1 00:00:00,000 --> 00:00:01,240 All right, let's dive in. 2 00:00:01,240 --> 00:00:04,480 Today, we're going to tackle prompt engineering for ChatGPT. 3 00:00:04,480 --> 00:00:05,920 Ooh, exciting. 4 00:00:05,920 --> 00:00:07,120 I know, right? 5 00:00:07,120 --> 00:00:09,880 And get this, we've got this awesome source, 6 00:00:09,880 --> 00:00:12,800 mastering prompt engineering with ChatGPT. 7 00:00:12,800 --> 00:00:16,120 And it says that ChatGPT can boost student interest 8 00:00:16,120 --> 00:00:19,800 in education by like 40%. 9 00:00:19,800 --> 00:00:21,160 That's huge. 10 00:00:21,160 --> 00:00:24,480 40% is a big jump, but it's not magic. 11 00:00:24,480 --> 00:00:26,960 Right, it all comes down to how we talk to it. 12 00:00:26,960 --> 00:00:29,160 So like, are we learning to speak AI? 13 00:00:29,160 --> 00:00:30,680 Is that what prompt engineering is? 14 00:00:30,680 --> 00:00:32,160 You could think of it like that, yeah. 15 00:00:32,160 --> 00:00:34,760 It's about learning the art of giving ChatGPT 16 00:00:34,760 --> 00:00:36,840 like crystal clear instructions. 17 00:00:36,840 --> 00:00:38,720 So you get exactly what you need, 18 00:00:38,720 --> 00:00:40,480 whether it's for writing, analysis, 19 00:00:40,480 --> 00:00:41,960 or even like creative tasks. 20 00:00:41,960 --> 00:00:43,600 So it's not about speaking AI, 21 00:00:43,600 --> 00:00:46,300 it's about being better at communicating, period, 22 00:00:46,300 --> 00:00:47,320 even to a machine. 23 00:00:47,320 --> 00:00:49,080 Exactly, and you know, ChatGPT, 24 00:00:49,080 --> 00:00:51,880 especially this version 4.0, it's incredibly powerful, 25 00:00:51,880 --> 00:00:54,880 but it needs the right kind of like guidance. 26 00:00:54,880 --> 00:00:56,760 And prompt engineering, that's what helps us, 27 00:00:56,760 --> 00:00:57,840 you know, give it that guidance, right? 28 00:00:57,840 --> 00:01:00,000 We can maximize the quality, the relevance, 29 00:01:00,000 --> 00:01:02,040 make sure the AI's responses are accurate, 30 00:01:02,040 --> 00:01:04,000 and it all boils down to like four things. 31 00:01:04,000 --> 00:01:04,960 Okay, lay it on me. 32 00:01:04,960 --> 00:01:08,680 Clarity, specificity, context, 33 00:01:08,680 --> 00:01:11,400 and just knowing what format you want the output in. 34 00:01:11,400 --> 00:01:13,240 Okay, that makes sense when you put it that way, 35 00:01:13,240 --> 00:01:15,440 but like, how does all this play out 36 00:01:15,440 --> 00:01:17,640 when you're actually using ChatGPT? 37 00:01:17,640 --> 00:01:21,000 Sure, so let's say you want ChatGPT 38 00:01:21,000 --> 00:01:23,200 to write like a product description. 39 00:01:23,200 --> 00:01:24,040 Okay, yeah. 40 00:01:24,040 --> 00:01:25,520 If you just put in something vague, 41 00:01:25,520 --> 00:01:27,400 you know, like rate a product description, 42 00:01:27,400 --> 00:01:29,520 you're probably gonna get something pretty bland, right? 43 00:01:29,520 --> 00:01:30,360 Yeah, probably. 44 00:01:30,360 --> 00:01:31,800 But if you give it specifics, 45 00:01:31,800 --> 00:01:34,560 like the product's features, target audience, even, 46 00:01:34,560 --> 00:01:36,520 like what kind of tone you want, 47 00:01:36,520 --> 00:01:38,560 the output's gonna be way better, 48 00:01:38,560 --> 00:01:39,960 way more tailored to what you need. 49 00:01:39,960 --> 00:01:41,200 It's like giving it the recipe, 50 00:01:41,200 --> 00:01:42,840 and then it bakes the cake the way you want. 51 00:01:42,840 --> 00:01:46,680 Exactly, so think about like those noise-canceling headphones. 52 00:01:46,680 --> 00:01:47,520 The ones for students? 53 00:01:47,520 --> 00:01:49,000 Yeah, yeah, so instead of just saying, 54 00:01:49,000 --> 00:01:50,720 hey, write a description, you could say, 55 00:01:50,720 --> 00:01:53,360 write a concise and engaging product description 56 00:01:53,360 --> 00:01:55,200 for noise-canceling headphones, 57 00:01:55,200 --> 00:01:58,760 and then specify, target it at college students. 