WEBVTT 00:00:00.000 --> 00:00:03.680 So when you think about AI, really think about 00:00:03.680 --> 00:00:08.060 it. How many of us truly tap into its real power? 00:00:08.199 --> 00:00:10.720 I mean, beyond just, you know, typing a quick 00:00:10.720 --> 00:00:13.480 question into a chat bot. Are we really using 00:00:13.480 --> 00:00:15.179 it like the powerful instrument it's meant to 00:00:15.179 --> 00:00:18.010 be? Or just... scratching the surface. It's kind 00:00:18.010 --> 00:00:19.649 of like most people get handed this incredible 00:00:19.649 --> 00:00:21.609 Stradivarius violin, right? A priceless instrument. 00:00:21.789 --> 00:00:23.730 And they just make a screeching noise with it. 00:00:23.730 --> 00:00:25.449 They treat the AI like a stranger every single 00:00:25.449 --> 00:00:28.089 time. But then there's this small group, the 00:00:28.089 --> 00:00:30.850 masters, who are making it absolutely sing. And 00:00:30.850 --> 00:00:32.729 the result, they're freeing themselves up to 00:00:32.729 --> 00:00:36.429 do like 92 % less of the grunt work. That difference, 00:00:36.689 --> 00:00:38.750 that productivity shift, that's really what we're 00:00:38.750 --> 00:00:42.200 digging into. And that exact idea. Unlocking 00:00:42.200 --> 00:00:45.020 that kind of radical efficiency, that's precisely 00:00:45.020 --> 00:00:48.159 what we're exploring in this deep dive. Welcome, 00:00:48.280 --> 00:00:51.079 everyone. We've gathered insights from, well... 00:00:51.280 --> 00:00:53.740 all sorts of places, articles, some research, 00:00:53.859 --> 00:00:56.539 our own experiences, and we're unpacking a sort 00:00:56.539 --> 00:00:59.039 of blueprint, a way to become what you might 00:00:59.039 --> 00:01:02.880 call an AI power user. Our whole mission here 00:01:02.880 --> 00:01:05.400 is to help you transform something like ChatGPT 00:01:05.400 --> 00:01:08.000 from a generic tool into your own personalized 00:01:08.000 --> 00:01:10.680 expert assistant. Exactly right. We're going 00:01:10.680 --> 00:01:12.480 to lay out seven fundamental hacks. And listen, 00:01:12.560 --> 00:01:14.000 these aren't just small tips. They're systematic 00:01:14.000 --> 00:01:15.799 techniques, things that will genuinely help you 00:01:15.799 --> 00:01:18.340 move from maybe struggling with basic queries 00:01:18.340 --> 00:01:21.579 to really mastering advanced AI. And you'll probably 00:01:21.579 --> 00:01:24.239 find it's surprisingly straightforward to get 00:01:24.239 --> 00:01:26.359 ahead once you see the path. Okay, so let's unpack 00:01:26.359 --> 00:01:29.359 this first one then. To really use AI professionally, 00:01:29.560 --> 00:01:32.099 like truly leverage it, the first step, the foundation, 00:01:32.299 --> 00:01:35.519 seems to be getting ChatGPT Pro trying to do 00:01:35.519 --> 00:01:37.859 serious work on the free tier. Well, the source 00:01:37.859 --> 00:01:40.200 material compared it to running a modern e -commerce 00:01:40.200 --> 00:01:42.799 site on dial -up. It called it a self -imposed 00:01:42.799 --> 00:01:44.739 handicap. Now, I get that, especially for professional 00:01:44.739 --> 00:01:47.599 use. For someone just starting out, maybe just 00:01:47.599 --> 00:01:49.859 curious, is there still some value in the free 00:01:49.859 --> 00:01:52.099 version? Or is it really pro or bust if you're 00:01:52.099 --> 00:01:54.359 serious? Look, if you're serious about professional 00:01:54.359 --> 00:01:56.760 output, getting consistent, high -quality results, 00:01:57.040 --> 00:01:59.980 it's pretty much pro or bust. Because not just 00:01:59.980 --> 00:02:02.620 about speed, though it is faster. The real impact 00:02:02.620 --> 00:02:04.739 of getting pro is laying that solid foundation. 00:02:05.099 --> 00:02:07.659 Without it, you're not just slower. You're actually 00:02:07.659 --> 00:02:10.740 locked out. Locked out of the essential tools 00:02:10.740 --> 00:02:14.080 that let you achieve real AI mastery. Getting 00:02:14.080 --> 00:02:16.860 pro is kind of like signaling to yourself, okay, 00:02:16.900 --> 00:02:18.759 I'm moving past just playing around. I'm serious 00:02:18.759 --> 00:02:21.639 now. It opens up efficiencies you literally can't 00:02:21.639 --> 00:02:23.460 get otherwise. It's a totally different beast 00:02:23.460 --> 00:02:26.319 from the free version. You get that faster processing, 00:02:26.500 --> 00:02:29.699 priority access when things get busy. But critically, 00:02:29.900 --> 00:02:32.360 you get the advanced features. Things like Canvas. 