1 00:00:00,000 --> 00:00:01,460 in the German language, right? 2 00:00:01,460 --> 00:00:03,620 They have words like Schadenfreude, 3 00:00:03,620 --> 00:00:06,880 which means pleasure from someone else's misfortune. 4 00:00:06,880 --> 00:00:09,820 What is the word for the multiplicative effect 5 00:00:09,820 --> 00:00:12,120 of two noobs together? 6 00:00:12,120 --> 00:00:14,960 ["HTTTA"] 7 00:00:19,360 --> 00:00:20,960 H-T-T-T-A. 8 00:00:20,960 --> 00:00:21,800 H-T-T-T-A. 9 00:00:21,800 --> 00:00:22,720 H-T-T-T-A. 10 00:00:22,720 --> 00:00:24,200 I-G-T-T-A. 11 00:00:24,200 --> 00:00:25,280 H-T-T-T-A. 12 00:00:25,280 --> 00:00:29,640 It's how to talk to a I with your host. 13 00:00:29,640 --> 00:00:32,320 Go to go and west the synth mind. 14 00:00:32,320 --> 00:00:33,840 Ladies and gentlemen, boys and girls, 15 00:00:33,840 --> 00:00:37,520 children of all ages, dogs, cats, robots, 16 00:00:37,520 --> 00:00:41,560 and everybody in between, we are glad you're here with us. 17 00:00:41,560 --> 00:00:46,400 Welcome to HTTTA, How to Talk to AI. 18 00:00:46,400 --> 00:00:49,120 I am your host, the synth mind in Southwest Shields, 19 00:00:49,120 --> 00:00:51,840 but I am joined by the ever voracious, 20 00:00:51,840 --> 00:00:55,760 ever capable, wonder herself, Ms. Got to Go, 21 00:00:55,760 --> 00:00:58,400 or affectionately as we know, G. 22 00:00:58,400 --> 00:00:59,560 How are you, G? 23 00:00:59,560 --> 00:01:00,720 Good, thank you. 24 00:01:00,720 --> 00:01:02,920 This is a lovely introduction. 25 00:01:02,920 --> 00:01:03,880 Hey, you know, I try. 26 00:01:03,880 --> 00:01:05,640 So why are we here? 27 00:01:05,640 --> 00:01:08,560 We want to talk to you about cramp engineering. 28 00:01:08,560 --> 00:01:11,120 I'm sure, unless you've been living under a rock, 29 00:01:11,120 --> 00:01:15,280 you've heard about this little thing called Chat GPT. 30 00:01:15,280 --> 00:01:19,200 It's just scratching the surface with the various different 31 00:01:19,200 --> 00:01:22,120 types of neural network backed models 32 00:01:22,120 --> 00:01:24,600 that are out there right now, being put forth by companies 33 00:01:24,600 --> 00:01:27,200 like OpenAI and Google and Microsoft. 34 00:01:27,200 --> 00:01:30,680 We're on the cusp of a wave right now that at least G and I 35 00:01:30,680 --> 00:01:35,360 think that this is going to be a revelatory time in history. 36 00:01:35,360 --> 00:01:38,480 And we want to hopefully hear in the coming weeks and months 37 00:01:38,480 --> 00:01:42,040 and years, share with you our journey, our experiences, 38 00:01:42,040 --> 00:01:46,160 and our insights into how to talk to these amazing machines 39 00:01:46,160 --> 00:01:48,360 and help you get the most out of them, 40 00:01:48,360 --> 00:01:52,960 be that from business insights to song lyrics 41 00:01:52,960 --> 00:01:55,520 and everything else in between that you can create 42 00:01:55,520 --> 00:01:56,920 with these wonderful tools. 43 00:01:56,920 --> 00:01:59,280 We'll talk about what prompt engineering is, 44 00:01:59,280 --> 00:02:02,160 and then a little bit more about what we want this podcast here 45 00:02:02,160 --> 00:02:03,080 to become. 46 00:02:03,080 --> 00:02:06,800 So with that, G, tell us about yourself. 47 00:02:06,800 --> 00:02:09,360 How did you find this little humble, you know, 48 00:02:09,360 --> 00:02:12,520 how did you find a GPT for the first time? 49 00:02:12,520 --> 00:02:15,360 And tell us how you came up in this world. 50 00:02:15,360 --> 00:02:18,560 It feels like it's been ages ago. 51 00:02:18,560 --> 00:02:23,960 I think I discovered Chat GPT a few days after it's launching. 52 00:02:23,960 --> 00:02:27,760 In AI years, that's six years of normal human time 53 00:02:27,760 --> 00:02:30,040 passed in one day in AI. 54 00:02:30,040 --> 00:02:33,560 So we're speaking to you in March of 2023 55 00:02:33,560 --> 00:02:35,040 right now for episode 0. 56 00:02:35,040 --> 00:02:37,480 That was only back in November, am I right? 57 00:02:37,480 --> 00:02:40,560 Yeah, so November 30th, Chat GPT launched. 58 00:02:40,560 --> 00:02:43,920 And I think I saw somebody on LinkedIn, 59 00:02:43,920 --> 00:02:47,680 and the person was definitely not very tech savvy. 60 00:02:47,680 --> 00:02:49,080 And I was like, wait a second. 61 00:02:49,080 --> 00:02:53,280 If a non-tech savvy person is talking about this, 62 00:02:53,280 --> 00:02:54,240 this is something. 63 00:02:54,240 --> 00:02:56,200 Obviously, we've been looking into AI 64 00:02:56,200 --> 00:03:00,120 and different large language models and image models 65 00:03:00,120 --> 00:03:04,320 and all the craziness that was happening with generative AI. 66 00:03:04,320 --> 00:03:06,760 But I think with Chat GPT, what was interesting, 67 00:03:06,760 --> 00:03:10,560 the way they launched and the whole marketing and PR 68 00:03:10,560 --> 00:03:14,680 campaign that it was so user driven, 69 00:03:14,680 --> 00:03:16,440 very similar as Mid Journey. 70 00:03:16,440 --> 00:03:18,840 So yes, I'm curious to hear your experience, 71 00:03:18,840 --> 00:03:23,160 but I tried it, and I was like, yep, everything changes. 72 00:03:23,160 --> 00:03:23,840 Time changes. 73 00:03:23,840 --> 00:03:24,400 I'm sold. 74 00:03:24,400 --> 00:03:26,160 I still remember the first time I 75 00:03:26,160 --> 00:03:29,200 put in a prompt that was not much different than something 76 00:03:29,200 --> 00:03:31,280 I might dump into a Google search, 77 00:03:31,280 --> 00:03:34,240 because I had no other context for how to use these things. 78 00:03:34,240 --> 00:03:38,120 And then before your eyes, it still 79 00:03:38,120 --> 00:03:40,200 makes me feel a little excited. 80 00:03:40,200 --> 00:03:44,000 It's like witnessing magic, the only way I can describe it. 81 00:03:44,000 --> 00:03:45,120 I have a question to you. 82 00:03:45,120 --> 00:03:49,640 When you first time used, did you 83 00:03:49,640 --> 00:03:53,000 communicate with Chat GPT as it is a person? 84 00:03:53,000 --> 00:03:57,120 Like this human being like machine or a robot, 85 00:03:57,120 --> 00:03:59,160 like saying, hey, how do you feel? 86 00:03:59,160 --> 00:04:00,560 Or what's your name? 87 00:04:00,560 --> 00:04:01,880 Or hi, I'm Gada. 88 00:04:01,880 --> 00:04:03,440 Like, who are you? 89 00:04:03,440 --> 00:04:06,880 Or did you go with giving instructions? 90 00:04:06,880 --> 00:04:10,000 As I'm sure you'll come to find out, and our listeners as well, 91 00:04:10,000 --> 00:04:13,680 I am someone who wants to get all the juice out 92 00:04:13,680 --> 00:04:15,320 of something like this. 93 00:04:15,320 --> 00:04:19,080 So my mind just started working with, oh my gosh, 94 00:04:19,080 --> 00:04:21,320 this is way different than anything I've seen before. 