58 00:01:58,760 --> 00:01:59,600 Oh, okay. 59 00:01:59,600 --> 00:02:01,560 Focus on their ability to improve focus 60 00:02:01,560 --> 00:02:03,280 while studying and highlight features 61 00:02:03,280 --> 00:02:05,120 like comfort and affordability. 62 00:02:05,120 --> 00:02:07,240 Ooh, I see the difference. 63 00:02:07,240 --> 00:02:08,080 Right. 64 00:02:08,080 --> 00:02:08,920 Like we're setting the scene. 65 00:02:08,920 --> 00:02:09,740 Yeah, you got it. 66 00:02:09,740 --> 00:02:11,160 And that's like the tip of the iceberg. 67 00:02:11,160 --> 00:02:13,560 There are all sorts of prompting techniques we can use, 68 00:02:13,560 --> 00:02:17,040 like zero shot prompting, one shot prompting, 69 00:02:17,040 --> 00:02:19,840 and even a few shot learning. 70 00:02:19,840 --> 00:02:23,360 Hold on, though, it's not a little complicated. 71 00:02:23,360 --> 00:02:24,720 Especially that few shot thing. 72 00:02:24,720 --> 00:02:25,760 Break that down for me. 73 00:02:25,760 --> 00:02:27,600 Okay, so few shot learning, 74 00:02:27,600 --> 00:02:29,800 it's basically like giving ChatGPT 75 00:02:29,800 --> 00:02:31,760 a few examples to learn from. 76 00:02:31,760 --> 00:02:33,280 Let's stick with the headphones, right? 77 00:02:33,280 --> 00:02:35,880 So you could give it a couple of really well-written 78 00:02:35,880 --> 00:02:38,320 product descriptions for similar products. 79 00:02:38,320 --> 00:02:39,160 Okay. 80 00:02:39,160 --> 00:02:41,620 And that shows ChatGPT the style, 81 00:02:41,620 --> 00:02:43,160 the format you're going for. 82 00:02:43,160 --> 00:02:45,680 Oh, so it's kind of like giving it a reference point 83 00:02:45,680 --> 00:02:46,880 before it just jumps in. 84 00:02:46,880 --> 00:02:47,720 Precisely. 85 00:02:47,720 --> 00:02:50,480 Now, zero shot, that's when you're using 86 00:02:50,480 --> 00:02:53,220 general knowledge for simple tasks. 87 00:02:53,220 --> 00:02:55,160 You wouldn't need to give it specific examples 88 00:02:55,160 --> 00:02:58,080 if you just asked it to translate a simple phrase. 89 00:02:58,080 --> 00:02:58,920 Got it. 90 00:02:58,920 --> 00:03:00,400 And then, what was the other one? 91 00:03:00,400 --> 00:03:01,240 One shot. 92 00:03:01,240 --> 00:03:03,520 One shot, you give it just one example 93 00:03:03,520 --> 00:03:04,920 to kind of prime it. 94 00:03:04,920 --> 00:03:06,000 Got it, got it. 95 00:03:06,000 --> 00:03:08,000 So it's like a scale of complexity. 96 00:03:08,000 --> 00:03:10,200 We go from these basic requests 97 00:03:10,200 --> 00:03:13,080 to stuff that's more, I don't know, nuanced. 98 00:03:13,080 --> 00:03:13,920 Right, exactly. 99 00:03:13,920 --> 00:03:16,200 But we gotta guide ChatGPT with examples. 100 00:03:16,200 --> 00:03:17,040 Exactly. 101 00:03:17,040 --> 00:03:19,640 And when we move to those more complex tasks, 102 00:03:19,640 --> 00:03:20,920 the quality of our prompts, 103 00:03:20,920 --> 00:03:22,440 that becomes even more crucial. 104 00:03:22,440 --> 00:03:23,600 Oh, for sure, for sure. 105 00:03:23,600 --> 00:03:26,000 Our source actually says that just tweaking your prompts 106 00:03:26,000 --> 00:03:29,320 a bit can improve the relevance by 35%. 107 00:03:29,320 --> 00:03:31,000 35%, wow. 108 00:03:31,000 --> 00:03:34,520 Just from tweaking a few words, that's a big difference. 109 00:03:34,520 --> 00:03:36,160 Yeah, it can be huge. 110 00:03:36,160 --> 00:03:39,040 Okay, but we've been focusing on writing this whole time. 111 00:03:39,040 --> 00:03:41,520 What about using ChatGPT for other stuff, 112 00:03:41,520 --> 00:03:44,200 like analyzing data or doing research? 113 00:03:44,200 --> 00:03:45,960 Well, that's where things get really interesting. 114 00:03:45,960 --> 00:03:48,240 ChatGPT can do so much. 115 00:03:48,240 --> 00:03:51,160 For example, in finance, you could use it to analyze 116 00:03:51,160 --> 00:03:53,760 market trends or even make investment recommendations. 117 00:03:53,760 --> 00:03:54,600 Oh, wow. 118 00:03:54,600 --> 00:03:58,840 Okay, so how would I use prompting to analyze stock market data? 119 00:03:58,840 --> 00:04:03,680 Let's say you wanna analyze Apple's stock performance. 