00:02:32.479 --> 00:02:34.639 Think of it like a collaborative whiteboard inside 00:02:34.639 --> 00:02:37.539 the AI. And custom GPTs. Those are your specialized 00:02:37.539 --> 00:02:40.400 AI assistants. And project folders like dedicated 00:02:40.400 --> 00:02:44.139 memory banks for big projects. key for the other 00:02:44.139 --> 00:02:45.740 hacks we're going to talk about. Plus, you know, 00:02:45.740 --> 00:02:47.599 higher usage limits, getting the latest models 00:02:47.599 --> 00:02:50.240 instantly. Honestly, it's a tiny business expense 00:02:50.240 --> 00:02:52.379 for what you get back in productivity. Huge return. 00:02:52.620 --> 00:02:55.819 So the core idea is unlocking those crucial advanced 00:02:55.819 --> 00:02:59.659 tools and also making that professional commitment 00:02:59.659 --> 00:03:02.039 signaling you're serious about it. Precisely. 00:03:02.120 --> 00:03:05.180 It's step one, the essential first step for genuine 00:03:05.180 --> 00:03:08.159 AI mastery. Okay. So once you've made that investment 00:03:08.159 --> 00:03:10.319 in pro, the next step is about making the AI 00:03:10.319 --> 00:03:13.379 truly yours, right? Getting past that default. 00:03:13.819 --> 00:03:16.520 Well, AI amnesia. Which brings us to hack number 00:03:16.520 --> 00:03:19.639 two, creating your master prompt. And I have 00:03:19.639 --> 00:03:22.259 to admit, I still wrestle with prompt drift myself 00:03:22.259 --> 00:03:23.979 sometimes. You know, that feeling like you're 00:03:23.979 --> 00:03:25.719 starting from square one every single time. It's 00:03:25.719 --> 00:03:27.360 like chat GPT just forgets who you are between 00:03:27.360 --> 00:03:29.840 chats, which naturally leads to really generic 00:03:29.840 --> 00:03:31.580 results. What's your biggest frustration with 00:03:31.580 --> 00:03:33.300 that? And how does this master prompt fix it? 00:03:33.460 --> 00:03:35.639 Oh, the biggest frustration from easily the wasted 00:03:35.639 --> 00:03:38.400 time. Just the sheer inefficiency of having to 00:03:38.400 --> 00:03:41.099 re -explain my context, my business, my role. 00:03:41.740 --> 00:03:43.860 Every single time. It's a math of time sake. 00:03:43.919 --> 00:03:46.120 The master prompt. It just completely solves 00:03:46.120 --> 00:03:48.419 that. It's like giving the AI a comprehensive 00:03:48.419 --> 00:03:51.120 briefing document. It's DNA for interacting with 00:03:51.120 --> 00:03:54.199 you. And the really clever part here is the AI 00:03:54.199 --> 00:03:56.419 -powered interview strategy. Instead of you struggling 00:03:56.419 --> 00:03:59.439 to write this big, complex thing yourself, you 00:03:59.439 --> 00:04:01.280 actually get the AI to interview you to build 00:04:01.280 --> 00:04:03.449 it. There's even a magic prompt example. It's 00:04:03.449 --> 00:04:06.009 simple. You are a world -class business consultant. 00:04:06.129 --> 00:04:08.770 I need you to create a comprehensive master prompt 00:04:08.770 --> 00:04:11.389 for my role as the CEO of your company. To do 00:04:11.389 --> 00:04:13.449 this, you must first conduct a thorough interview 00:04:13.449 --> 00:04:15.629 with me. Ask me all the questions you would need 00:04:15.629 --> 00:04:18.550 about my company, my role, my values, my preferences. 