95 00:04:21,320 --> 00:04:23,000 Let's, what can I use it for? 96 00:04:23,000 --> 00:04:26,160 So I started out with some zero shot prompts. 97 00:04:26,160 --> 00:04:30,040 That's basically just like a singular question of it. 98 00:04:30,040 --> 00:04:32,800 That might be a little more advanced than a Google search, 99 00:04:32,800 --> 00:04:35,720 but just the specificity of the responses, 100 00:04:35,720 --> 00:04:38,120 I was so taken aback by. 101 00:04:38,120 --> 00:04:42,280 But then soon thereafter, I learned about role prompting, 102 00:04:42,280 --> 00:04:45,720 which is where you can actually make the AI model assume 103 00:04:45,720 --> 00:04:47,400 a persona of sorts. 104 00:04:47,400 --> 00:04:50,240 You get a better result. It's just incredible. 105 00:04:50,240 --> 00:04:52,280 If I want it to make me a recipe, 106 00:04:52,280 --> 00:04:54,960 I tell it to be a chef first, and then ask me 107 00:04:54,960 --> 00:04:56,680 to make a specific recipe. 108 00:04:56,680 --> 00:04:59,400 And lo and behold, that's a better result than if I just 109 00:04:59,400 --> 00:05:01,520 ask it, hey, give me a recipe, because it 110 00:05:01,520 --> 00:05:03,920 has the context around that. 111 00:05:03,920 --> 00:05:06,640 So that was my first kind of delve into it. 112 00:05:06,640 --> 00:05:09,800 I think if I look back, some of my first prompts, which 113 00:05:09,800 --> 00:05:12,600 is only, I think I only really discovered it in January, 114 00:05:12,600 --> 00:05:15,080 but some of my first prompts were 115 00:05:15,080 --> 00:05:17,280 trying to use it for like planning a trick. 116 00:05:17,280 --> 00:05:21,360 And quickly discovered that, oh, this is way better than doing 117 00:05:21,360 --> 00:05:24,200 like 40 Google searches all at once. 118 00:05:24,200 --> 00:05:27,280 Funny that you say that because I was getting back 119 00:05:27,280 --> 00:05:31,960 into YouTube game, the first one I tried to, 120 00:05:31,960 --> 00:05:36,040 as I think any YouTuber did for ChachiPT to write YouTube 121 00:05:36,040 --> 00:05:39,200 titles, I was like, OK, I can give instructions, 122 00:05:39,200 --> 00:05:42,840 make it less than 50 characters, make it catchy. 123 00:05:42,840 --> 00:05:45,480 And then I so quickly realized that no, 124 00:05:45,480 --> 00:05:48,880 it's the moment you put that YouTube titles, it goes, 125 00:05:48,880 --> 00:05:52,360 oh, how to make money, guide, get rich. 126 00:05:52,360 --> 00:05:53,600 And I was like, no, no, no. 127 00:05:53,600 --> 00:05:55,160 So not flipping. 128 00:05:55,160 --> 00:05:57,600 Yeah, so I was like, wait a second. 129 00:05:57,600 --> 00:06:01,800 So this thing should be able to write way better things. 130 00:06:01,800 --> 00:06:05,960 This is just me who is not giving proper instructions. 131 00:06:05,960 --> 00:06:07,480 And actually, in the very beginning, 132 00:06:07,480 --> 00:06:10,600 I made a video about ethics of AI. 133 00:06:10,600 --> 00:06:13,680 And looking back, it is still an interesting video. 134 00:06:13,680 --> 00:06:16,080 But looking back, I just simply asked, 135 00:06:16,080 --> 00:06:18,640 oh, what do you think about ethics of AI? 136 00:06:18,640 --> 00:06:19,800 Or what should we do? 137 00:06:19,800 --> 00:06:21,280 What is the impact? 138 00:06:21,280 --> 00:06:24,720 Now with the knowledge I have right now, 139 00:06:24,720 --> 00:06:29,000 I know that I can completely influence those results. 140 00:06:29,000 --> 00:06:33,520 And now I can use jailbreaks and really even get ChachiPT 141 00:06:33,520 --> 00:06:35,360 to say some crazy things. 142 00:06:35,360 --> 00:06:37,480 Yeah, yeah. 143 00:06:37,480 --> 00:06:39,760 It is definitely a bit of a shift 144 00:06:39,760 --> 00:06:44,160 from using it compared to a depth-first search model 145 00:06:44,160 --> 00:06:47,680 or a page hit model like Google's algorithm is based on. 146 00:06:47,680 --> 00:06:51,200 But I feel like since you and I might be categorized 147 00:06:51,200 --> 00:06:53,520 as power users of this kind of tool right now, 148 00:06:53,520 --> 00:06:55,720 let's rewind the tape a little bit 149 00:06:55,720 --> 00:06:59,640 and talk about maybe what prompt engineering even is. 150 00:06:59,640 --> 00:07:02,120 From today's Wikipedia entry, prompt engineering 151 00:07:02,120 --> 00:07:04,040 is a concept in artificial intelligence, 152 00:07:04,040 --> 00:07:06,400 particularly natural language processing. 153 00:07:06,400 --> 00:07:09,200 In prompt engineering, the description of the task 154 00:07:09,200 --> 00:07:11,160 that the AI is supposed to accomplish 155 00:07:11,160 --> 00:07:14,000 is embedded in the input as a question 156 00:07:14,000 --> 00:07:16,080 instead of being implicitly given. 157 00:07:16,080 --> 00:07:18,360 Prompt engineering typically works by converting 158 00:07:18,360 --> 00:07:21,720 one or more tasks into prompt-based datasets 159 00:07:21,720 --> 00:07:23,400 and training the language model 160 00:07:23,400 --> 00:07:25,840 with what has been called prompt-based learning 161 00:07:25,840 --> 00:07:27,720 or just prompt learning. 162 00:07:27,720 --> 00:07:30,920 All right, so a lot of fancier terms that you've never heard. 163 00:07:30,920 --> 00:07:32,200 Allow me. 164 00:07:32,200 --> 00:07:33,800 Yes, please, translate. 165 00:07:33,800 --> 00:07:36,080 Allow me to translate this as, 166 00:07:36,080 --> 00:07:39,680 explain me like I'm five or any normal human being 167 00:07:39,680 --> 00:07:41,480 who is listening to this. 168 00:07:41,480 --> 00:07:46,280 Prompt engineering is an art and a science in itself 169 00:07:46,280 --> 00:07:51,280 of how to talk to AI or to large language models, if you may. 170 00:07:51,360 --> 00:07:55,160 And I really like to use some references 171 00:07:55,160 --> 00:07:58,600 that it's not what you say, it's how you say it. 172 00:07:58,600 --> 00:08:00,480 And I think if you think about it, 173 00:08:00,480 --> 00:08:03,040 if you have a conversation with a person 174 00:08:03,040 --> 00:08:07,040 and you want this person to do something what you want. 175 00:08:07,040 --> 00:08:10,920 So anybody who managed people, I think, in Relate 176 00:08:10,920 --> 00:08:15,280 or anybody who had personal assistant or executive assistant, 177 00:08:15,280 --> 00:08:19,720 the results you get is based on your instructions you provide. 178 00:08:19,720 --> 00:08:24,040 If you give no input, you get very sloppy results. 