120 00:04:03,680 --> 00:04:06,220 Okay, instead of just asking it generically, 121 00:04:06,220 --> 00:04:07,200 you could be really specific. 122 00:04:07,200 --> 00:04:10,000 You could say like, analyze Apple's stock performance 123 00:04:10,000 --> 00:04:12,760 over the last year, focusing on the key factors 124 00:04:12,760 --> 00:04:14,440 influencing its fluctuations. 125 00:04:14,440 --> 00:04:15,280 Okay. 126 00:04:15,280 --> 00:04:16,540 Then give me a concise summary 127 00:04:16,540 --> 00:04:19,000 and highlight any potential growth opportunities. 128 00:04:19,000 --> 00:04:19,840 Oh, I see, I see. 129 00:04:19,840 --> 00:04:22,240 So you're guiding it towards specific insights, 130 00:04:22,240 --> 00:04:24,040 not just a data dump. 131 00:04:24,040 --> 00:04:25,000 Precisely. 132 00:04:25,000 --> 00:04:27,360 And this structured approach, 133 00:04:27,360 --> 00:04:30,080 it leads to much more meaningful analysis. 134 00:04:30,080 --> 00:04:30,920 Right, right. 135 00:04:30,920 --> 00:04:34,120 And the source material says that just having 136 00:04:34,120 --> 00:04:36,800 these clear instructions can boost the accuracy 137 00:04:36,800 --> 00:04:39,920 of the responses by like 85%. 138 00:04:39,920 --> 00:04:41,320 85%. 139 00:04:41,320 --> 00:04:42,160 Yeah. 140 00:04:42,160 --> 00:04:43,120 Wow, that's amazing. 141 00:04:43,120 --> 00:04:44,240 I'm starting to see why everyone's talking 142 00:04:44,240 --> 00:04:45,400 about prompt engineering. 143 00:04:45,400 --> 00:04:46,880 Yeah, it's a game changer. 144 00:04:46,880 --> 00:04:47,880 For sure. 145 00:04:47,880 --> 00:04:50,200 But like, I feel like we're just scratching the surface here. 146 00:04:50,200 --> 00:04:51,400 Oh, we are, we totally are. 147 00:04:51,400 --> 00:04:54,280 There's like a whole world of advanced techniques 148 00:04:54,280 --> 00:04:56,600 for building like really structured conversations 149 00:04:56,600 --> 00:04:58,240 with chat GPT. 150 00:04:58,240 --> 00:04:59,880 And we can see how it's being applied 151 00:04:59,880 --> 00:05:00,720 in different industries, 152 00:05:00,720 --> 00:05:03,680 like marketing, customer service, education. 153 00:05:03,680 --> 00:05:04,760 Oh, I can only imagine. 154 00:05:04,760 --> 00:05:05,520 Right. 155 00:05:05,520 --> 00:05:06,560 It's a lot to unpack. 156 00:05:06,560 --> 00:05:08,280 This is already a lot to take in. 157 00:05:08,280 --> 00:05:09,120 I know, right? 158 00:05:09,120 --> 00:05:10,480 But I'm ready for more. 159 00:05:10,480 --> 00:05:11,720 Where do we go from here? 160 00:05:11,720 --> 00:05:13,680 Well, now that we've gotten the basics down, 161 00:05:13,680 --> 00:05:15,840 you know, we've seen how it applies to different stuff. 162 00:05:15,840 --> 00:05:17,320 I think it's time to explore 163 00:05:17,320 --> 00:05:19,200 some of those advanced techniques I mentioned. 164 00:05:19,200 --> 00:05:20,040 Okay. 165 00:05:20,040 --> 00:05:21,480 So you can really level up our prompting game, 166 00:05:21,480 --> 00:05:24,320 especially for like the really tough scenarios. 167 00:05:24,320 --> 00:05:25,160 Right. 168 00:05:25,160 --> 00:05:26,000 Lead the way. 169 00:05:26,000 --> 00:05:26,840 Let's do it. 170 00:05:26,840 --> 00:05:27,680 Okay. 171 00:05:27,680 --> 00:05:29,120 So you ready to kick things up a notch? 172 00:05:29,120 --> 00:05:30,080 Oh, yeah. 173 00:05:30,080 --> 00:05:30,920 Lay it on me. 174 00:05:30,920 --> 00:05:31,760 All right. 175 00:05:31,760 --> 00:05:33,520 So we've covered like the basics of prompting. 176 00:05:33,520 --> 00:05:37,680 Now, things can get a little more interesting. 177 00:05:37,680 --> 00:05:38,520 Okay. 178 00:05:38,520 --> 00:05:40,960 There are some really cool advanced techniques out there. 179 00:05:40,960 --> 00:05:43,000 Our source talks about something called 180 00:05:43,000 --> 00:05:44,640 chain of thought prompting. 181 00:05:44,640 --> 00:05:45,600 Chain of thought. 182 00:05:45,600 --> 00:05:46,920 What's that all about? 183 00:05:46,920 --> 00:05:49,400 So like sometimes, you know, we got to break down 184 00:05:49,400 --> 00:05:52,840 big task into smaller steps, right? 