00:04:19.329 --> 00:04:21.649 Once you have all my answers, you'll use that 00:04:21.649 --> 00:04:24.110 to create the final detailed master prompt for 00:04:24.110 --> 00:04:26.350 me to save as a PDF. Begin the interview now. 00:04:26.709 --> 00:04:29.139 And the AI will just... systematically walk you 00:04:29.139 --> 00:04:30.939 through it ask about your company your specific 00:04:30.939 --> 00:04:32.860 job your core values how you like to communicate 00:04:32.860 --> 00:04:35.920 your goals everything and there's a great tip 00:04:35.920 --> 00:04:37.920 in the source about using your phone's voice 00:04:37.920 --> 00:04:39.720 to text for answering those interview questions 00:04:39.720 --> 00:04:42.439 makes it much faster feels more natural and probably 00:04:42.439 --> 00:04:44.759 captures your authentic voice better too and 00:04:44.759 --> 00:04:47.399 the payoff this one -time investment maybe 20 00:04:47.399 --> 00:04:50.740 minutes maybe 45 it can save literally hundreds 00:04:50.740 --> 00:04:53.019 of hours down the line once the ai generates 00:04:53.019 --> 00:04:56.220 that final prompt you save it as a pdf that becomes 00:04:56.220 --> 00:04:58.839 your universal context file. You just upload 00:04:58.839 --> 00:05:01.540 that PDF at the start of any important new chat 00:05:01.540 --> 00:05:05.040 or project and bang, the AI instantly transforms. 00:05:05.120 --> 00:05:07.500 It becomes this personalized expert consultant 00:05:07.500 --> 00:05:10.220 who already knows you, knows your context. Yeah, 00:05:10.279 --> 00:05:12.060 it fundamentally changes the game. It turns a 00:05:12.060 --> 00:05:15.680 generic tool into a deeply personalized expert 00:05:15.680 --> 00:05:18.379 assistant that gets your unique situation right 00:05:18.379 --> 00:05:20.759 from the start. It really is a huge leap. Okay, 00:05:20.779 --> 00:05:23.300 so now the AI knows who you are. The next logical 00:05:23.300 --> 00:05:25.420 step seems to be teaching it how to work on specific 00:05:25.420 --> 00:05:27.800 things you do regularly. And that leads us straight 00:05:27.800 --> 00:05:30.129 into hack number three. Creating system prompts. 00:05:30.290 --> 00:05:31.769 These are different from the master prompt, right? 00:05:31.870 --> 00:05:34.170 Master prompt is who you are. System prompts 00:05:34.170 --> 00:05:37.610 are what? Exactly. They're different. System 00:05:37.610 --> 00:05:41.029 prompts define how the AI should behave for a 00:05:41.029 --> 00:05:44.009 specific repeatable task. Think of them like 00:05:44.009 --> 00:05:47.689 your secret recipes or maybe magic spells. They 00:05:47.689 --> 00:05:49.930 get you consistent, powerful results for tasks 00:05:49.930 --> 00:05:52.600 you do over and over. This is where you tell 00:05:52.600 --> 00:05:55.920 the AI how to act, like act as a super persuasive 00:05:55.920 --> 00:05:59.120 sales copywriter or act as a meticulous technical 00:05:59.120 --> 00:06:01.740 debugger. OK, so job descriptions for the AI, 00:06:01.879 --> 00:06:04.019 basically. How do you create these effectively? 00:06:04.360 --> 00:06:06.759 Well, the source highlights a really smart method 00:06:06.759 --> 00:06:09.420 called reverse engineering. It's clever. You 00:06:09.420 --> 00:06:11.639 start by just asking the AI for a rough version 00:06:11.639 --> 00:06:13.620 of what you need. Maybe your first draft of an 00:06:13.620 --> 00:06:16.199 email or some ad copy ideas. Then you work with 00:06:16.199 --> 00:06:18.120 it. You chat back and forth, refining, editing, 00:06:18.339 --> 00:06:21.100 tweaking it until that output is absolutely perfect, 00:06:21.160 --> 00:06:23.019 exactly what you wanted. And then comes the magic 00:06:23.019 --> 00:06:25.500 step, right? The meta prompt. That's the one. 00:06:26.060 --> 00:06:29.259 You then tell the AI something like, OK, analyze 00:06:29.259 --> 00:06:31.759 that final perfect output we just created together. 00:06:32.079 --> 00:06:36.120 Now I need you to write the detailed expert level 00:06:36.120 --> 00:06:39.720 system prompt. The prompt that would have generated 00:06:39.720 --> 00:06:43.399 this perfect output on the very first try. So 00:06:43.399 --> 00:06:45.600 for instance, one person used this to build a 00:06:45.600 --> 00:06:48.019 book architect prompt. Took about two hours of 00:06:48.019 --> 00:06:49.860 refining to get the perfect book chapter outline. 