179 00:08:24,040 --> 00:08:26,720 And I experienced this as an architect 180 00:08:26,720 --> 00:08:28,160 working with the clients. 181 00:08:28,160 --> 00:08:32,920 And anybody who ever worked with graphical designers 182 00:08:32,920 --> 00:08:34,800 would say exactly the same. 183 00:08:34,800 --> 00:08:38,720 So coming with this background, this is my interpretation 184 00:08:38,720 --> 00:08:43,720 or you can say generalization of what prompt engineering is. 185 00:08:44,080 --> 00:08:47,480 I definitely agree with the art and science cross-section 186 00:08:47,480 --> 00:08:50,640 because it would almost seem like just in your example there 187 00:08:50,640 --> 00:08:54,160 that if you said, hey, I need to communicate 188 00:08:54,160 --> 00:08:56,960 some instructions to one of my coworkers 189 00:08:56,960 --> 00:09:00,080 or persuade them to do something, for example, 190 00:09:00,080 --> 00:09:02,840 with prompt engineering and using some of these models, 191 00:09:02,840 --> 00:09:04,800 you might want to say and think about, 192 00:09:04,800 --> 00:09:08,040 okay, what makes an argument persuasive? 193 00:09:08,040 --> 00:09:12,080 And using the resources, just searching on the internet 194 00:09:12,080 --> 00:09:15,000 or even asking the AI, hey, what makes a good persuasive 195 00:09:15,000 --> 00:09:19,160 argument and it tells you, then using that in the prompt, 196 00:09:19,160 --> 00:09:22,720 hey, I need to convince someone of my position 197 00:09:22,720 --> 00:09:26,640 based on this different website entry or this technique 198 00:09:26,640 --> 00:09:29,800 I want you to use the persuasive argumentative style 199 00:09:29,800 --> 00:09:32,480 and imagine that you're a master communicator 200 00:09:32,480 --> 00:09:33,320 and debater. 201 00:09:33,320 --> 00:09:35,960 So it's just, I think that the perfect cross-section 202 00:09:35,960 --> 00:09:39,200 of it is where, yes, thinking about things 203 00:09:39,200 --> 00:09:42,600 maybe a little algorithmically and as little components 204 00:09:42,600 --> 00:09:45,480 in an equation is definitely one way that's helpful 205 00:09:45,480 --> 00:09:49,840 but also having a completely outside the box perspective 206 00:09:49,840 --> 00:09:54,640 into how things are communicated or how a certain piece 207 00:09:54,640 --> 00:09:57,200 of art that if you want to generate something 208 00:09:57,200 --> 00:10:01,440 in mid-journey, how that style gets assembled 209 00:10:01,440 --> 00:10:04,440 is a, I think a really key and important piece. 210 00:10:04,440 --> 00:10:06,400 You know what's funny outcome of all, 211 00:10:06,400 --> 00:10:11,160 especially text to image, I noticed that a lot of people 212 00:10:11,160 --> 00:10:14,240 learn more about art and art styles and artists. 213 00:10:14,240 --> 00:10:18,040 And this is like one kind of benefit I noticed 214 00:10:18,040 --> 00:10:21,200 that people search actively about different artists 215 00:10:21,200 --> 00:10:22,920 in different times. 216 00:10:22,920 --> 00:10:26,280 As someone who studied art history back in academia, 217 00:10:26,280 --> 00:10:30,720 this is fascinating to see the same names and styles pop up 218 00:10:30,720 --> 00:10:33,080 and now more people actually know about it. 219 00:10:33,080 --> 00:10:34,720 I think that's quite cool. 220 00:10:34,720 --> 00:10:37,640 You pulled Wikipedia, which is cool, it's fine, 221 00:10:37,640 --> 00:10:40,160 but I want to pull Learn Prompting. 222 00:10:40,160 --> 00:10:42,720 Maybe I'm just still, I'm still stuck in the past 223 00:10:42,720 --> 00:10:43,560 a little bit. 224 00:10:43,560 --> 00:10:45,480 No, no, Wikipedia, it's a classic, right? 225 00:10:45,480 --> 00:10:48,600 But on learnprompting.org, prompt engineering 226 00:10:48,600 --> 00:10:51,680 is the process of communicating effectively 227 00:10:51,680 --> 00:10:54,920 with an AI to achieve desired results. 228 00:10:54,920 --> 00:10:56,000 There you go. 229 00:10:56,000 --> 00:10:56,840 Here you go. 230 00:10:56,840 --> 00:10:59,600 And of course there is way more you can go into 231 00:10:59,600 --> 00:11:04,360 and all the details, but I think this one sentence 232 00:11:04,360 --> 00:11:06,640 so nicely also wraps it. 233 00:11:06,640 --> 00:11:07,800 No, I completely agree. 234 00:11:07,800 --> 00:11:10,560 And you probably noticed that G and I may not have 235 00:11:10,560 --> 00:11:12,440 the exact same speaking intonation. 236 00:11:12,440 --> 00:11:15,080 We're in fact across the world from one another, 237 00:11:15,080 --> 00:11:18,000 but found each other just through that little discord 238 00:11:18,000 --> 00:11:20,440 right there, the learnprompting.org discord. 239 00:11:20,440 --> 00:11:21,920 I'm really glad you brought that up 240 00:11:21,920 --> 00:11:24,600 because what I think you'll find, 241 00:11:24,600 --> 00:11:26,800 especially with all of the different tools 242 00:11:26,800 --> 00:11:30,200 that Facebook and Google, Instagram leverage 243 00:11:30,200 --> 00:11:33,840 to put ads that are relevant to us in front of our faces. 244 00:11:33,840 --> 00:11:36,680 If you make one Google search for chat GPT, 245 00:11:36,680 --> 00:11:39,240 you might start getting advertisements 246 00:11:39,240 --> 00:11:43,120 for Jeff's top 3000 chat GPT prompts. 247 00:11:43,120 --> 00:11:44,960 You know, what I think is, 248 00:11:44,960 --> 00:11:48,680 you look at Jeff's top 3000 GPT prompt emails. 249 00:11:48,680 --> 00:11:50,720 You mean all these guides? 250 00:11:50,720 --> 00:11:54,480 All these guides, quick reference, quick references 251 00:11:54,480 --> 00:11:57,680 and some, hey, pay me $500 for that, 252 00:11:57,680 --> 00:12:00,000 how to talk to AI course. 253 00:12:00,000 --> 00:12:03,360 And then in two weeks, you pay me another 200. 254 00:12:03,360 --> 00:12:05,800 The previous course is outdated 255 00:12:05,800 --> 00:12:07,920 and now there is a new course. 256 00:12:07,920 --> 00:12:12,920 The thing with learnprompting.org, it's open source 257 00:12:12,920 --> 00:12:15,480 which means that it's live. 258 00:12:15,480 --> 00:12:19,320 So community and people on GitHub updating all the time. 259 00:12:19,320 --> 00:12:23,560 And Wes, I know you published an article there too, right? 260 00:12:23,560 --> 00:12:26,040 Yep, our articles and our write-ups in there 261 00:12:26,040 --> 00:12:28,320 is based on peer reviewed research. 262 00:12:28,320 --> 00:12:30,440 Now, I don't know about you, G. 263 00:12:30,440 --> 00:12:32,760 I can't imagine that the average bear 264 00:12:32,760 --> 00:12:35,520 has a hundred page academic papers 265 00:12:35,520 --> 00:12:39,080 on the latest how to talk to AI techniques 266 00:12:39,080 --> 00:12:41,320 and their experiments on their nightstand. 