:17:59.299 library. Seriously, do this. When you craft good, 00:17:59.339 --> 00:18:01.200 effective prompts that really work well for a 00:18:01.200 --> 00:18:03.440 task you do often, save them. Doesn't matter 00:18:03.440 --> 00:18:05.660 where Notion, Evernote, Google Keep, a simple 00:18:05.660 --> 00:18:07.819 text file, just create a repository for your 00:18:07.819 --> 00:18:10.059 successful prompts. Yeah, build templates for 00:18:10.059 --> 00:18:12.819 your common tasks. Maybe writing certain types 00:18:12.819 --> 00:18:15.700 of emails, generating content ideas, problem 00:18:15.700 --> 00:18:18.740 solving frameworks, planning outlines. This simple 00:18:18.740 --> 00:18:21.079 act of saving and reusing successful prompts 00:18:21.079 --> 00:18:23.359 will save you an incredible amount of time and 00:18:23.359 --> 00:18:26.029 mental effort in the long run. It's compound 00:18:26.029 --> 00:18:29.109 efficiency. So these advanced tips, they're really 00:18:29.109 --> 00:18:31.869 about building good habits, right? And establishing 00:18:31.869 --> 00:18:34.349 systems for getting consistent, high -quality 00:18:34.349 --> 00:18:37.069 results more easily. Yes, exactly. They empower 00:18:37.069 --> 00:18:39.549 you to work smarter, not just harder, and more 00:18:39.549 --> 00:18:41.869 efficiently, and ultimately with greater impact 00:18:41.869 --> 00:18:44.390 when you're collaborating with AI tools. Mid 00:18:44.390 --> 00:18:46.690 -roll sponsor read. Today's deep dive really 00:18:46.690 --> 00:18:49.049 underscores something important. Mastering AI. 00:18:49.289 --> 00:18:52.190 It isn't about some secret technical wizardry 00:18:52.190 --> 00:18:54.710 or knowing complex algorithms. It's fundamentally 00:18:54.710 --> 00:18:57.509 about clear, effective communication. It's about 00:18:57.509 --> 00:19:00.089 how we interact with the tool. Yeah, we've gone 00:19:00.089 --> 00:19:02.250 from dissecting those five critical mistakes 00:19:02.250 --> 00:19:04.789 that just waste your time and energy to arming 00:19:04.789 --> 00:19:07.789 you with some really powerful techniques. The 00:19:07.789 --> 00:19:10.190 five boxes method for that essential structure, 00:19:10.730 --> 00:19:13.230 prompt chaining for breaking down those big complex 00:19:13.230 --> 00:19:15.769 projects, meta prompting for when you need an 00:19:15.769 --> 00:19:17.609 AI -powered coach to help you craft the prompt 00:19:17.609 --> 00:19:20.670 itself, and the intelligent workflow for tackling 00:19:20.670 --> 00:19:23.289 your most challenging multi -step tasks in a 00:19:23.289 --> 00:19:25.559 strategic way. Ultimately, what we're seeing 00:19:25.559 --> 00:19:28.519 is that AI rewards clear thinking and better 00:19:28.519 --> 00:19:30.940 questions, not just sheer effort or banging your 00:19:30.940 --> 00:19:32.980 head against the wall. The more specific, the 00:19:32.980 --> 00:19:35.240 more structured your prompts are, the more amazing 00:19:35.240 --> 00:19:37.359 and tailored your results will inevitably become. 00:19:37.900 --> 00:19:40.240 So where do you start with all this? We challenge 00:19:40.240 --> 00:19:43.069 you. Pick just one technique from this deep dive, 00:19:43.230 --> 00:19:45.829 just one to begin with right now. Pick a task 00:19:45.829 --> 00:19:47.970 you do pretty regularly. Maybe it's drafting 00:19:47.970 --> 00:19:50.130 a certain kind of email, writing a social media 00:19:50.130 --> 00:19:53.250 post, or outlining a simple plan. And try applying 00:19:53.250 --> 00:19:55.670 the five boxes method to it next time. I think 00:19:55.670 --> 00:19:58.710 you will be genuinely amazed at how much better, 00:19:58.950 --> 00:20:01.170 how much more aligned with your actual needs 00:20:01.170 --> 00:20:05.059 your AI outputs become almost immediately. The 00:20:05.059 --> 00:20:08.579 key, as always, is to start simple, get a win, 00:20:09.000 --> 00:20:11.420 and then steadily build those skills over time. 00:20:11.740 --> 00:20:14.140 The future really does belong to people who can 00:20:14.140 --> 00:20:16.039 communicate effectively with AI. It's becoming 00:20:16.039 --> 00:20:18.440 a fundamental skill. And with these kinds of 00:20:18.440 --> 00:20:20.180 techniques in your toolkit, you're not just using 00:20:20.180 --> 00:20:22.960 technology, you are truly starting to master 00:20:22.960 --> 00:20:24.960 it. So here's something to ponder as we wrap 00:20:24.960 --> 00:20:27.630 up. How could applying even a single one of these 00:20:27.630 --> 00:20:30.369 techniques transform just one of your daily tasks, 00:20:30.930 --> 00:20:33.710 making it not just faster perhaps, but genuinely 00:20:33.710 --> 00:20:36.329 better, more impactful? That's it for this deep 00:20:36.329 --> 00:20:38.049 dive. Thanks so much for joining us today. Until 00:20:38.049 --> 00:20:38.450 next time.