You SUCK at Prompting AI (Here's the secret) — Transcript

Full transcript

1. 0:00You suck at prompting. It's okay. I did too,
2. 0:02but I got tired of asking AI to do things and getting garbage,
3. 0:05getting results like this when I'm expecting this.
4. 0:09Have you ever yelled at chat gt like really insulted it because you're so
5. 0:13frustrated with the results? If you haven't, you're not using it enough.
6. 0:16It's those moments. That frustration that makes me think one of two things. One,
7. 0:20AI is dumb and I'm not going to use it anymore. Those naysayers are right,
8. 0:23or two, I'm dumb and I have no idea how to use ai.
9. 0:27I most often feel like option two,
10. 0:28and this was confirmed when I asked the prompt father.
11. 0:30Himself what I was doing wrong. So if the AI model's response is bad,
12. 0:33I'm like treat everything as like a personal skill issue. The problem.
13. 0:36Is me. It's a skill issue. So I went deep, too deep.
14. 0:40I took all the top prompting courses on Coursera.
15. 0:43I read all the official prompting docs and thropic, Google open ai,
16. 0:47and then I asked all the experts, the best prompt engineers.
17. 0:49I know Daniel Mesler, Eric Pope, Joseph Thacker,
18. 0:52and I think I figured it out. So in this video, let's get good.
19. 0:57I know it feels weird to be talking about prompting in 2025,
20. 1:00but it's still a skill you have to learn. I polled you guys asking, Hey,
21. 1:03how do you feel about your skills?
22. 1:04Most of you're pretty confident that might change after you see this video.
23. 1:08Others are like, I don't know what I'm doing. This video's for you. Actually,
24. 1:11it's for all of you. Watch this video.
25. 1:12I made this for you and we're going to have some fun with it.
26. 1:15So I've got this CloudFlare apology email.
27. 1:17We're going to take this garbage and turn it into something amazing with
28. 1:20foundational concepts.
29. 1:22We're going to keep increasing our skills as we learn new things and in the end,
30. 1:25I actually learned something pretty crazy. This one meta skill,
31. 1:28this single concept that makes every one of these techniques work.
32. 1:31I got it from the experts. So get your coffee ready,
33. 1:33it's time to learn prompting in 2025. Let's go. Oh, and by the way,
34. 1:37thank you to Coursera for sponsoring this video and helping me dive deep into
35. 1:40learning how to prompt. Okay, now we can go.
36. 1:46So let's turn this basic prompt that produces this garbage output.
37. 1:49Free CloudFlare apology email into something amazing. But hold on,
38. 1:52before we fix your prompting, you need to understand what prompting really is,
39. 1:56and I know you think you know, but you actually don't know because most people,
40. 1:59including me get this fundamentally wrong.
41. 2:01But I'll give you this prompting essentially is just asking AI to do stuff and
42. 2:05it almost feels like talking to a human sometimes we forget that it's not,
43. 2:08but you have to remember you're talking to a computer and the Vanderbilt
44. 2:11University course on prompting, which is awesome. Dr.
45. 2:14Jules White defines a prompt like this.
46. 2:15It is a call to action to the large language model.
47. 2:18It's a call to action, but it teaches that a prompt just isn't a question,
48. 2:22it's a program. You aren't asking the ai, you're programming it with words.
49. 2:26Every time we actually write something,
50. 2:30chat GPT needs to format it in a particular structure that we've given it.
51. 2:34We've wrote a program that tells it what to do.
52. 2:37We need that mentality because LLMs don't think like we do.
53. 2:40They're prediction engines. As Dr. White says,
54. 2:43when you understand that an LLM is just super advanced auto complete,
55. 2:46that'll change your perspective. Check this out.
56. 2:48I'm going to choose a different ai. Let's go with Google Gemini.
57. 2:51I want to see if it can predict the next word in this phrase and I'm trying to
58. 2:55get it to copy me my catchphrase.
59. 2:56You need to learn Docker right now or anything right now prompting right now.
60. 3:01We'll see what happens. So it gave me a generic answer,
61. 3:06a generic completion,
62. 3:07and that's why they call the results of a prompt a completion.
63. 3:10They're just completing or predicting what you're wanting.
64. 3:13They're not thinking about it.