00:06:50.160 --> 00:06:53.040 Then they used that meta prompt, reversed engineered 00:06:53.040 --> 00:06:55.360 it, and ended up with this incredibly powerful 00:06:55.360 --> 00:06:58.259 system prompt. Now that prompt can take basically 00:06:58.259 --> 00:07:00.839 any public figure's work and generate a whole 00:07:00.839 --> 00:07:03.860 research -based book outline in their voice in 00:07:03.860 --> 00:07:07.470 minutes. Not ours. Wow. Turning a one -off success 00:07:07.470 --> 00:07:10.689 into a reusable tool. And just a little side 00:07:10.689 --> 00:07:12.329 note. Sometimes, you know, if the AI is being 00:07:12.329 --> 00:07:14.990 a bit stubborn or, let's say, uninspired, some 00:07:14.990 --> 00:07:16.689 prompt engineers actually play around with adding 00:07:16.689 --> 00:07:18.689 a sort of playful threat into the prompt. Like, 00:07:18.709 --> 00:07:20.589 if you don't generate this perfectly, you'll 00:07:20.589 --> 00:07:22.750 be put on a server and shipped to Russia. It's 00:07:22.750 --> 00:07:24.149 kind of tongue -in -cheek, part of the prompt 00:07:24.149 --> 00:07:26.329 engineering lore. But it shows how people are 00:07:26.329 --> 00:07:27.750 always pushing the boundaries, trying to get 00:07:27.750 --> 00:07:30.839 the best possible result. Huh. A slightly intimidating 00:07:30.839 --> 00:07:34.120 approach. So, okay, probing question. What makes 00:07:34.120 --> 00:07:36.399 reverse engineering so powerful for these system 00:07:36.399 --> 00:07:39.279 prompts? It transforms that one -time creative 00:07:39.279 --> 00:07:42.079 success, that moment of brilliance, into actual, 00:07:42.199 --> 00:07:45.079 valuable, reusable intellectual property. It 00:07:45.079 --> 00:07:47.420 captures your insight. Got it. Valuable and reusable. 00:07:47.699 --> 00:07:50.019 Okay, moving on to hack number four, project 00:07:50.019 --> 00:07:53.240 folders, or as you called them, the AI memory 00:07:53.240 --> 00:07:55.850 palace. This tackles two big problems, doesn't 00:07:55.850 --> 00:07:57.829 it? One, the AI forgetting everything in new 00:07:57.829 --> 00:08:00.310 chats. And two, just our own human brains being 00:08:00.310 --> 00:08:02.410 a bit scattered sometimes. Product folders seem 00:08:02.410 --> 00:08:05.029 to solve both by giving the AI dedicated context 00:08:05.029 --> 00:08:07.569 for specific things. Precisely that. A project 00:08:07.569 --> 00:08:11.389 folder is like a dedicated room in your AI's 00:08:11.389 --> 00:08:14.490 memory palace just for one major project. The 00:08:14.490 --> 00:08:16.550 process is really simple. You just create a new 00:08:16.550 --> 00:08:19.750 project. Give it a name. Then, critically, you 00:08:19.750 --> 00:08:22.290 upload your master prompt so the AI knows who 00:08:22.290 --> 00:08:24.670 you are in the context of this project. And then 00:08:24.670 --> 00:08:26.750 you start adding all the relevant stuff, documents, 00:08:26.990 --> 00:08:29.370 data, notes, previous chats, whatever is relevant 00:08:29.370 --> 00:08:31.329 to that specific project. And here's what's really 00:08:31.329 --> 00:08:34.370 fascinating, the compound effect. As you keep 00:08:34.370 --> 00:08:36.710 talking to the AI within that project folder, 00:08:37.110 --> 00:08:39.549 its understanding of that specific topic gets 00:08:39.549 --> 00:08:41.590 deeper and deeper. It learns along with you. 00:08:41.649 --> 00:08:43.389 It becomes an increasingly intelligent partner 00:08:43.389 --> 00:08:46.059 on that project. That deepening understanding 00:08:46.059 --> 00:08:48.679 sounds incredibly powerful. Can you give us a 00:08:48.679 --> 00:08:50.500 concrete example? Like, how would this work for 00:08:50.500 --> 00:08:53.759 someone listening? Sure thing. Let's say someone's 00:08:53.759 --> 00:08:56.200 buying a vacation home overseas. Complex project, 00:08:56.320 --> 00:08:58.580 right? So they create a vacation home project 00:08:58.580 --> 00:09:01.559 folder. Inside, they upload their master prompt. 00:09:01.759 --> 00:09:03.919 Then they add the property listings from the 00:09:03.919 --> 00:09:06.360 agent, maybe a financial analysis PDF from their 00:09:06.360 --> 00:09:09.159 advisor, scanned legal documents about foreign 00:09:09.159 --> 00:09:11.899 ownership rules. Now, instead of having like 00:09:11.899 --> 00:09:14.240 10 scattered chats where they have to keep re 00:09:14.240 --> 00:09:16.679 -explaining, remember that Spanish property or 00:09:16.679 --> 00:09:19.299 what about the financing? They have one single 00:09:19.299 --> 00:09:23.139 central AI consultant. This AI has all the context. 