185 00:05:52,840 --> 00:05:54,360 Makes it easier to manage. 186 00:05:54,360 --> 00:05:55,240 Yeah, for sure. 187 00:05:55,240 --> 00:05:56,840 Well, chain of thought prompting, 188 00:05:56,840 --> 00:05:58,960 it's kind of like doing that for chat GPT. 189 00:05:58,960 --> 00:06:00,800 It helps it think through a problem 190 00:06:00,800 --> 00:06:03,520 more logically step by step. 191 00:06:03,520 --> 00:06:05,680 So instead of giving it this like huge, 192 00:06:05,680 --> 00:06:07,760 complicated prompt all at once, 193 00:06:07,760 --> 00:06:09,520 we break it down into smaller pieces. 194 00:06:09,520 --> 00:06:10,360 Exactly. 195 00:06:10,360 --> 00:06:13,520 So imagine you're asking chat GPT to like, 196 00:06:13,520 --> 00:06:16,520 I don't know, create a financial forecast for a startup. 197 00:06:16,520 --> 00:06:17,360 Okay. 198 00:06:17,360 --> 00:06:18,180 Yeah. 199 00:06:18,180 --> 00:06:19,040 If you just hit it with that all at once, 200 00:06:19,040 --> 00:06:20,720 it might be too much. 201 00:06:20,720 --> 00:06:21,720 Yeah, I can see that. 202 00:06:21,720 --> 00:06:23,480 But if you break it down, you know, 203 00:06:23,480 --> 00:06:24,960 first ask for market analysis, 204 00:06:24,960 --> 00:06:27,960 then revenue projections, then expense estimations. 205 00:06:27,960 --> 00:06:28,800 I'm okay. 206 00:06:28,800 --> 00:06:30,720 It makes it way easier for the AI to digest 207 00:06:30,720 --> 00:06:33,640 and you'll get a more accurate, more structured output. 208 00:06:33,640 --> 00:06:35,760 So it's like we're scaffolding the thinking process. 209 00:06:35,760 --> 00:06:36,600 Yeah. 210 00:06:36,600 --> 00:06:37,800 Wouldn't that take forever? 211 00:06:37,800 --> 00:06:39,840 Like writing all those little prompts? 212 00:06:39,840 --> 00:06:41,160 Well, it might seem that way at first, 213 00:06:41,160 --> 00:06:43,680 but honestly, the improvement you get in the results, 214 00:06:43,680 --> 00:06:45,080 it's often worth it. 215 00:06:45,080 --> 00:06:45,920 Oh, okay. 216 00:06:45,920 --> 00:06:48,040 If you get the hang of it, it actually gets pretty fast. 217 00:06:48,040 --> 00:06:48,880 Okay. 218 00:06:48,880 --> 00:06:49,720 That makes sense. 219 00:06:49,720 --> 00:06:52,080 So what other advanced techniques are there? 220 00:06:52,080 --> 00:06:53,560 Besides chain of thought, I mean. 221 00:06:53,560 --> 00:06:54,400 Oh, there's a bunch. 222 00:06:54,400 --> 00:06:58,400 This source talks about a cool framework called C-R-E-A-T-E. 223 00:07:00,400 --> 00:07:02,560 C-R-E-A-T-E. 224 00:07:02,560 --> 00:07:03,400 Okay. 225 00:07:03,400 --> 00:07:04,240 Break that down for me. 226 00:07:04,240 --> 00:07:08,120 It stands for constraints, role, examples, action, task, 227 00:07:08,120 --> 00:07:08,960 and examples. 228 00:07:08,960 --> 00:07:10,080 Okay. I'm following so far. 229 00:07:10,080 --> 00:07:12,560 It's kind of like a roadmap 230 00:07:12,560 --> 00:07:14,600 for building really effective prompts. 231 00:07:14,600 --> 00:07:15,440 I like roadmaps. 232 00:07:15,440 --> 00:07:16,280 Me too. 233 00:07:16,280 --> 00:07:18,840 So constraints, role, examples, 234 00:07:18,840 --> 00:07:20,520 those are pretty self-explanatory, right? 235 00:07:20,520 --> 00:07:21,360 Yeah. 236 00:07:21,360 --> 00:07:22,880 But what about like action and task? 237 00:07:22,880 --> 00:07:24,000 Those seem kind of similar. 238 00:07:24,000 --> 00:07:25,400 Good catch. 239 00:07:25,400 --> 00:07:28,320 So action is the specific verb you want ChatGPT to do. 240 00:07:28,320 --> 00:07:31,240 Like write, analyze, or summarize. 241 00:07:31,240 --> 00:07:32,080 Okay. 242 00:07:32,080 --> 00:07:34,000 Task is like the overall objective, 243 00:07:34,000 --> 00:07:35,680 like write a marketing plan. 244 00:07:35,680 --> 00:07:36,520 Yeah. 245 00:07:36,520 --> 00:07:38,120 Or create a Python script. 246 00:07:38,120 --> 00:07:39,040 Oh, okay. 247 00:07:39,040 --> 00:07:41,480 So it's like giving ChatGPT a job title 248 00:07:41,480 --> 00:07:43,320 and like a list of responsibilities. 