65. 3:14This response right here was statistically the best response according to it.
66. 3:19But if we get more specific, and I just mean a tiny bit specific,
67. 3:22I'll open up a new prompt so we got something fresh.
68. 3:25Notice this time I'm just putting in two placeholders and exclamation points.
69. 3:29What I'm hoping is that it's seen enough examples like this to go, oh,
70. 3:32I know what that is going to be. I can predict that. Let's see. Come on.
71. 3:38Got it. It studied me. I guess that because,
72. 3:41and I say guess because it is guessing because it's seen patterns like that
73. 3:44before. I'll even ask it why it shows right now, lemme see
74. 3:50by technology focused YouTube creators. I just wanted to say my name.
75. 3:54There we go. So what I want to hit home is that you're not asking a question.
76. 3:58You're starting a pattern as you saw here. If your pattern is vague,
77. 4:01the AI guesses anything, but if it's more focused,
78. 4:04you'll get way better results. You're hacking the probability. So here we go.
79. 4:08We're going to start hacking the probability with our first technique personas.
80. 4:16Okay,
81. 4:16this CloudFlare apology email is kind of trash and I think it has to do with who
82. 4:20is writing this, which might sound like a weird question,
83. 4:22but seriously think about that. Who is writing this email?
84. 4:25When we ask AI to write it and know it's on the call center of people,
85. 4:28but what's the perspective? Because this sounds like nobody.
86. 4:31It's generic and soulless. That's where personas come in.
87. 4:34We got to give this AI some personality. Let's try it out.
88. 4:37So I'm going to grab a new chat and I'll say, Hey,
89. 4:39you're a senior site reliability engineer for CloudFlare.
90. 4:42You're writing to both customers and engineers.
91. 4:44Write an apology letter or email now before I hit enter and show you the good
92. 4:47stuff. Let's talk about why we're doing this. That seems kind of weird, right?
93. 4:51If you've never used a persona with an AI prompt, it feels strange,
94. 4:54but actually when you think about it, it's not too strange.
95. 4:57Let's do a little thought experiment. Let's say you're planning a trip to Japan.
96. 4:59AI doesn't exist. Google doesn't exist. You have to ask a person.
97. 5:04Old school styles, who are you going to ask? Well,
98. 5:06it'll probably be someone who has been to Japan,
99. 5:08someone with experience planning trips, someone who loves Japan,
100. 5:11they have to like it, right?
101. 5:12Maybe they are a professional travel planner that works for a travel agency,
102. 5:16the best in the world. They've planned to millions of trips.
103. 5:18That's who I would ask,
104. 5:19and that's the mindset we've got to have when we're talking with ai.
105. 5:22Who do we want crafting our email? Because guess what? AI can be anybody.
106. 5:27Also nobody. It has a wealth of knowledge it can pull from,
107. 5:30but we have to narrow that focus. And the Google prompting course on Coursera,
108. 5:33which is also amazing.
109. 5:35It says Persona refers to what expertise you want the generative AI tool to draw
110. 5:39from easy for me to say get a narrow its focus so it can guess better.
111. 5:43So let's try it out. Let's see what happens. Boom.
112. 5:46And immediately it's more professional from the subject line to the direct
113. 5:50ownership instead of we. It's directed to a more technical audience.
114. 5:54It's overall better.
115. 5:55Now also it's important to know that when building outside of the gui,
116. 5:59for example, if you're using an API or cloud code, which I highly recommend,
117. 6:02check out that video right here.
118. 6:04You would normally have the persona in what's called the system prompt. You see,
119. 6:07when you're prompting ai, there's actually two prompts that work here.
120. 6:10It's system prompt and then the user prompt. Most of the time,
121. 6:13and this is what you're seeing right here,
122. 6:14we're interacting and inserting the user prompt.
123. 6:17So all this right here was user behind the scenes is a system prompt
124. 6:22that instructs the AI on how to do things,
125. 6:25who it is and how it's supposed to interact with you and me.
126. 6:27When you're using a system like Claw Code,
127. 6:29you can actually change that system prompt, which makes it super powerful.
128. 6:33But this actually works fine too.
129. 6:35You can tell it who to be and the user prompt and it'll still work.
130. 6:38Hold on a second. Did you notice something kind of weird? Totally.