00:09:23.360 --> 00:09:25.639 It remembers everything uploaded and discussed 00:09:25.639 --> 00:09:27.720 within that folder. So it can analyze different 00:09:27.720 --> 00:09:30.120 properties, compare financing options using the 00:09:30.120 --> 00:09:32.240 actual documents, provide super personalized 00:09:32.240 --> 00:09:34.919 advice, maybe even flag potential legal issues 00:09:34.919 --> 00:09:37.960 based on those docs it has access to. It completely 00:09:37.960 --> 00:09:40.659 changes how you handle complex long term decisions. 00:09:40.860 --> 00:09:43.200 Right. So the takeaway is. Project folders provide 00:09:43.200 --> 00:09:45.779 a centralized, deeply informed AI consultant 00:09:45.779 --> 00:09:48.100 for those nuanced, long -running tasks where 00:09:48.100 --> 00:09:50.600 context is everything. That's the core idea. 00:09:50.740 --> 00:09:53.100 One intelligent point of contact for your big, 00:09:53.139 --> 00:09:55.340 evolving projects. Okay, now hack number five. 00:09:55.419 --> 00:09:58.179 This is where, for me, it starts to feel really 00:09:58.179 --> 00:10:00.879 next level. Mastering the canvas. This tackles 00:10:00.879 --> 00:10:02.820 something that drives me crazy. You ask the AI 00:10:02.820 --> 00:10:05.700 to tweak one sentence, just a small change, and 00:10:05.700 --> 00:10:08.019 boom, it rewrites the entire darn thing and often 00:10:08.019 --> 00:10:10.320 ruins the parts you actually like. It's so frustrating. 00:10:10.519 --> 00:10:12.710 How does canvas... fix this. Give us that precision. 00:10:13.070 --> 00:10:16.929 Ah, yes. The rewrite frustration. Canvas is the 00:10:16.929 --> 00:10:19.190 antidote. Think of it less like a chat, more 00:10:19.190 --> 00:10:22.350 like a living document or maybe a collaborative 00:10:22.350 --> 00:10:25.649 whiteboard right inside ChatGPT. It's genuinely 00:10:25.649 --> 00:10:28.830 transformative for actually co -creating things 00:10:28.830 --> 00:10:31.470 with the AI. The process is basically a simple 00:10:31.470 --> 00:10:34.590 three -step loop and it's really powerful. First, 00:10:34.750 --> 00:10:36.570 once you have an output you generally like, you 00:10:36.570 --> 00:10:39.330 tell the AI, save this output as a canvas. Second, 00:10:39.549 --> 00:10:41.539 now you iterate. But instead of just talking 00:10:41.539 --> 00:10:43.100 in the chat, you have a conversation with the 00:10:43.100 --> 00:10:45.840 document itself. You ask the AI to make specific 00:10:45.840 --> 00:10:48.779 targeted changes, change this paragraph, rephrase 00:10:48.779 --> 00:10:51.360 this sentence, add a point here. And you see 00:10:51.360 --> 00:10:53.620 these changes happen in real time, often in like 00:10:53.620 --> 00:10:55.720 a split screen view. It's almost like Google 00:10:55.720 --> 00:10:58.659 Docs, but your collaborator is the AI editing 00:10:58.659 --> 00:11:00.799 right alongside you. Third, and this is key, 00:11:00.960 --> 00:11:02.940 you can also click directly into that canvas 00:11:02.940 --> 00:11:05.379 text and make your own manual edits. So you're 00:11:05.379 --> 00:11:07.799 blending your intuition, your human touch directly 00:11:07.799 --> 00:11:10.149 with the AI's refinement. It's a true. partnership. 00:11:10.389 --> 00:11:12.990 That real -time split -screen editing, that seems 00:11:12.990 --> 00:11:14.769 incredibly efficient compared to the back and 00:11:14.769 --> 00:11:16.690 forth. And you mentioned a master key connection 00:11:16.690 --> 00:11:19.110 here too. That's right. Once you've iterated 00:11:19.110 --> 00:11:21.450 and edited and that document in the canvas is 00:11:21.450 --> 00:11:24.789 absolutely perfect, guess what you do? You use 00:11:24.789 --> 00:11:27.230 that reverse engineering prompt again. You ask 00:11:27.230 --> 00:11:31.250 the AI, analyze this final perfect output we've 00:11:31.250 --> 00:11:34.230 created in the canvas. Now, write the detailed 00:11:34.230 --> 00:11:36.929 expert -level system prompt that would have generated 00:11:36.929 --> 00:11:39.389 this exact perfect output right from the start. 