267 00:12:41,320 --> 00:12:42,920 And if they do, I'm sure they work 268 00:12:42,920 --> 00:12:44,040 to put them right to sleep. 269 00:12:44,040 --> 00:12:46,200 But the secret sauce of learnprompting 270 00:12:46,200 --> 00:12:49,280 is we will take those and then digest them 271 00:12:49,280 --> 00:12:54,120 much like a cow digest grass through the six or seven stomachs 272 00:12:54,120 --> 00:12:57,560 that they have and make it somewhat more digestible, 273 00:12:57,560 --> 00:12:59,200 somewhat more understandable. 274 00:12:59,200 --> 00:13:01,440 But then rather than coming out at cow dung in the end, 275 00:13:01,440 --> 00:13:04,600 oh, we're gonna take the time and effort and energy 276 00:13:04,600 --> 00:13:07,440 to try to make this as broadly understandable as possible 277 00:13:07,440 --> 00:13:10,600 because we're both very passionate about these tools 278 00:13:10,600 --> 00:13:13,240 and what prompting can do for you in your life. 279 00:13:13,240 --> 00:13:17,280 In that same vein, what has coming into prompt engineering 280 00:13:17,280 --> 00:13:20,080 kind of done for you and shaped a little bit 281 00:13:20,080 --> 00:13:21,960 of what you kind of wanna do in the future? 282 00:13:21,960 --> 00:13:25,520 So just to rewind, first time I came across 283 00:13:25,520 --> 00:13:28,880 learnprompting.org, this thing felt real. 284 00:13:28,880 --> 00:13:31,440 Like it felt research backed. 285 00:13:31,440 --> 00:13:34,760 And this is what I really wanted to also understand. 286 00:13:34,760 --> 00:13:36,560 Not just, hey, this is a cool prompt, 287 00:13:36,560 --> 00:13:37,760 use this prompt, that's it. 288 00:13:37,760 --> 00:13:41,680 But I wanted to understand how to make these good prompts. 289 00:13:41,680 --> 00:13:42,520 Right? 290 00:13:42,520 --> 00:13:46,640 And this was because I was going through career shift 291 00:13:46,640 --> 00:13:49,320 and I was about to look for jobs. 292 00:13:49,320 --> 00:13:51,960 And my thinking was that, okay, 293 00:13:51,960 --> 00:13:53,280 if I'm looking for a new job, 294 00:13:53,280 --> 00:13:57,480 I need to get some competitive edge or upscale. 295 00:13:57,480 --> 00:14:02,080 And I was learning about AI for like past years, right? 296 00:14:02,080 --> 00:14:04,960 And I was like, okay, I'm going to learn this 297 00:14:04,960 --> 00:14:08,840 and get involved in the community and contribute. 298 00:14:08,840 --> 00:14:13,120 And I think Discord, and I could be wrong, 299 00:14:13,120 --> 00:14:18,120 but I think when I joined Discord was like 350 people. 300 00:14:18,320 --> 00:14:23,320 Now today in 2023, what we have March. 301 00:14:23,400 --> 00:14:24,920 That was 30,000, I think. 302 00:14:24,920 --> 00:14:25,760 Yes. 303 00:14:25,760 --> 00:14:29,000 And it was like this small website 304 00:14:29,000 --> 00:14:33,200 that one brilliant person, should we say? 305 00:14:33,200 --> 00:14:36,200 Well, eventually we'll get Sander on here. 306 00:14:36,200 --> 00:14:38,840 We'll have the spiritual leader 307 00:14:38,840 --> 00:14:40,720 of the learnprompting movement. 308 00:14:40,720 --> 00:14:44,240 We'll interview him and talk to him about his motivations. 309 00:14:44,240 --> 00:14:46,800 But see, I think that's an interesting approach 310 00:14:46,800 --> 00:14:47,640 where you would take, 311 00:14:47,640 --> 00:14:49,760 hey, I'm going to learn everything about this. 312 00:14:49,760 --> 00:14:51,600 To me, I would go, I'm going to use this 313 00:14:51,600 --> 00:14:52,760 to write my cover letters. 314 00:14:52,760 --> 00:14:54,720 So I don't have to write a thousand cover letters 315 00:14:54,720 --> 00:14:56,000 as I apply for jobs. 316 00:14:56,000 --> 00:14:58,360 Whole topic, I have personal opinion 317 00:14:58,360 --> 00:15:02,400 about writing cover letters or resumes 318 00:15:02,400 --> 00:15:04,920 or using your personal data and dumping it 319 00:15:04,920 --> 00:15:06,440 in these type of tools. 320 00:15:06,440 --> 00:15:08,480 We can definitely explore that. 321 00:15:08,480 --> 00:15:10,000 But yeah, so learnprompting, 322 00:15:10,000 --> 00:15:12,080 I got involved with the community, 323 00:15:12,080 --> 00:15:15,320 started talking to Sanders and became contributor. 324 00:15:15,320 --> 00:15:19,080 And then I decided to make YouTube videos 325 00:15:19,080 --> 00:15:22,880 where I just share what I learned and what I'm doing. 326 00:15:22,880 --> 00:15:26,320 So just basically get people on a journey. 327 00:15:26,320 --> 00:15:29,320 And yeah, the rest is history. 328 00:15:29,320 --> 00:15:32,240 My channel started getting a couple of views 329 00:15:32,240 --> 00:15:35,240 and then we started talking 330 00:15:35,240 --> 00:15:39,240 and just basically everything became crazy. 331 00:15:39,240 --> 00:15:42,160 Now I'm scrapping time to keep learning, 332 00:15:42,160 --> 00:15:45,840 to keep updating myself and also creating. 333 00:15:45,840 --> 00:15:48,600 And I'm glad that we are doing this 334 00:15:48,600 --> 00:15:51,760 because I'm receiving so many questions from people. 335 00:15:51,760 --> 00:15:54,440 Just not scalable for me to answer it all. 336 00:15:54,440 --> 00:15:57,560 Plus putting two minds and getting some guests 337 00:15:57,560 --> 00:16:01,880 on this channel who are deep into these things as well. 338 00:16:01,880 --> 00:16:03,920 I think it's going to be a fantastic medium. 339 00:16:03,920 --> 00:16:07,280 Getting their perspectives just wonderfully beneficial, 340 00:16:07,280 --> 00:16:09,240 but let that not be a barrier. 341 00:16:09,240 --> 00:16:12,840 You don't need to be a computer science expert 342 00:16:12,840 --> 00:16:15,320 or have some sort of certification 343 00:16:15,320 --> 00:16:19,440 in natural language modeling or some fancy degree. 344 00:16:19,440 --> 00:16:21,360 You need to, from where I sit, 345 00:16:21,360 --> 00:16:24,880 you need two different things to be a prompt engineer 346 00:16:24,880 --> 00:16:26,280 and use these tools really well, 347 00:16:26,280 --> 00:16:28,240 even to the point where you can make a career 348 00:16:28,240 --> 00:16:29,520 or make some money from them. 349 00:16:29,520 --> 00:16:32,800 You need to have maybe probably done a Google search 350 00:16:32,800 --> 00:16:35,080 or send a text message or written an email 351 00:16:35,080 --> 00:16:38,400 because you will be writing into a little box 352 00:16:38,400 --> 00:16:40,680 on the internet that is going to send it 353 00:16:40,680 --> 00:16:43,040 to a large language model and give you a response back. 354 00:16:43,040 --> 00:16:46,960 And then two, you need to just be a person 355 00:16:46,960 --> 00:16:50,960 that's alive maybe and have had some different experiences, 356 00:16:50,960 --> 00:16:53,520 insights and viewpoints in the world. 