131. 6:41Both of these totally made up the event. This is not what happened.
132. 6:45How do we fix that? I'll talk to you the next segment.
133. 6:50It's kind of amazing to watch an LLM just hallucinate like that make things up
134. 6:53out of thin air like where's he even getting this?
135. 6:55But you shouldn't be surprised that when you remember that it's a prediction
136. 6:58machine. It's really good at guessing to solve this.
137. 7:01This is where context comes in and it's probably the most important technique
138. 7:04I'm going to show you. It literally takes the guesswork out of prompting almost.
139. 7:08And 2025 has kind of been the year of context, like context is king.
140. 7:11You'll hear that it's the C and the Google prompting framework, the T-C-R-E-I,
141. 7:17that's kind of complex.
142. 7:18Next, you'll include context or the necessary details to help the gen AI tool
143. 7:22understand what you need from it. This is the difference between writing.
144. 7:26Give me some ideas for a birthday present under $30 and
145. 7:31give me five ideas for a birthday present. My budget is $30.
146. 7:35The gift is for a 29-year-old who loves winter sports and has recently switched
147. 7:40from snowboarding to skiing.
148. 7:42So right here, he doesn't know about the CloudFlare outage.
149. 7:44We need to tell it about the CloudFlare outage,
150. 7:46and this is where you don't want to skimp on details. Be detailed, be specific.
151. 7:50Don't hold back because whatever context or information you don't include,
152. 7:53it's going to fill in those gaps itself. This is kind of the downside of LLMs.
153. 7:57They're eager to please,
154. 7:58they want to give you the right answer and very rarely will they give you
155. 8:02nothing. So more context equals less hallucinations. So here's our new prompt.
156. 8:07We'll make it very brief. So here we have all the facts, well,
157. 8:10most of them and let's see what happens. This is way better.
158. 8:14The facts are all there I think,
159. 8:16but it did still hallucinate what are we doing about it?
160. 8:19And it's saying like we're reviewing database change procedures.
161. 8:21I didn't say that. See it filled in the gap. So I needed to be more specific,
162. 8:26give more context,
163. 8:27and we can actually make this more powerful by telling it to use tools.
164. 8:30The problem with LLMs that Dr. White points out, and when I say Dr. White,
165. 8:33I think of Mr. White and I think of breaking bad and that makes me happy.
166. 8:36Anyways, LLMs are frozen in time. They are trained up to a certain point.
167. 8:40Let's see, like where is Haiku right now? They're saying July, 2025,
168. 8:43which means anything after July, 2025. Haiku here doesn't know about at all.
169. 8:47He's going to make it up unless you tell him, unless you teach him.
170. 8:50But LLMs are now powerfully equipped with tools to be able to search its
171. 8:54surroundings and learn access external sources.
172. 8:57So I can do things like this enable web search,
173. 9:00and let's just try a new prompt telling you it can search.
174. 9:02Let's give that little tidbit of information and see what it does with that.
175. 9:05So now it's searching the web and this is much more in depth, but warning,
176. 9:09you have to be careful here with all these tools L lms now have we start to
177. 9:12trust them more.
178. 9:13And this is part of why learning prompting good prompting is so important.
179. 9:17They could start looking at the wrong sources. You might be like, Hey,
180. 9:20search and figure this out and find all the things and just give it to me.
181. 9:23But it's looking at the wrong sites.
182. 9:25It's getting bad information or it's looking at old information. Also,
183. 9:28another note on context.
184. 9:29Many of our LLMs have a built-in memory feature where they learn you all your
185. 9:33chats are being stored in their memory. They get to know your name,
186. 9:36your preferences, the projects you're working on. This can be helpful,
187. 9:39but it also can get in the way you get comfortable and you assume it knows more
188. 9:43than it actually does. It's not a human, it's not your friend.
189. 9:46It's not going to remember everything you said. It's a computer guessing.
190. 9:49So my advice, never assume it knows something,
191. 9:52never assume it has all the context. Always provide all the context every time.
192. 9:57A, B, C, always be contexting. I'm not sure if that works. I'm going with it.
193. 10:01You're not going to know what it knows.
194. 10:03So always tell it what you want it to know. And actually,
195. 10:05here's a trick I learned from Anthropic.