249 00:07:43,320 --> 00:07:44,160 Exactly. 250 00:07:44,160 --> 00:07:45,000 And then examples part, 251 00:07:45,000 --> 00:07:46,160 where real magic happens. 252 00:07:46,160 --> 00:07:47,880 Remember we were talking about few shot learning? 253 00:07:47,880 --> 00:07:48,720 Yeah. 254 00:07:48,720 --> 00:07:50,160 This framework uses that. 255 00:07:50,160 --> 00:07:52,200 You know, you build examples right into your prompts 256 00:07:52,200 --> 00:07:54,160 to give ChatGPT even more guidance. 257 00:07:54,160 --> 00:07:56,840 So it's all about giving it the tools to succeed. 258 00:07:56,840 --> 00:07:57,680 Exactly. 259 00:07:57,680 --> 00:07:59,240 And speaking of success, 260 00:07:59,240 --> 00:08:01,640 our source mentions how ChatGPT is already being used 261 00:08:01,640 --> 00:08:03,000 in like tons of industry. 262 00:08:03,000 --> 00:08:03,840 Oh, really? 263 00:08:03,840 --> 00:08:04,680 Yeah. 264 00:08:04,680 --> 00:08:05,520 It's crazy. 265 00:08:05,520 --> 00:08:06,360 Like in customer service, 266 00:08:06,360 --> 00:08:08,320 some companies are using it to make chatbots 267 00:08:08,320 --> 00:08:10,760 that can give people support super fast. 268 00:08:10,760 --> 00:08:11,600 Oh yeah. 269 00:08:11,600 --> 00:08:12,440 Those things are everywhere. 270 00:08:12,440 --> 00:08:13,480 I know, right? 271 00:08:13,480 --> 00:08:15,240 And sometimes you can't even tell if you're talking 272 00:08:15,240 --> 00:08:16,880 to a bot or a real person. 273 00:08:16,880 --> 00:08:17,720 It's true. 274 00:08:17,720 --> 00:08:18,560 It's true. 275 00:08:18,560 --> 00:08:19,600 That's kind of freaky sometimes. 276 00:08:19,600 --> 00:08:20,720 Yeah. 277 00:08:20,720 --> 00:08:23,480 But it shows how powerful these AI systems are getting. 278 00:08:23,480 --> 00:08:24,320 Yeah. 279 00:08:24,320 --> 00:08:26,120 And it's all thanks to prompt engineering, right? 280 00:08:26,120 --> 00:08:27,400 Well, partly, yeah. 281 00:08:27,400 --> 00:08:28,720 It's a big piece of the puzzle. 282 00:08:28,720 --> 00:08:32,560 So it's not just about like mimicking human conversation. 283 00:08:32,560 --> 00:08:36,080 They can actually, I don't know, do real work. 284 00:08:36,080 --> 00:08:36,920 Totally. 285 00:08:36,920 --> 00:08:37,760 Like in marketing, 286 00:08:37,760 --> 00:08:40,720 companies are using ChatGPT to write ad copy, 287 00:08:40,720 --> 00:08:43,280 even personalize email campaigns. 288 00:08:43,280 --> 00:08:44,440 Crazy stuff, right? 289 00:08:44,440 --> 00:08:45,440 It is. 290 00:08:45,440 --> 00:08:48,920 But I can't help but think about like the downsides, 291 00:08:48,920 --> 00:08:50,680 you know, what about ethics and stuff? 292 00:08:50,680 --> 00:08:51,720 Oh yeah. 293 00:08:51,720 --> 00:08:52,560 Good point. 294 00:08:52,560 --> 00:08:53,720 That's a big conversation. 295 00:08:53,720 --> 00:08:54,560 For sure. 296 00:08:54,560 --> 00:08:56,560 There are definitely ethical things to think about, right? 297 00:08:56,560 --> 00:08:57,720 Like in healthcare, 298 00:08:57,720 --> 00:09:00,040 what if AI starts making medical decisions 299 00:09:00,040 --> 00:09:01,480 that should be made by humans? 300 00:09:01,480 --> 00:09:02,320 Whoa. 301 00:09:02,320 --> 00:09:03,360 Okay, yeah. 302 00:09:03,360 --> 00:09:04,200 That's a little scary. 303 00:09:04,200 --> 00:09:05,040 Right. 304 00:09:05,040 --> 00:09:07,520 It's all about finding that balance, you know? 305 00:09:07,520 --> 00:09:11,720 Using AI strengths, but recognizing its limitations. 306 00:09:11,720 --> 00:09:12,720 So how do we do that? 307 00:09:12,720 --> 00:09:14,920 Well, that's where prompt engineering comes in, right? 308 00:09:14,920 --> 00:09:16,840 We need to make sure we're guiding AI 309 00:09:16,840 --> 00:09:18,920 in a way that's ethical, responsible, 310 00:09:18,920 --> 00:09:21,720 and ultimately benefits humans. 311 00:09:21,720 --> 00:09:22,800 That makes sense. 312 00:09:22,800 --> 00:09:26,400 This makes me think about like the future of work. 313 00:09:26,400 --> 00:09:28,960 Will robots take all our jobs? 314 00:09:28,960 --> 00:09:29,920 The big question. 