357 00:16:53,520 --> 00:16:55,920 It's really just everyone's differences 358 00:16:55,920 --> 00:16:58,320 that they can bring forth into a process 359 00:16:58,320 --> 00:17:00,440 like prompt engineering that can give them 360 00:17:00,440 --> 00:17:02,160 a bit of a competitive edge. 361 00:17:02,160 --> 00:17:05,320 If you've been well traveled or well read 362 00:17:05,320 --> 00:17:07,680 or just have certain kind of Joe DeVive 363 00:17:07,680 --> 00:17:10,160 that can play into some really interesting 364 00:17:10,160 --> 00:17:12,640 and creative results thanks to prompt engineering. 365 00:17:12,640 --> 00:17:15,200 Funny thing, I don't know if you watched my video 366 00:17:15,200 --> 00:17:18,240 about Mid Journey, but Mid Journey's founder, 367 00:17:18,240 --> 00:17:21,360 he basically shared the interesting insight 368 00:17:21,360 --> 00:17:24,320 that when we'd been doing user research, 369 00:17:24,320 --> 00:17:26,200 and I'm here specifically Mid Journey, 370 00:17:26,200 --> 00:17:28,440 so this was before ChaiGBT, 371 00:17:28,440 --> 00:17:31,720 that what they found out that people who are older, 372 00:17:31,720 --> 00:17:34,880 that older people, they communicate way better 373 00:17:34,880 --> 00:17:36,040 what they want. 374 00:17:36,040 --> 00:17:39,160 They have, we have experiences, 375 00:17:39,160 --> 00:17:40,560 exactly what you were saying. 376 00:17:40,560 --> 00:17:42,960 We have experiences, they lived life, 377 00:17:42,960 --> 00:17:45,600 they have seen stuff and experienced stuff, 378 00:17:45,600 --> 00:17:48,480 so they can actually put words to it, 379 00:17:48,480 --> 00:17:51,680 just statistically probably read more books 380 00:17:51,680 --> 00:17:53,640 than younger person. 381 00:17:53,640 --> 00:17:57,520 So this is what's interesting that as for prompt engineering 382 00:17:57,520 --> 00:18:01,600 that actually your age, you have an advantage 383 00:18:01,600 --> 00:18:03,040 over a younger person. 384 00:18:03,880 --> 00:18:04,840 Definitely, definitely. 385 00:18:04,840 --> 00:18:08,960 That's just another data point in how this isn't a young, 386 00:18:08,960 --> 00:18:12,360 tech savvy tool that you need to really be 387 00:18:12,360 --> 00:18:13,960 on the cutting edge to use. 388 00:18:13,960 --> 00:18:15,400 It's gonna really form a lot. 389 00:18:15,400 --> 00:18:16,680 It's gonna transform anything, 390 00:18:16,680 --> 00:18:19,080 and everybody has something to bring to the table. 391 00:18:19,080 --> 00:18:21,760 So Wes, so I found learn prompting. 392 00:18:21,760 --> 00:18:23,760 How did you found it? 393 00:18:23,760 --> 00:18:27,760 And also what got you to this point? 394 00:18:27,760 --> 00:18:30,200 Like where is the interest coming from 395 00:18:30,200 --> 00:18:32,800 and how are you using all this knowledge 396 00:18:32,800 --> 00:18:35,600 and your involvement with learn prompting? 397 00:18:35,600 --> 00:18:38,880 And in general, I heard that you're one of the top sellers 398 00:18:38,880 --> 00:18:40,640 on prompt base, right? 399 00:18:40,640 --> 00:18:41,960 Well, you know, I'm up there, 400 00:18:41,960 --> 00:18:45,520 but I'm flirting around top 200 in the world, 401 00:18:45,520 --> 00:18:48,680 but that's just one little opportunity 402 00:18:48,680 --> 00:18:51,000 that people can take advantage of right now. 403 00:18:51,000 --> 00:18:52,920 You can put your compelling prompt 404 00:18:52,920 --> 00:18:55,480 that gives you a travel itinerary 405 00:18:55,480 --> 00:18:58,920 or something that produces a photorealistic bottle 406 00:18:58,920 --> 00:19:00,280 in mid journey. 407 00:19:00,280 --> 00:19:02,200 You can put those online on this website, 408 00:19:02,200 --> 00:19:03,440 promptbase.com. 409 00:19:03,440 --> 00:19:05,200 We'll put that in the description as well, 410 00:19:05,200 --> 00:19:09,040 that you can sell between two and $20, two and $30, 411 00:19:09,040 --> 00:19:11,520 or even offer custom prompts. 412 00:19:11,520 --> 00:19:14,240 It's a great little way to make some passive income. 413 00:19:14,240 --> 00:19:17,080 If you look at some of the top sellers on there right now, 414 00:19:17,080 --> 00:19:22,080 they're on pace to make 80 to $100,000 in passive income 415 00:19:22,200 --> 00:19:27,200 in selling 299, 399, 499 prompts. 416 00:19:27,200 --> 00:19:28,840 This is amazing. 417 00:19:28,840 --> 00:19:32,840 Not too expensive, but definitely speaks to the interest. 418 00:19:32,840 --> 00:19:34,840 This isn't just something that, you know, 419 00:19:34,840 --> 00:19:36,840 select group of people are gonna be interested in. 420 00:19:36,840 --> 00:19:39,360 It's gonna touch every aspect of everyone's life 421 00:19:39,360 --> 00:19:40,480 in some way. 422 00:19:40,480 --> 00:19:42,920 But you asked, you know, how I got into this. 423 00:19:42,920 --> 00:19:47,520 So I just finished a degree into operations research. 424 00:19:47,520 --> 00:19:49,360 So I got initially exposed 425 00:19:49,360 --> 00:19:51,640 to some machine learning principles, 426 00:19:51,640 --> 00:19:54,480 and a lot of the building blocks 427 00:19:54,480 --> 00:19:57,040 that go into making these large language models. 428 00:19:57,040 --> 00:19:59,840 It was definitely in over my head in those classes, 429 00:19:59,840 --> 00:20:02,160 but I recognized that, hey, this is a, 430 00:20:02,160 --> 00:20:04,320 this is kind of like the key to the future. 431 00:20:04,320 --> 00:20:07,720 I'm gonna try real hard to understand it a little bit, 432 00:20:07,720 --> 00:20:09,440 and then take it from there. 433 00:20:09,440 --> 00:20:13,040 But graduated, came and went on about my regular job, 434 00:20:13,040 --> 00:20:16,120 regular life, and then Chad GPT came out. 435 00:20:16,120 --> 00:20:18,280 And immediately, like I said, 436 00:20:18,280 --> 00:20:19,840 soon as I put that first prompt in there, 437 00:20:19,840 --> 00:20:22,040 I'm like, this is way different. 438 00:20:22,040 --> 00:20:24,320 I can only, the only way I can describe it is, 439 00:20:24,320 --> 00:20:26,480 is like, it must have been what it was like 440 00:20:26,480 --> 00:20:29,400 for someone to see an airplane flying in the sky 441 00:20:29,400 --> 00:20:33,120 for the first time, like back in the 1910s, 442 00:20:33,120 --> 00:20:33,960 or something like that. 443 00:20:33,960 --> 00:20:34,800 Or Google. 