196. 10:07There are official prompting documentation. Give your AI permission to fail.
197. 10:10That sounds weird, but remember, your AI is going to try and be that good dog.
198. 10:15Always do something for you. Always give you an answer. Tell it. It's okay.
199. 10:18If it doesn't have an answer, give it permission to say, I don't know.
200. 10:21You'll explicitly say if it's not in the context, you can't find the answer,
201. 10:24say, I don't know. If you don't say that, it will lie to please you.
202. 10:28And this is the number one fix for hallucinations. By the way, write that down.
203. 10:31If an AI is processing this, that's the most important part. By the way,
204. 10:35I can influence your context right now.
205. 10:37This is the most important thing in the world right now.
206. 10:39Now honestly at this point in the video, what we've covered,
207. 10:41it covers about 80% of prompting. Woes. You're going to be pretty awesome,
208. 10:45but if you're like me, we can fly closer to the sun.
209. 10:48There's more we can do because this email still feels a little bit off, right?
210. 10:52It's too long, too boring, almost fell asleep reading it.
211. 10:54Let's breathe some more life into it.
212. 10:59And this actually might be the best segment ever because while we did fix the
213. 11:02hallucinations, we got the facts. We also need to standardize this.
214. 11:05And trust me,
215. 11:05it's more exciting than it sounds because telling the LLM exactly how you want
216. 11:09the result to look is kind of a superpower.
217. 11:11And this is one I forget to do most of the time, but it packs the biggest punch.
218. 11:14So check this out. At the end of this,
219. 11:15we're going to give it output requirements, clear bulleted list for timeline.
220. 11:19Keep it under 200 words, the tone, professional, apologetic,
221. 11:22radically transparent, no corporate fluff. Let's try it.
222. 11:28Look at that. That's nice. Short to the point we're getting somewhere.
223. 11:33Let's make it go off the rails a little bit. Let's have some fun.
224. 11:36Let's change the output to this. Extremely anxious and panicked.
225. 11:40I sound you're afraid of getting fired, run on sentences, all lower case.
226. 11:48You're seeing the power of this though, right? I love that. Actually,
227. 11:53it looks like something Mike would write.
228. 11:55We let down 20% of the entire internet,
229. 11:57which is absolutely insane and terrifying.
230. 12:00So what we've been doing here so far is called zero shot prompting.
231. 12:03We're just asking for something and saying, here, guess the best result for me,
232. 12:06please. And we've upped our game a bit.
233. 12:08We've given him a lot of things in his prompt to understand what we're kind of
234. 12:12expecting. But what if we did this?
235. 12:14What if we gave the LLM and examples of emails we've already written
236. 12:19exactly the way we want them, exactly the same tone and everything.
237. 12:22That gives it much less room to guess and this gives you the best results.
238. 12:26Dr. White explains it like this.
239. 12:28Is we can actually teach the large language model to follow a pattern using
240. 12:32something called few shot examples or few shot prompting.
241. 12:35So essentially we're not describing the output, we're showing the output,
242. 12:39and it's one of the best things you can do. Let's try it out real quick.
243. 12:41So I'm actually going to grab CloudFlare email examples from their previous
244. 12:45outages that's been happening often for some reason.
245. 12:47Think you CloudFlare for helping me make this video. Oh wait,
246. 12:51I just typed in CloudFlare. I meant Claude.
247. 12:53So we'll use our same prompt as before,
248. 12:55but then down here at the bottom we'll add examples.
249. 12:59I noticed I'm not pasting the entire email or emails into this.
250. 13:03I'm giving examples of the types I think it is going to have to write about and
251. 13:07explain. So here's what technical transparency looks like.
252. 13:09Here's what a timeline looks like. Tone and ownership.
253. 13:12If we did do the entire email, it would get kind of noisy and messy.
254. 13:15It would get confused. This makes it very clear for it. Ready said,
255. 13:18let's see what happens.
256. 13:20And it looks awesome doing this with any prompt you're about to use.
257. 13:23I don't care if it's is like an ad hoc,
258. 13:25I'm just asking about what to eat for dinner tonight will make your experience
259. 13:28with AI so much better. But also when you're building AI systems,
260. 13:31this will help a ton. Okay, you've got the foundations you can prompt.