315 00:09:29,920 --> 00:09:32,200 Yeah, some people are worried about that for sure. 316 00:09:32,200 --> 00:09:33,040 Rightly so. 317 00:09:33,040 --> 00:09:33,880 Yeah. 318 00:09:33,880 --> 00:09:35,880 But others think it's an opportunity, you know? 319 00:09:35,880 --> 00:09:37,840 Like humans and AI, we could work together 320 00:09:37,840 --> 00:09:39,520 in new and cool ways. 321 00:09:39,520 --> 00:09:40,360 That's a good point. 322 00:09:40,360 --> 00:09:41,920 We're still the ones writing the prompts, right? 323 00:09:41,920 --> 00:09:42,680 Exactly. 324 00:09:42,680 --> 00:09:44,600 It's not about AI taking over, it's about, you know, 325 00:09:44,600 --> 00:09:45,760 like teamwork. 326 00:09:45,760 --> 00:09:47,400 Teamwork makes the dream work, as they say. 327 00:09:47,400 --> 00:09:48,320 Exactly. 328 00:09:48,320 --> 00:09:52,280 Imagine designers using chat GPT to brainstorm ideas, 329 00:09:52,280 --> 00:09:55,880 or writers, you know, using it to get over writer's block. 330 00:09:55,880 --> 00:09:57,680 The possibilities are endless. 331 00:09:57,680 --> 00:09:59,000 My mind is blown. 332 00:09:59,000 --> 00:09:59,960 So much to think about. 333 00:09:59,960 --> 00:10:01,080 I know, right? 334 00:10:01,080 --> 00:10:02,040 It's a lot to unpack. 335 00:10:02,040 --> 00:10:05,920 So as we go forward, what are like the key takeaways? 336 00:10:05,920 --> 00:10:07,680 What do people need to keep in mind? 337 00:10:07,680 --> 00:10:10,160 Well, one thing is, you know, prompt engineering. 338 00:10:10,160 --> 00:10:11,360 It's not a one-time thing. 339 00:10:11,360 --> 00:10:12,200 Right. 340 00:10:12,200 --> 00:10:13,920 A skill you keep developing. 341 00:10:13,920 --> 00:10:14,760 Right. 342 00:10:14,760 --> 00:10:16,600 The more you practice, the better you get 343 00:10:16,600 --> 00:10:18,480 at writing good prompts. 344 00:10:18,480 --> 00:10:21,080 And, you know, AI, it's always evolving. 345 00:10:21,080 --> 00:10:22,560 There's always going to be something new to learn. 346 00:10:22,560 --> 00:10:23,400 It's like anything, right? 347 00:10:23,400 --> 00:10:24,240 Yeah. 348 00:10:24,240 --> 00:10:25,080 Practice makes perfect. 349 00:10:25,080 --> 00:10:25,920 Yeah, pretty much. 350 00:10:25,920 --> 00:10:28,160 So before we move on, is there anything else 351 00:10:28,160 --> 00:10:29,280 you think people should know? 352 00:10:29,280 --> 00:10:33,480 Yeah, I think it's important to remember that AI, it's a tool. 353 00:10:33,480 --> 00:10:34,320 Okay. 354 00:10:34,320 --> 00:10:38,200 And like any tool, it can be used for good or for bad, right? 355 00:10:38,200 --> 00:10:39,040 That's true. 356 00:10:39,040 --> 00:10:40,440 As we get better at this stuff, 357 00:10:40,440 --> 00:10:42,400 we got to use this power responsibly. 358 00:10:42,400 --> 00:10:43,880 Think about the ethics, you know? 359 00:10:43,880 --> 00:10:44,720 Yeah. 360 00:10:44,720 --> 00:10:46,640 Try to make prompts that are fair and inclusive. 361 00:10:46,640 --> 00:10:47,480 Right. 362 00:10:47,480 --> 00:10:48,320 AI for good. 363 00:10:48,320 --> 00:10:49,160 Exactly. 364 00:10:49,160 --> 00:10:51,480 It's up to us to shape the future of AI 365 00:10:51,480 --> 00:10:53,000 and make sure it benefits everyone. 366 00:10:53,000 --> 00:10:54,280 Ooh, good point. 367 00:10:54,280 --> 00:10:55,400 I like that. 368 00:10:55,400 --> 00:10:58,680 It's a reminder that we're not just along for the ride. 369 00:10:58,680 --> 00:11:00,000 We're the ones in the driver's seat. 370 00:11:00,000 --> 00:11:01,280 Exactly. 371 00:11:01,280 --> 00:11:02,280 And you know what? 372 00:11:02,280 --> 00:11:04,680 It's been awesome diving deep with you today. 373 00:11:04,680 --> 00:11:06,080 Likewise. 374 00:11:06,080 --> 00:11:07,560 This is fascinating stuff. 375 00:11:07,560 --> 00:11:08,600 It is. 376 00:11:08,600 --> 00:11:10,160 We're just getting started, though. 377 00:11:10,160 --> 00:11:12,080 There's so much more to explore. 378 00:11:12,080 --> 00:11:12,920 It's funny. 