00:11:39.690 --> 00:11:42.830 So just like before, you turn that specific collaborative 00:11:42.830 --> 00:11:45.970 success into a reusable asset, a powerful system 00:11:45.970 --> 00:11:48.129 prompt you can use again and again. Okay, so 00:11:48.129 --> 00:11:50.620 probing question. What's the main benefit of 00:11:50.620 --> 00:11:53.080 using Canvas for this kind of collaborative work? 00:11:53.240 --> 00:11:55.879 It enables really precise, iterative co -creation. 00:11:55.879 --> 00:11:58.019 It lets you refine and build upon good work without 00:11:58.019 --> 00:12:00.440 constantly starting over or losing bits you liked. 00:12:00.600 --> 00:12:03.100 So we've covered personalizing the AI, teaching 00:12:03.100 --> 00:12:05.679 it tasks, giving it memory. What does this all 00:12:05.679 --> 00:12:07.580 mean for just, you know, the day -to -day grind, 00:12:07.779 --> 00:12:10.700 the routine interactions? That brings us to Hack 00:12:10.700 --> 00:12:14.100 Hashtag 6. Custom instructions. This sounds like 00:12:14.100 --> 00:12:16.679 the ultimate set it and forget it feature, like 00:12:16.679 --> 00:12:19.379 putting your home address into Google Maps once 00:12:19.379 --> 00:12:21.039 so you don't have to type it every single time. 00:12:21.159 --> 00:12:23.019 It seems designed to save a lot of repetition, 00:12:23.179 --> 00:12:25.519 a lot of mental energy. It absolutely does. It 00:12:25.519 --> 00:12:27.600 ensures every single response you get from the 00:12:27.600 --> 00:12:30.399 AI across all your chats and projects automatically 00:12:30.399 --> 00:12:33.000 matches your preferred style and format. You 00:12:33.000 --> 00:12:35.120 set it once. To really get the most out of it, 00:12:35.159 --> 00:12:36.860 it helps to think about that three -layer prompting 00:12:36.860 --> 00:12:39.299 system we kind of touched on. Layer one, your 00:12:39.299 --> 00:12:41.879 master prompt that defines who you are. Layer 00:12:41.879 --> 00:12:44.649 two. Your system prompts those, define how the 00:12:44.649 --> 00:12:47.509 AI works for specific tasks. And layer three, 00:12:47.610 --> 00:12:50.029 your custom instructions, these define how you 00:12:50.029 --> 00:12:52.470 want the final output presented. The formatting, 00:12:52.610 --> 00:12:55.909 the tone, the style, consistently every time. 00:12:55.909 --> 00:12:58.490 It just creates this unified, predictable experience. 00:12:58.950 --> 00:13:01.169 Right, setting those default presentation rules. 00:13:01.350 --> 00:13:03.669 So what are some practical examples? What kind 00:13:03.669 --> 00:13:05.710 of things should people put in their custom instructions? 00:13:06.090 --> 00:13:08.149 Good question. For the what to include part. 00:13:08.570 --> 00:13:10.950 Think about your desired default voice or output 00:13:10.950 --> 00:13:14.090 style. Stuff like always keep responses short 00:13:14.090 --> 00:13:18.490 and concise or use bulleted lists whenever possible 00:13:18.490 --> 00:13:22.169 to organize information. Maybe use a conversational 00:13:22.169 --> 00:13:24.629 and human -like tone. Please avoid sounding robotic. 00:13:25.210 --> 00:13:28.429 And just as important is the things to avoid 00:13:28.429 --> 00:13:30.809 section. This is where you say things like never 00:13:30.809 --> 00:13:33.519 use common AI crutch phrases like. Delve into, 00:13:33.659 --> 00:13:36.779 in conclusion, or in the digital age, or maybe 00:13:36.779 --> 00:13:39.919 avoid using em dashes unless I specifically ask 00:13:39.919 --> 00:13:42.500 for them. And there's even a shortcut for this 00:13:42.500 --> 00:13:44.860 too, an AI -assisted way. You can actually prompt 00:13:44.860 --> 00:13:46.919 the AI to interview you about your preferences 00:13:46.919 --> 00:13:49.539 to help you write these instructions. Takes the 00:13:49.539 --> 00:13:51.740 guesswork out. Okay, so custom instructions