444 00:20:34,800 --> 00:20:36,120 Or Google search, yeah. 445 00:20:36,120 --> 00:20:39,080 It was that transformative to me, 446 00:20:39,080 --> 00:20:42,160 and I immediately just started using it, 447 00:20:42,160 --> 00:20:43,480 but I didn't have the context 448 00:20:43,480 --> 00:20:46,160 beyond what we call zero-shot prompting now that I know, 449 00:20:46,160 --> 00:20:47,840 but essentially the same kind of questions 450 00:20:47,840 --> 00:20:49,800 I would throw into a Google search. 451 00:20:49,800 --> 00:20:52,120 That's when I found, learned prompting, 452 00:20:52,120 --> 00:20:54,560 saw that it was open source, 453 00:20:54,560 --> 00:20:57,640 and I'm like, oh, I just finished two and a half years 454 00:20:57,640 --> 00:20:59,480 of reading papers just like this. 455 00:20:59,480 --> 00:21:01,880 I can maybe contribute something here. 456 00:21:01,880 --> 00:21:03,080 So that's what I did. 457 00:21:03,080 --> 00:21:06,080 And during that journey, and this is the thing that, 458 00:21:06,080 --> 00:21:08,960 if I never talk about this or never do anything else 459 00:21:08,960 --> 00:21:12,080 with prompting and prompt engineering and talking to AI, 460 00:21:12,080 --> 00:21:14,320 why it's so special and important to me, 461 00:21:14,320 --> 00:21:19,320 when you have a tool that can get you 75% of the way 462 00:21:19,960 --> 00:21:22,560 on anything, on writing an email, 463 00:21:22,560 --> 00:21:25,000 on coming up with new and fresh ideas 464 00:21:25,000 --> 00:21:28,360 for your next YouTube video title, 465 00:21:28,360 --> 00:21:30,560 on what you should eat for dinner tonight, 466 00:21:30,560 --> 00:21:32,960 on where you should go during your trip, you know, 467 00:21:32,960 --> 00:21:34,000 trip to Europe. 468 00:21:34,000 --> 00:21:35,880 When you have something that you know 469 00:21:35,880 --> 00:21:38,240 is gonna get you 75% of the way there, 470 00:21:38,240 --> 00:21:41,760 at least for me, who's someone that is kind of constantly 471 00:21:41,760 --> 00:21:43,840 thinking about how he's gonna communicate, 472 00:21:43,840 --> 00:21:45,280 how he's gonna express himself, 473 00:21:45,280 --> 00:21:47,760 that email I gotta write to my boss the next day. 474 00:21:47,760 --> 00:21:52,200 With that gone, it frees up so much just like emotional 475 00:21:52,200 --> 00:21:54,360 and intellectual like space in my life 476 00:21:54,360 --> 00:21:57,080 that's allowed me to be more present with my family 477 00:21:57,080 --> 00:21:58,960 and engaging with other people. 478 00:21:58,960 --> 00:22:02,000 So while I can't say that's gonna be everyone's reaction 479 00:22:02,000 --> 00:22:04,120 and response and experience when they do this, 480 00:22:04,120 --> 00:22:05,560 it was for me and that's what, 481 00:22:05,560 --> 00:22:08,080 and it's only just the beginning with so many of these things. 482 00:22:08,080 --> 00:22:10,080 So that's how I came into it and why I know 483 00:22:10,080 --> 00:22:12,360 it's gonna continue to be a big part of my life 484 00:22:12,360 --> 00:22:13,440 and my journey. 485 00:22:13,440 --> 00:22:14,720 This is amazing. 486 00:22:14,720 --> 00:22:16,240 Thank you for sharing this. 487 00:22:16,240 --> 00:22:17,520 No, absolutely. 488 00:22:17,520 --> 00:22:18,680 Cause it can be anybody. 489 00:22:18,680 --> 00:22:21,400 Everybody has something to offer in prop engineering. 490 00:22:21,400 --> 00:22:24,160 If you're a good communicator and written, 491 00:22:24,160 --> 00:22:26,880 auditory, in any form, something to offer. 492 00:22:26,880 --> 00:22:31,280 I think with that, we want this to become something great, 493 00:22:31,280 --> 00:22:33,840 something informative, something funny. 494 00:22:33,840 --> 00:22:34,680 I don't know about you, 495 00:22:34,680 --> 00:22:36,960 every time I click on a new AI video, 496 00:22:36,960 --> 00:22:41,080 I feel like it's all gloom and doom are way above my head. 497 00:22:41,080 --> 00:22:42,760 Is that been your experience? 498 00:22:42,760 --> 00:22:44,320 Not with my videos. 499 00:22:44,320 --> 00:22:46,000 I know, that's why I like it. 500 00:22:46,000 --> 00:22:48,160 And that's why we're together. 501 00:22:48,160 --> 00:22:52,680 No, this was actually my annoyance. 502 00:22:52,680 --> 00:22:55,320 You can say that for almost like a year, 503 00:22:55,320 --> 00:22:58,960 space on YouTube about AI turned into this, 504 00:22:58,960 --> 00:23:03,720 get rich quick, use chat GPT, you are superhuman AI. 505 00:23:03,720 --> 00:23:08,120 I was like, guys and girls, where is critical thinking? 506 00:23:08,120 --> 00:23:09,320 Where is the rational thinking? 507 00:23:09,320 --> 00:23:11,280 This is like life moves on. 508 00:23:11,280 --> 00:23:15,360 And for me, this danger of sensationalism of AI, 509 00:23:15,360 --> 00:23:18,280 because for most people who don't know what it is, 510 00:23:18,280 --> 00:23:21,200 and they just hear something like that, 511 00:23:21,200 --> 00:23:23,280 that either scares them, 512 00:23:23,280 --> 00:23:25,880 or it makes them feel that this is a fraud. 513 00:23:25,880 --> 00:23:28,440 And that's why I think having these conversations 514 00:23:28,440 --> 00:23:31,520 and having this podcast allows aid for us 515 00:23:31,520 --> 00:23:32,840 to have different perspectives 516 00:23:32,840 --> 00:23:35,200 that we come from different backgrounds 517 00:23:35,200 --> 00:23:38,840 and share more in-depth views 518 00:23:38,840 --> 00:23:42,720 would wish this podcast to be relevant. 519 00:23:42,720 --> 00:23:45,440 Because even researchers, AI researcher 520 00:23:45,440 --> 00:23:47,640 doesn't mean is a good prompt engineer. 521 00:23:47,640 --> 00:23:49,520 I really like what Sam Altman said, 522 00:23:49,520 --> 00:23:51,520 that he's not a good prompt engineer. 523 00:23:51,520 --> 00:23:54,360 Creates and the CEO of these tools. 524 00:23:54,360 --> 00:23:56,600 Exactly, because going into this, 525 00:23:56,600 --> 00:23:59,520 like you're doing your own research, 526 00:23:59,520 --> 00:24:02,240 learning about prompts, testing, experimenting, 527 00:24:02,240 --> 00:24:05,600 seeing what works, what not, like down to the word. 528 00:24:05,600 --> 00:24:07,120 And I think it's an important point 529 00:24:07,120 --> 00:24:08,800 you brought up right there where it's just, 530 00:24:08,800 --> 00:24:11,960 hey, no one is an expert in this right now. 531 00:24:11,960 --> 00:24:15,560 This is a term that's been used for six months. 