379 00:11:12,920 --> 00:11:14,200 You know, we started talking about crafting, 380 00:11:14,200 --> 00:11:17,240 like the perfect prompt, giving chat GPT 381 00:11:17,240 --> 00:11:18,760 a set of instructions, 382 00:11:18,760 --> 00:11:21,560 but it feels like we've stumbled onto something way bigger. 383 00:11:21,560 --> 00:11:22,400 Yeah. 384 00:11:22,400 --> 00:11:23,240 Yeah, I agree. 385 00:11:23,240 --> 00:11:26,280 Like a whole new way of thinking about how we communicate, 386 00:11:26,280 --> 00:11:28,640 how we like tap into creativity. 387 00:11:28,640 --> 00:11:30,280 It's a real shift, yeah. 388 00:11:30,280 --> 00:11:32,760 Prompt engineering, it's not just giving AI orders. 389 00:11:32,760 --> 00:11:34,760 It's about understanding what it can do, 390 00:11:34,760 --> 00:11:37,880 anticipating its responses, you know, 391 00:11:37,880 --> 00:11:39,360 guiding it toward what we want. 392 00:11:39,360 --> 00:11:42,080 So it's more like, I don't know, collaboration. 393 00:11:42,080 --> 00:11:43,440 We're working with the AI, 394 00:11:43,440 --> 00:11:44,880 not just telling it what to do. 395 00:11:44,880 --> 00:11:46,200 Exactly. 396 00:11:46,200 --> 00:11:47,960 And our source makes a big point about this, 397 00:11:47,960 --> 00:11:49,720 especially for creative stuff. 398 00:11:49,720 --> 00:11:53,920 Chat GPT can be like an awesome brainstorming partner. 399 00:11:53,920 --> 00:11:54,760 Oh yeah. 400 00:11:54,760 --> 00:11:56,040 Yeah, throwing out ideas, 401 00:11:56,040 --> 00:11:58,160 exploring all sorts of creative avenues. 402 00:11:58,160 --> 00:11:59,000 It's amazing. 403 00:11:59,000 --> 00:12:00,440 It's like having a brainstorming buddy 404 00:12:00,440 --> 00:12:03,440 who can process like a zillion ideas per second. 405 00:12:03,440 --> 00:12:05,160 Uh-huh, right, exactly. 406 00:12:05,160 --> 00:12:07,560 And it's not just about quantity, it's quality too. 407 00:12:07,560 --> 00:12:08,400 Okay. 408 00:12:08,400 --> 00:12:10,640 So quantity can help us get unstuck 409 00:12:10,640 --> 00:12:13,640 when we're feeling like creatively blocked, you know? 410 00:12:13,640 --> 00:12:15,000 It can challenge our assumptions 411 00:12:15,000 --> 00:12:17,360 and push us to think in new ways. 412 00:12:17,360 --> 00:12:19,720 That reminds me of how writers use prompts. 413 00:12:19,720 --> 00:12:20,600 Yeah. 414 00:12:20,600 --> 00:12:22,880 A single word can spark a whole story. 415 00:12:22,880 --> 00:12:23,720 Yeah. 416 00:12:23,720 --> 00:12:25,400 Now imagine AI doing that. 417 00:12:25,400 --> 00:12:26,240 Right. 418 00:12:26,240 --> 00:12:27,200 Like amplifying that power. 419 00:12:27,200 --> 00:12:29,920 Right, helping us build whole worlds and narratives. 420 00:12:29,920 --> 00:12:31,080 That's the future. 421 00:12:31,080 --> 00:12:31,920 It is. 422 00:12:31,920 --> 00:12:35,720 And get this, our source mentions how prompt engineering 423 00:12:35,720 --> 00:12:38,480 can be used for stuff that like normally needs 424 00:12:38,480 --> 00:12:39,880 a lot of technical know-how. 425 00:12:39,880 --> 00:12:40,920 Oh yeah. 426 00:12:40,920 --> 00:12:41,760 Give me an example. 427 00:12:41,760 --> 00:12:43,120 Well think about coding. 428 00:12:43,120 --> 00:12:48,040 Chat GPT can write code in like tons of programming languages. 429 00:12:48,040 --> 00:12:48,880 Oh, seriously? 430 00:12:48,880 --> 00:12:49,720 Yeah. 431 00:12:49,720 --> 00:12:51,760 Now this doesn't mean anyone can become a software engineer 432 00:12:51,760 --> 00:12:52,720 overnight, right? 433 00:12:52,720 --> 00:12:53,560 Right, right. 434 00:12:53,560 --> 00:12:56,360 But if you understand some basic coding ideas, 435 00:12:56,360 --> 00:12:59,720 you can use Chat GPT to make like simple programs or scripts, 436 00:12:59,720 --> 00:13:00,560 you know? 437 00:13:00,560 --> 00:13:01,400 Okay. 438 00:13:01,400 --> 00:13:04,280 Maybe automate some tasks, build a basic website even. 439 00:13:04,280 --> 00:13:05,120 Oh wow. 440 00:13:05,120 --> 00:13:07,000 Like it opens up all these possibilities. 441 00:13:07,000 --> 00:13:07,840 Exactly. 