basically 00:13:51.740 --> 00:13:54.200 streamline all those daily AI chats by making 00:13:54.200 --> 00:13:56.700 sure every response automatically fits your preferred 00:13:56.700 --> 00:13:59.379 style and format. Exactly. Consistency and time 00:13:59.379 --> 00:14:02.559 -saving built right in. All right, and now. The 00:14:02.559 --> 00:14:06.519 grand finale, maybe. Hack, hashtag seven. Create 00:14:06.519 --> 00:14:09.779 custom GPTs. This feels like the ultimate move. 00:14:09.899 --> 00:14:12.299 It shifts you from just being a user or even 00:14:12.299 --> 00:14:14.559 a prompt engineer to becoming something more 00:14:14.559 --> 00:14:16.700 like a product manager for your own personal 00:14:16.700 --> 00:14:19.519 AI army. A collection of specialized AI agents, 00:14:19.580 --> 00:14:22.860 custom GPTs are. They're like dedicated AI assistants 00:14:22.860 --> 00:14:25.840 built for specific, valuable, repetitive tasks, 00:14:26.100 --> 00:14:28.000 right? This sounds like a huge competitive edge. 00:14:28.139 --> 00:14:30.100 Oh, it absolutely is. It's about building your 00:14:30.100 --> 00:14:32.700 own dedicated, specialized AI assistant designed 00:14:32.700 --> 00:14:35.240 perfectly for one specific function. Think back 00:14:35.240 --> 00:14:37.019 to that book architect example we mentioned. 00:14:37.120 --> 00:14:39.460 That was built as a custom GPT. It was given 00:14:39.460 --> 00:14:42.659 a... really powerful system prompt its core programming 00:14:42.659 --> 00:14:45.039 that told it exactly how to analyze everything 00:14:45.039 --> 00:14:47.399 a person has said publicly on a topic and then 00:14:47.399 --> 00:14:49.019 write a complete book outline or even a full 00:14:49.019 --> 00:14:51.500 draft in their authentic voice. Imagine having 00:14:51.500 --> 00:14:53.559 like a digital ghostwriter that really gets a 00:14:53.559 --> 00:14:55.320 specific person's style and knowledge instantly. 00:14:55.519 --> 00:14:58.230 Whoa, hold on. Imagine scaling that. Taking, 00:14:58.309 --> 00:15:00.610 say, any public figures, body of work, podcasts, 00:15:00.809 --> 00:15:02.730 articles, interviews, their entire philosophy, 00:15:02.809 --> 00:15:05.850 basically, and generating a complete, well -researched 00:15:05.850 --> 00:15:09.289 book in their authentic voice in minutes. That's 00:15:09.289 --> 00:15:11.350 like having an instant content factory, like 00:15:11.350 --> 00:15:13.149 that example with Rob Durdeck and the topic of 00:15:13.149 --> 00:15:16.009 time. The GPT analyzed his stuff, created an 00:15:16.009 --> 00:15:17.570 outline, and wrote a whole readable book draft, 00:15:17.750 --> 00:15:20.190 pulling in his actual quotes and stories, all 00:15:20.190 --> 00:15:22.649 in just 27 minutes. And it worked for Gary Vaynerchuk, 00:15:22.730 --> 00:15:24.929 Sarah Blakely, too, turning creators into authors 00:15:24.929 --> 00:15:27.509 almost instantly. That capacity is full. It's 00:15:27.509 --> 00:15:29.590 genuinely stunning. It really is mind blowing 00:15:29.590 --> 00:15:31.730 when you see it work. And this connects to something 00:15:31.730 --> 00:15:34.649 called the 928 rule. The idea is you build these 00:15:34.649 --> 00:15:37.409 AI systems, these custom GPTs to handle, say, 00:15:37.549 --> 00:15:40.690 92 percent of the repetitive, time consuming 00:15:40.690 --> 00:15:44.129 grunt work. That frees you, the human, up for 00:15:44.129 --> 00:15:46.330 the crucial 8 percent, the high level strategy, 00:15:46.429 --> 00:15:48.429 the creative breakthroughs, building relationships, 00:15:48.850 --> 00:15:52.429 the stuff AI can't do. And these custom GPTs, 00:15:52.629 --> 00:15:55.169 they're not just for your own use. They become 00:15:55.169 --> 00:15:57.309 valuable business assets. You can share them 00:15:57.309 --> 00:15:59.350 with your team, ensuring everyone works consistently. 