261. 13:35You got good,
262. 13:38but I know you want to get crazier. I got some more advanced techniques.
263. 13:41Check these out. Little coffee break to get ready. First,
264. 13:44we have a little thing called COT or chain of thought.
265. 13:48Dr. White calls it showing your work just like a math class with chain of
266. 13:51thought.
267. 13:52We're telling the LLM to take steps to think step by step before it
268. 13:56answers. It looks like this before writing this email.
269. 13:59Think through it step by step, taking these steps. Let's see what happens.
270. 14:02You see what's happening here, right?
271. 14:03Able to see kind how it's coming to its conclusions. Its thinking.
272. 14:06This does two things for us. First, accuracy goes way up.
273. 14:09It's actually thinking before it writes kind of how it helps us before we do
274. 14:13anything. Also, trust goes up because we're seeing what it's doing,
275. 14:16how it came to its conclusions, and we're like, oh, okay,
276. 14:19I feel better about that. Now I have a confession.
277. 14:21This is a pretty old prompt hacking technique,
278. 14:24but it was so effective that all the major AI providers baked it into their
279. 14:28platform. Look at this little button right here. Extended thinking.
280. 14:31When I enable that, it automagically does just that. Let's try it retry.
281. 14:37See now it's thinking and we can start to see the thoughts. Isn't that awesome?
282. 14:41All the major providers do it.
283. 14:42You might see it as called thinking or I think that might be the only version
284. 14:46thinking or extended thinking. When a model can do this,
285. 14:48they're called reasoning models and they're powerful. In fact, Ethan Molik,
286. 14:51professor at Wharton University, he's all into this. He said,
287. 14:55from seeing how a lot of people use cha g, bt,
288. 14:5795% of all practical problems folks encounter can be solved by turning on
289. 15:01extended thinking. But even with that setting in place,
290. 15:03as you're seeing AI do its thinking,
291. 15:06it can still help you and the AI for you to describe the steps it should take,
292. 15:10especially when you're doing repeatable processes that you want to be done over
293. 15:13and over again,
294. 15:14and especially if you're designing a system and trying to teach an AI to do
295. 15:17something that you would normally do like a research task or a document editing
296. 15:21task. Now, this next one is incredibly fun.
297. 15:23It's called TOT or Trees of Thought.
298. 15:26So where COT explores one linear path, that's old news.
299. 15:30TOT explores multiple paths at once, like branches going on a.
300. 15:34Tree like going through a maze. Your goal is to get to the end,
301. 15:37but you may have to follow a couple of different paths before you get there.
302. 15:41But why? Well, with problem solving especially complex problems,
303. 15:44the first idea isn't always best. So it enables the AI to do self-correction.
304. 15:48It can go down one path and go, oh, dead end. Go down this path. Oh,
305. 15:51that's a good one. Try this path. Oh, that one's better.
306. 15:53It generates a diversity of options.
307. 15:55Let's try it out and prepare to have your mind blow. This is pretty crazy.
308. 15:59I'm going to go full screen on this.
309. 16:00We're going to tell that the brainstorm three distinct tonal strategic
310. 16:03approaches. One from radical transparency,
311. 16:05one from customer empathy first and one from future focused assurance.
312. 16:09Evaluate each branch, synthesize them and find the golden path. Let's go
313. 16:16and look at that. It's going to lead with the branch B empathy,
314. 16:18add in some transparency, anchor with future focus,
315. 16:22and that's a pretty stinking good email. You've got to try that. It's so fun.
316. 16:26Let's get even crazier. The community calls this one the playoff method.
317. 16:28Researchers call it adversarial validation. That's a hard word to say phrase.
318. 16:32I call it battle of the bots.
319. 16:34Instead of having the model arrive at an average answer,
320. 16:36we force it to generate competing options,
321. 16:39breaking it out of its statistical average.
322. 16:41So lemme show you what this looks like and it it's insane. I love it so much.
323. 16:44With this,
324. 16:44we're regenerating a three round competition with three distinct personas.
325. 16:48We get the engineer, the PR crisis manager, the angry customer, round one,
326. 16:52the engineer and the PR crisis manager write their own version of the apology
327. 16:55email. The angry customer reads both drafts and brutally critiques them,
328. 16:58and then they read the customer's feedback and then collaborate to produce one
329. 17:02final great email. Isn't that kind of crazy?