442 00:13:07,840 --> 00:13:09,720 It makes tech more accessible for everyone. 443 00:13:09,720 --> 00:13:10,640 It's empowering, right? 444 00:13:10,640 --> 00:13:13,200 It gives people the tools to create, to innovate. 445 00:13:13,200 --> 00:13:14,040 Yeah. 446 00:13:14,040 --> 00:13:15,600 And it all circles back to what we were saying 447 00:13:15,600 --> 00:13:19,360 about prompt engineering being this like huge shift, 448 00:13:19,360 --> 00:13:20,200 you know? 449 00:13:20,200 --> 00:13:23,120 It's changing how we think about like everything. 450 00:13:23,120 --> 00:13:23,960 Yeah. 451 00:13:23,960 --> 00:13:26,400 Problem solving, creativity, even work itself. 452 00:13:26,400 --> 00:13:28,000 It really is. 453 00:13:28,000 --> 00:13:29,560 Wow, this has been an awesome deep dive. 454 00:13:29,560 --> 00:13:30,400 I know, right? 455 00:13:30,400 --> 00:13:32,160 We've covered so much. 456 00:13:32,160 --> 00:13:36,360 From the basics of like writing a prompt to how it impacts, 457 00:13:36,360 --> 00:13:39,080 you know, creative work, the future of work. 458 00:13:39,080 --> 00:13:41,680 I feel like I have a whole new understanding of AI now. 459 00:13:41,680 --> 00:13:42,520 Me too. 460 00:13:42,520 --> 00:13:44,680 And I think, you know, it's important to say, 461 00:13:44,680 --> 00:13:46,120 this is just the start. 462 00:13:46,120 --> 00:13:46,960 Right. 463 00:13:46,960 --> 00:13:48,720 The world of AI, it's always changing, 464 00:13:48,720 --> 00:13:50,480 new discoveries all the time. 465 00:13:50,480 --> 00:13:52,040 It's kind of scary how fast it's moving. 466 00:13:52,040 --> 00:13:53,480 It can be, yeah. 467 00:13:53,480 --> 00:13:54,840 But it's also exciting. 468 00:13:54,840 --> 00:13:58,640 The key is to stay curious, keep learning, you know? 469 00:13:58,640 --> 00:14:00,120 Be open to new stuff. 470 00:14:00,120 --> 00:14:00,960 Right, right. 471 00:14:00,960 --> 00:14:02,400 There's so many resources out there, 472 00:14:02,400 --> 00:14:04,800 books, online courses, communities, you know? 473 00:14:04,800 --> 00:14:06,000 We can all learn from each other. 474 00:14:06,000 --> 00:14:06,840 Yeah, definitely. 475 00:14:06,840 --> 00:14:08,200 And it's important to remember, AI 476 00:14:08,200 --> 00:14:10,920 isn't just for like the tech geniuses. 477 00:14:10,920 --> 00:14:11,680 Oh, absolutely. 478 00:14:11,680 --> 00:14:15,360 It can benefit everyone, artists, entrepreneurs, anyone 479 00:14:15,360 --> 00:14:17,560 who's, you know, curious about the future. 480 00:14:17,560 --> 00:14:18,720 Exactly. 481 00:14:18,720 --> 00:14:20,520 So as we wrap up our deep dive today, 482 00:14:20,520 --> 00:14:21,880 I want to challenge our listener. 483 00:14:21,880 --> 00:14:22,560 OK, almost. 484 00:14:22,560 --> 00:14:25,480 Don't just watch this technology happen, you know? 485 00:14:25,480 --> 00:14:26,720 Be a part of it. 486 00:14:26,720 --> 00:14:28,320 Experiment with AI tools. 487 00:14:28,320 --> 00:14:30,440 Learn how to write awesome prompts. 488 00:14:30,440 --> 00:14:33,400 And then use that knowledge to make something great. 489 00:14:33,400 --> 00:14:36,280 Yeah, make something that helps you and helps the world too. 490 00:14:36,280 --> 00:14:37,760 Exactly. 491 00:14:37,760 --> 00:14:41,080 It's been a pleasure diving deep with you and our listeners. 492 00:14:41,080 --> 00:14:44,240 Until next time, keep exploring, keep learning, 493 00:14:44,240 --> 00:14:46,960 and keep pushing the boundaries of what AI can do. 494 00:14:46,960 --> 00:14:47,560 Absolutely. 495 00:14:47,560 --> 00:14:49,200 It's been a great conversation. 496 00:14:49,200 --> 00:14:52,520 And I hope everyone listening is inspired to dive 497 00:14:52,520 --> 00:14:53,520 into the world of AI. 498 00:14:53,520 --> 00:14:56,360 It's a fascinating, ever-changing landscape. 499 00:14:56,360 --> 00:14:58,720 And we're all just starting to explore its potential. 500 00:14:58,720 --> 00:15:00,920 So go out there, experiment, learn, and who knows, 501 00:15:00,920 --> 00:15:29,880 maybe you'll be the one to make the next big breakthrough.