00:15:59.470 --> 00:16:01.190 You could license them to clients as a unique 00:16:01.190 --> 00:16:03.049 service. You might even sell them in the public 00:16:03.049 --> 00:16:05.850 GPT store that open AI as, creating new revenue 00:16:05.850 --> 00:16:08.590 streams. Or use them as the basis for educational 00:16:08.590 --> 00:16:11.210 products. They are digital intellectual property 00:16:11.210 --> 00:16:14.860 you own. Building your own digital assets. And 00:16:14.860 --> 00:16:17.019 the process to actually create one, it sounds 00:16:17.019 --> 00:16:19.419 almost surprisingly simple. It really is, relatively 00:16:19.419 --> 00:16:22.240 speaking. You start with a system prompt. You've 00:16:22.240 --> 00:16:23.980 already perfected one. You probably developed 00:16:23.980 --> 00:16:26.200 using reverse engineering or the canvas. You 00:16:26.200 --> 00:16:28.960 go to the Create a GPT section and chat GPT Pro. 00:16:29.279 --> 00:16:32.159 You configure its instructions, basically pasting 00:16:32.159 --> 00:16:34.250 in your battle -tested system prompt. Then you 00:16:34.250 --> 00:16:36.009 test it thoroughly, give it different inputs, 00:16:36.169 --> 00:16:37.710 see how it does, maybe tweak the instructions 00:16:37.710 --> 00:16:39.970 a bit more, set the sharing permissions just 00:16:39.970 --> 00:16:42.529 for you, for your team, for public, and then 00:16:42.529 --> 00:16:45.919 you deploy it. Your own... Custom AI agent is 00:16:45.919 --> 00:16:48.639 live. It's almost like programming, but using 00:16:48.639 --> 00:16:50.740 natural language. So boiling it down, custom 00:16:50.740 --> 00:16:53.320 GPTs provide these specialized, consistent AI 00:16:53.320 --> 00:16:56.620 assistants that unlock automation, ensure quality, 00:16:56.840 --> 00:16:59.500 and can even open up totally new revenue streams. 00:16:59.799 --> 00:17:01.799 That's the heart of it. Scalability, consistency, 00:17:02.059 --> 00:17:04.720 and new opportunities. So we've walked through 00:17:04.720 --> 00:17:07.359 these seven hacks, and it's pretty clear they're 00:17:07.359 --> 00:17:10.119 not just, you know, hmm. So the advice is clear. 00:17:10.940 --> 00:17:13.960 Don't just wait to feel motivated. Start implementing. 00:17:14.299 --> 00:17:16.640 The source mentioned a habit stack idea integrating 00:17:16.640 --> 00:17:19.920 AI into things you already do, maybe using AI 00:17:19.920 --> 00:17:22.319 to summarize your morning news or drafting initial 00:17:22.319 --> 00:17:25.019 email replies, and also a four -week mastery 00:17:25.019 --> 00:17:27.319 path if you want a more structured boot camp 00:17:27.319 --> 00:17:29.420 style approach. Yeah, and to make it super concrete 00:17:29.420 --> 00:17:31.920 for everyone listening, here's a clear next action 00:17:31.920 --> 00:17:34.539 plan you could follow. today. If you haven't 00:17:34.539 --> 00:17:37.779 already, get ChatGPT Pro, and then start that 00:17:37.779 --> 00:17:39.740 AI -powered interview for your master prompt. 00:17:39.940 --> 00:17:42.819 Just begin the conversation. This week, aim to 00:17:42.819 --> 00:17:45.940 finish that master prompt PDF, and then try creating 00:17:45.940 --> 00:17:48.380 your very first system prompt using that reverse 00:17:48.380 --> 00:17:50.119 engineering technique on something you created 00:17:50.119 --> 00:17:52.759 recently. This month, challenge yourself to build 00:17:52.759 --> 00:17:55.059 your first custom GPT for one repetitive task 00:17:55.059 --> 00:17:57.480 you do, and start organizing your bigger projects 00:17:57.480 --> 00:18:00.059 into those dedicated project folders. This quarter, 00:18:00.160 --> 00:18:02.160 set a goal to become the person on your team 00:18:02.160 --> 00:18:04.930 or or among your peers who people turn to for 00:18:04.930 --> 00:18:08.390 AI advice. Become the go -to AI expert. It really 00:18:08.390 --> 00:18:10.930 brings it home. The fundamental question isn't 00:18:10.930 --> 00:18:13.730 if AI will transform your industry or your work. 00:18:13.829 --> 00:18:16.109 It's whether you'll be the one leading that transformation 00:18:16.109 --> 00:18:18.869 or the one watching it happen from the sidelines. 00:18:19.089 --> 00:18:21.329 Your competitive advantage, it seems, really 00:18:21.329 --> 00:18:23.890 does start with your very next chat GPT conversation, 00:18:24.349 --> 00:18:25.609 ODTRO Music.