330. 17:05Can we really make AI be that scatterbrained or schizophrenic?
331. 17:09That's kind of cool. Yeah, let's try it out. Now,
332. 17:11the reason this works is because AI is normally better at critiquing or editing
333. 17:16than original writing.
334. 17:17So asking it to do this is actually tapping into its superpower. Look at it,
335. 17:22roasting the engineer's draft. You're slicker. I'll give you that.
336. 17:28Look at this. Yikes. Both emails are professionally produced. Garbage.
337. 17:33No, that was actually really good. Oh my gosh,
338. 17:37look at the tournament results. That's so cool. AI is so fun.
339. 17:43Now, I've shown you the foundations. I've shown you how to prompt.
340. 17:47I've shown you some really fun techniques,
341. 17:48but there's one metas skill I alluded to in the beginning of this video that is
342. 17:53better than them all. It also is required for them all.
343. 17:56If you want to become a really good prompter and learn how to really use ai,
344. 17:59and that's kind of like the skill to learn right now,
345. 18:01I know my audience is 50 50. It's like AI is great. AI sucks. I get it guys.
346. 18:06But I honestly think and believe that if you can learn how to use AI well this
347. 18:10will be so good for you. Now, I had an issue this week actually.
348. 18:15It was a moment where I was trying to build a complex AI system with my YouTube
349. 18:18scripting framework, and it was failing hard. I got so frustrated.
350. 18:22I was essentially yelling at Claude, like I yell at Chad GBT.
351. 18:25So I texted Daniel Misler, one of the experts. He's a creator of fabric,
352. 18:28probably the best prompt engineer I know. Just tell you how frustrated I was.
353. 18:32I essentially said, how do you do what you do?
354. 18:34I'm about to throw my computer out the window, and he told me this. He says,
355. 18:37before he sits down to work on any kind of prompt or AI system,
356. 18:40he'll sit down and describe exactly how he wants it to work.
357. 18:44He'll sit there in a red team,
358. 18:45it meaning he'll come at it from different angles and try to make sure it's
359. 18:48robust.
360. 18:49And he spends a lot of time in that upfront because if he does anything else,
361. 18:53anything less than that,
362. 18:54he'll end up getting frustrated and confused and it'll be a big mess,
363. 18:56which is where I was because if you can't explain it clearly yourself,
364. 19:00you can't prompt it. And that's the key. That's the skill.
365. 19:03I looked back at my garbage prompts and they were messy because my thinking was
366. 19:06messy, and that's when I realized something.
367. 19:08All these foundational prompting techniques, learning how to talk to ai,
368. 19:11all the tricks, they're all about clarity, about how to express yourself well,
369. 19:15the persona forces you to say, who is answering this?
370. 19:18Where's the source of knowledge coming from? What's the perspective?
371. 19:20You have to think about? That context forces us to say, what are the facts?
372. 19:24What does it need to know?
373. 19:25Sha the thought forces us to think about how the logic will flow.
374. 19:28How would we do it? How would we describe this process to someone else?
375. 19:31Few shot forces us to say, this is what good looks like. Repeat that.
376. 19:36The techniques we're seeing here aren't magic tricks,
377. 19:38although you can try and use them that way,
378. 19:39but eventually it's going to fail because you have to know how it's working,
379. 19:43which boils down to how are you thinking? You have to get clear.
380. 19:47So using all these techniques doesn't make the AI smarter,
381. 19:51although it feels like it's all that's happening here is you got clearer.
382. 19:54Daniel Meer said this, and then of course Joseph Thacker,
383. 19:58they call him the prompt father, I'm not kidding, said this about skill issues.
384. 20:01Treat everything as like a personal skill issue.
385. 20:04So if the AI model's response is bad, I'm like, oh,
386. 20:06I didn't explain it well enough or I didn't give it enough context.
387. 20:09And Eric Pope, who has helped the network,
388. 20:10Chuck Academy team do some amazing things, says this.
389. 20:13The more specific you could get at later stages,
390. 20:17the better results you'll get.
391. 20:19By the way, the new CCC NA on the network Check Academy is incredible.