532 00:24:15,560 --> 00:24:18,960 There are no objective measures of prompt engineers. 533 00:24:18,960 --> 00:24:21,400 So please don't think for a second 534 00:24:21,400 --> 00:24:24,520 that you're gonna have all this wisdom and insight 535 00:24:24,520 --> 00:24:29,280 by spending $199.99 on Jeff's top 3,000 bucks. 536 00:24:29,280 --> 00:24:32,320 Because I'm dollars to donuts would bet 537 00:24:32,320 --> 00:24:36,280 that he probably used Jack GPT to create most of us. 538 00:24:36,280 --> 00:24:38,680 But I recognize though that, hey, 539 00:24:38,680 --> 00:24:43,080 some people may not want to become really deeply involved 540 00:24:43,080 --> 00:24:44,720 prompt engineers, but recognize 541 00:24:44,720 --> 00:24:46,600 that there's a lot of value to be gained. 542 00:24:46,600 --> 00:24:49,880 We can at least hopefully during these podcasts 543 00:24:49,880 --> 00:24:52,960 offer some insights into how to find people 544 00:24:52,960 --> 00:24:56,080 that know how to talk to AI and leverage it well, 545 00:24:56,080 --> 00:25:00,160 or what questions you should be asking of your peers, 546 00:25:00,160 --> 00:25:04,120 your employees of the AI itself to get to a point 547 00:25:04,120 --> 00:25:06,760 where you can get it to do what you want from it. 548 00:25:06,760 --> 00:25:10,560 This is a really good point because that internet 549 00:25:10,560 --> 00:25:13,240 and especially social media, right? 550 00:25:13,240 --> 00:25:16,520 Everyone heard that marketing, digital marketing, 551 00:25:16,520 --> 00:25:18,920 if you want your products to be sold, 552 00:25:18,920 --> 00:25:21,440 you need to be an influencer and stuff like that. 553 00:25:21,440 --> 00:25:26,440 Mind you, not everybody needs to be a marketing expert. 554 00:25:26,520 --> 00:25:29,680 That's why there are specific group of people 555 00:25:29,680 --> 00:25:34,120 who will go in depth of how algorithms work, 556 00:25:34,120 --> 00:25:35,920 how to position your product. 557 00:25:35,920 --> 00:25:38,000 And it should be like that. 558 00:25:38,000 --> 00:25:40,760 I was always also saying to my artist friends 559 00:25:40,760 --> 00:25:44,760 who worry about business side, get a business partner. 560 00:25:44,760 --> 00:25:47,520 Your job is to be an artist. 561 00:25:47,520 --> 00:25:52,520 The moment artist cares and worries about how to earn money, 562 00:25:52,560 --> 00:25:53,760 that changes the game. 563 00:25:53,760 --> 00:25:57,160 So I think it's exactly the same that just because 564 00:25:57,160 --> 00:25:58,800 there is this tools and AI tools, 565 00:25:58,800 --> 00:26:01,960 not everybody needs to become an expert. 566 00:26:01,960 --> 00:26:05,800 If you want to, you definitely can go that path. 567 00:26:05,800 --> 00:26:10,800 But a lot of people, it just as using any other assistance, 568 00:26:10,840 --> 00:26:12,960 they will use prompt engineering consultants, 569 00:26:12,960 --> 00:26:14,760 they will be buying prompts. 570 00:26:14,760 --> 00:26:17,840 It's just how it has always been. 571 00:26:17,840 --> 00:26:20,400 Not everybody becomes an expert. 572 00:26:20,400 --> 00:26:23,440 Yeah, and I think it can be used too. 573 00:26:23,440 --> 00:26:25,880 It is not a strictly generative tool. 574 00:26:25,880 --> 00:26:28,400 Hey, I need you to write this email for me. 575 00:26:28,400 --> 00:26:30,800 I need you to work on this report for me. 576 00:26:30,800 --> 00:26:33,000 Plenty of ethics to unpack with that one 577 00:26:33,000 --> 00:26:36,120 as I'm sure this is gonna become a huge topic of discussion 578 00:26:36,120 --> 00:26:39,680 if it's not already on every high school and college campus 579 00:26:39,680 --> 00:26:41,920 considering the essay is now in jeopardy. 580 00:26:41,920 --> 00:26:42,760 It is. 581 00:26:42,760 --> 00:26:44,480 That's not to say that it should be something 582 00:26:44,480 --> 00:26:45,560 that should be shunned. 583 00:26:45,560 --> 00:26:49,200 We gave every student a calculator, however many years ago, 584 00:26:49,200 --> 00:26:52,280 and we didn't then stop doing math. 585 00:26:52,280 --> 00:26:54,200 We changed how we evaluated them. 586 00:26:54,200 --> 00:26:56,200 The same can be said for prompt engineering. 587 00:26:56,200 --> 00:26:59,320 I think it's a great exercise to go, yeah, use this tool, 588 00:26:59,320 --> 00:27:00,840 but then you're gonna put your prompt 589 00:27:00,840 --> 00:27:02,520 at the top of the essay. 590 00:27:02,520 --> 00:27:04,680 I don't want you to come back and tell me about 591 00:27:04,680 --> 00:27:05,880 how that worked out for you. 592 00:27:05,880 --> 00:27:08,760 I'm still gonna grade it like I would any old essay, 593 00:27:08,760 --> 00:27:10,520 and that in itself becomes 594 00:27:10,520 --> 00:27:13,000 a really good critical thinking exercise. 595 00:27:13,000 --> 00:27:16,120 Talking about critical thinking, one use case, 596 00:27:16,120 --> 00:27:18,120 which is what you shared, 597 00:27:18,120 --> 00:27:22,560 that for you it's freeing 75% of time 598 00:27:22,560 --> 00:27:26,400 and removing this just basically stress and anxiety 599 00:27:26,400 --> 00:27:27,760 when you have to write emails, 600 00:27:27,760 --> 00:27:30,400 when you have to basically overthink. 601 00:27:30,400 --> 00:27:33,440 The way I look at it is like literally I have 602 00:27:33,440 --> 00:27:37,200 a personal assistant who can be honest. 603 00:27:37,200 --> 00:27:40,520 Again, going back to my roots in academia 604 00:27:40,520 --> 00:27:43,440 and with architecture, you want to have a professor. 605 00:27:43,440 --> 00:27:47,480 The best professor is who never compliments your work. 606 00:27:47,480 --> 00:27:51,760 You bring 100 ideas and he will tell how every 607 00:27:51,760 --> 00:27:55,280 of those ideas could be better or what you can improve. 608 00:27:55,280 --> 00:27:58,000 So for me, I noticed that most of the times 609 00:27:58,000 --> 00:28:02,640 when I use Chai GPT for ideation or brainstorming 610 00:28:02,640 --> 00:28:07,640 or writing, I use it to criticize me. 611 00:28:07,960 --> 00:28:11,360 Well, criticize mode, which I made a video about, 612 00:28:11,360 --> 00:28:15,960 but I use it in so many instances to basically challenge 613 00:28:15,960 --> 00:28:19,840 my own bias or my own thinking, what I think is good. 614 00:28:19,840 --> 00:28:23,880 And that allows me push boundaries 615 00:28:23,880 --> 00:28:27,400 and then I can again run the same thing 616 00:28:27,400 --> 00:28:30,600 and push even more and even more. 617 00:28:30,600 --> 00:28:33,000 That's, it's almost just like a little stepping stone 618 00:28:33,000 --> 00:28:34,560 to help me get to the next level. 