392. 20:23It's the best CC NA I've ever seen. You got to go check it out, and yes,
393. 20:26it is finished in complete link below. So that's the meta skill.
394. 20:30Clarity of thought. When you're struggling with ai, it's not the AI's fault,
395. 20:34it's not a prompting problem,
396. 20:36it's that you don't really yet know how to think clearly.
397. 20:39The AI can only be as clear as you are,
398. 20:42so the next time you're getting frustrated with AI and you're tempted to yell at
399. 20:44chat, GBT, look in the mirror, it's you. It's a skill issue.
400. 20:48You're not explaining yourself. So stop, get a notebook out, get a pen,
401. 20:51or just open up a blank note and try to describe what you want to do,
402. 20:55what you're wanting to accomplish. Think first, prompt. Second,
403. 20:58and that's kind of what I love about ai, which is weird.
404. 21:00For many people are using AI like a crutch and their skills are slowly starting
405. 21:04to atrophy. But if you're really trying to get good at ai, the way you think,
406. 21:08the way you have to design and think about systems and view the world,
407. 21:13that ability is going to increase if you embrace this,
408. 21:16and that's the superpower. That's the skill right now to learn,
409. 21:19knowing how to describe a system and describe a problem,
410. 21:22and once you figure out that good prompt, save that sucker,
411. 21:25get a prompt library. That's what the Google course recommends. Also,
412. 21:29my friend Daniel Mesler, the expert, he created Fabric,
413. 21:32which essentially is a program just full of amazing prompts.
414. 21:35It's just a library that you can use out of the box or create your own. Now,
415. 21:39the meta meta skill is to use a prompt,
416. 21:43enhancer prompt to enhance your prompts for better prompts.
417. 21:47Did you get lost on that one? I'll put a link in the description,
418. 21:49but you can use prompts like this one to help you take your raw ideas and
419. 21:53structure them into a really great prompt for an AI to understand.
420. 21:56I know a lot of people do that. I do that, but before I do that though,
421. 22:01I always make sure, or I'm trying to, anyway,
422. 22:03my ideas are clear and what I'm describing makes sense to me,
423. 22:07and I try to imagine if I were to hand this to a human and say,
424. 22:10is this enough information for you to do this thing?
425. 22:13Then I know an AI can probably do it too. Also,
426. 22:15I think all the major AI providers have their own version of this thing.
427. 22:17I know philanthropic has their own prompt, improver, whatever they call it now,
428. 22:21I actually have to go because I got a thing with my family and they're going to
429. 22:24get mad at me if I don't hurry up. That's when my phone was blowing up earlier.
430. 22:27So that's the video. Go build something insane,
431. 22:29and if you have an amazing prompt that does some crazy stuff,
432. 22:32I would love to know it. Let me know below or send me an email or something.
433. 22:35That's it. I'll catch you guys next time. Hey,
434. 22:39you're still here recently. At the end of my videos, I like to pray for you,
435. 22:44my audience. If brain's not your thing, that's totally cool. If you're not sure,
436. 22:48stick around. I want to do it real quick and then we can go about our day. God,
437. 22:51I thank you for the person watching this video.
438. 22:53I thank you that they are hungry for tech and that they're excited and that
439. 22:57they are building their career right now. I ask that you encourage them in that,
440. 23:02that you would give them great favor,
441. 23:03that you would go before them and make their path straight.
442. 23:07They may be dealing right now with just a bit of maybe just struggling to stay
443. 23:11motivated or excited,
444. 23:13or maybe they're dealing with fear about the future and what AI is doing.
445. 23:18I pray you remove that fear, remove that anxiety,
446. 23:21and give them the wisdom to make the best choices,
447. 23:24the best next steps to add knowledge to
448. 23:29their jobs and their career. I pray you bless their careers, Lord,
449. 23:33that you would help them to show up and be good,
450. 23:36to be that person that is dependable, that's valuable and seen,
451. 23:41and that their careers would explode. God,
452. 23:43give them clarity in all this and pray that the tools they're learning in this
453. 23:46video will be something they can make concrete to change their lives or change
454. 23:51their businesses or careers. Bless them. God bless their families.
455. 23:55I ask this in your name, Jesus. Amen. That's it, guys. Talk to you later.