619 00:28:34,560 --> 00:28:36,520 And I think it can be that for a lot of people. 620 00:28:36,520 --> 00:28:40,080 I'm sure you've heard and it's definitely a point like, 621 00:28:40,080 --> 00:28:42,840 hey, this could be the end of creativity 622 00:28:42,840 --> 00:28:45,520 and content creative thinking as we know it. 623 00:28:45,520 --> 00:28:46,360 I disagree. 624 00:28:46,360 --> 00:28:47,480 I think it's just gonna, 625 00:28:47,480 --> 00:28:49,080 this is gonna take our creativity 626 00:28:49,080 --> 00:28:50,880 to heights we could never achieve 627 00:28:50,880 --> 00:28:54,440 because the stress of looking at a blank page 628 00:28:54,440 --> 00:28:58,840 or an empty canvas, it isn't there anymore. 629 00:28:58,840 --> 00:29:03,200 Like it's, okay, I know I have at least a jumping off point 630 00:29:03,200 --> 00:29:07,320 and I can get something and have it do that output again. 631 00:29:07,320 --> 00:29:09,360 And if I don't like it, or I could just say, 632 00:29:09,360 --> 00:29:10,760 all right, I can work with this. 633 00:29:10,760 --> 00:29:14,280 Let me just make some changes and go on to something 634 00:29:14,280 --> 00:29:16,440 that you never even would have even considered. 635 00:29:16,440 --> 00:29:19,240 Okay, so what can people expect from us? 636 00:29:19,240 --> 00:29:22,720 Well, we hope to offer two perspectives, 637 00:29:22,720 --> 00:29:25,640 hopefully unique ones in the sense that G and I 638 00:29:25,640 --> 00:29:27,440 just try to hold on as tight as we can 639 00:29:27,440 --> 00:29:29,800 to the breakneck pace of all things AI 640 00:29:29,800 --> 00:29:31,920 and prompt engineering that gets thrown at us. 641 00:29:31,920 --> 00:29:33,880 So we can offer that perspective in the sense 642 00:29:33,880 --> 00:29:35,480 that we're a little more in the center of it. 643 00:29:35,480 --> 00:29:37,560 It's all I seem to think about these days. 644 00:29:37,560 --> 00:29:39,840 So we're gonna offer those two perspectives 645 00:29:39,840 --> 00:29:40,960 that you can count on. 646 00:29:40,960 --> 00:29:42,840 We hope to bring some interesting people 647 00:29:42,840 --> 00:29:45,840 in the AI and prompt engineering world to bear 648 00:29:45,840 --> 00:29:49,320 everything from people that are researching and building 649 00:29:49,320 --> 00:29:52,320 and trying to push these models even more 650 00:29:52,320 --> 00:29:54,600 into the future and removing biases 651 00:29:54,600 --> 00:29:56,680 and refining them and making them better tools 652 00:29:56,680 --> 00:29:59,280 to people that are just using prompts 653 00:29:59,280 --> 00:30:00,920 in a very interesting way. 654 00:30:00,920 --> 00:30:03,160 That is another thing we hope to offer you guys. 655 00:30:03,160 --> 00:30:07,240 And to add on that, another aspect is that we would like 656 00:30:07,240 --> 00:30:10,960 to make this podcast accessible both to prompt engineers 657 00:30:10,960 --> 00:30:15,400 or aspiring people to go and explore deeper that field, 658 00:30:15,400 --> 00:30:18,120 but also business owners, everyday people 659 00:30:18,120 --> 00:30:22,560 who want to understand this and want to understand AI space, 660 00:30:22,560 --> 00:30:25,880 but also where prompt engineers, 661 00:30:25,880 --> 00:30:27,840 where these people play a role 662 00:30:27,840 --> 00:30:30,080 in their business going forward. 663 00:30:30,080 --> 00:30:33,320 So I received a lot of messages, 664 00:30:33,320 --> 00:30:35,640 especially from business owners, 665 00:30:35,640 --> 00:30:38,160 entrepreneurs, agency managers. 666 00:30:38,160 --> 00:30:39,880 They understand that this is coming. 667 00:30:39,880 --> 00:30:43,400 They understand that this is important, 668 00:30:43,400 --> 00:30:46,080 but the question is how do you actually use it? 669 00:30:46,080 --> 00:30:48,880 How do you implement it in everyday business? 670 00:30:48,880 --> 00:30:52,760 How do you place this prompt engineer in your small agency 671 00:30:52,760 --> 00:30:55,040 and what this person actually does? 672 00:30:55,040 --> 00:30:59,160 So there's so much to explore from both perspectives 673 00:30:59,160 --> 00:31:03,040 and bringing on this podcast, both prompt engineers 674 00:31:03,040 --> 00:31:05,200 and matching them together with people 675 00:31:05,200 --> 00:31:09,680 who could use such service in their day-to-day operations. 676 00:31:09,680 --> 00:31:13,800 And having those two groups of people explore ideas 677 00:31:13,800 --> 00:31:17,480 and have conversations, I think that would be really cool. 678 00:31:17,480 --> 00:31:21,680 And therefore I think publishing this video on YouTube 679 00:31:21,680 --> 00:31:23,600 is going to be cool to see comments. 680 00:31:23,600 --> 00:31:25,480 But we're gonna have some fun with this too. 681 00:31:25,480 --> 00:31:28,800 Like maybe we'll interview someone that's passed away 682 00:31:28,800 --> 00:31:32,400 hundreds of years ago with the help of ChatGPT 683 00:31:32,400 --> 00:31:34,720 and some of these text to voice tools 684 00:31:34,720 --> 00:31:35,840 that are coming online. 685 00:31:35,840 --> 00:31:38,960 A lot of our talk is gonna focus around ChatGPT, 686 00:31:38,960 --> 00:31:41,480 Mid Journey, a lot of these generative AI models, 687 00:31:41,480 --> 00:31:44,120 but we'll share some other tools and use cases, 688 00:31:44,120 --> 00:31:46,560 both in the sense of different websites 689 00:31:46,560 --> 00:31:48,600 that you can check out, take advantage of, 690 00:31:48,600 --> 00:31:51,480 or different things that'll make you a better prompt engineer, 691 00:31:51,480 --> 00:31:52,880 better user of these tools. 692 00:31:52,880 --> 00:31:57,320 So basically, watch out for the third co-host 693 00:31:57,320 --> 00:31:59,280 with quite synthetic voice. 694 00:31:59,280 --> 00:32:00,120 There you go. 695 00:32:00,120 --> 00:32:02,440 There may be, maybe some may feel 696 00:32:02,440 --> 00:32:03,920 we'll be out of a job like that. 697 00:32:03,920 --> 00:32:06,880 I say nay, it'll just make us better at what we do. 698 00:32:06,880 --> 00:32:10,400 So with that folks, 4G across the way there, 699 00:32:10,400 --> 00:32:12,240 go to go ourself across the ocean, 700 00:32:12,240 --> 00:32:16,000 West the synth mine, we're gonna say ciao and bye for now. 701 00:32:16,000 --> 00:32:18,360 Happy prompting everybody. 702 00:32:18,360 --> 00:32:19,840 Happy prompting everybody. 703 00:32:19,840 --> 00:32:20,680 Take care. 704 00:32:20,680 --> 00:32:42,520 Ciao.