5 Hacks To Use ChatGPT So Well It’s Almost Unfair — Transcript

Full transcript

1. 0:00Most people treat Chat GPT like a new
2. 0:02intern, but I use it to try five
3. 0:05strategies. It's like having five of the
4. 0:08world's smartest minds work 24/7 for
5. 0:12only $20 a month to help me move faster,
6. 0:16decide better, and focus on what truly
7. 0:18matters. But, I only figured this out
8. 0:21after spending years in tech and AI as a
9. 0:24CEO, board member, and investor,
10. 0:26>> [music]
11. 0:26>> where I watched elite teams use AI in
12. 0:29ways most people never see. So, today
13. 0:32I'm sharing those five strategies to use
14. 0:35Chat GPT so well, it feels illegal. The
15. 0:38first strategy is called the clone. When
16. 0:41I was a kid, my grandpa had
17. 0:44>> [music]
18. 0:44>> late-stage Alzheimer's. Every few
19. 0:47minutes, he would look at my
20. 0:48grandmother, the woman he'd been married
21. 0:51for over 60 years, and he would ask,
22. 0:53"Who are you?" [music] And then he would
23. 0:55go ahead and tell me a story, forgetting
24. 0:58that he just told [music] me the exact
25. 0:59same story 5 minutes ago. His brain
26. 1:02didn't have memory of his last 5
27. 1:05minutes. Chat GPT suffers from something
28. 1:08very similar. Every time you open a new
29. 1:10chat session, you're starting from zero
30. 1:13context. But, there is a way to give
31. 1:16Chat GPT an almost telepathic ability.
32. 1:20It can keep the memory and context for
33. 1:23you alive without you explaining every
34. 1:26single time. For that, I apply a
35. 1:28three-step framework called ESP. E is
36. 1:32extract. Open the voice mode in Chat GPT
37. 1:36and say,
38. 1:37"Listen, you're my chief of staff.
39. 1:39Interview me for 20 minutes, 30 minutes,
40. 1:42my background, my tone, my strategy, my
41. 1:45blind spots."
42. 1:47Then, talk a while. It be messy. Don't
43. 1:50self-edit. This step is about raw
44. 1:54capture only. S is synthesize. Now, this
45. 1:58is where you turn the raw data into an
46. 2:00operating system. So, upload the
47. 2:02transcript of whatever you said to
48. 2:04ChatGPT and ask ChatGPT to organize it
49. 2:08into a summary on who you are and how
50. 2:10you think. Now, go to settings, custom
51. 2:13instructions, and paste your summary. By
52. 2:15the way, this is also the place where
53. 2:17you tell ChatGPT exactly how you would
54. 2:20want it to behave. Always format data as
55. 2:23tables. Don't send me walls of text. But
56. 2:26now, ChatGPT knows who you are, how you
57. 2:29want your responses shaped, and what to
58. 2:32avoid by default. And here's where it
59. 2:35gets really wild. P is for [music]
60. 2:38projects. Think of a project as a space
61. 2:41where you store your context. It lives
62. 2:44with your instructions, memories, files,
63. 2:47repeated tasks. For instance, for your
64. 2:50work projects, upload your resume, PDFs,
65. 2:54and data that's relevant. You can even
66. 2:56customize the instructions inside each
67. 2:58project settings. Just remember not to
68. 3:01mix domains. So, for example, I have my
69. 3:04consulting project and then my YouTube
70. 3:07project [music] and my health project
71. 3:10separate. This turns your AI from a
72. 3:12temporary chat into a persistent brain.
73. 3:16But that's only the first battle. You
74. 3:18still need an army to do the heavy
75. 3:20lifting. That is our next strategy. The
76. 3:23second strategy is called the swarm.
77. 3:26There's a great story in ancient
78. 3:28>> [music]
79. 3:28>> Indian mythology about a demon who was
80. 3:31impossible to defeat with a single
81. 3:33weapon. So, the great goddess manifested
82. 3:36eight arms wielding eight different
83. 3:39weapons at once. She was known as
84. 3:42Ashtabhuja Durga. Eight arms, eight
85. 3:45weapons, eight battles fought [music] in
86. 3:48parallel.
87. 3:49And most of us are fighting our own
88. 3:51modern demons with just our two bare
89. 3:54hands. Our brain can only handle one
90. 3:57task at a time. You need to hire three
91. 4:00specific advanced capabilities of
92. 4:03ChatGPT. Number one, your hands. Stop
93. 4:07doing tasks one by one. Use the agent
94. 4:09mode. Dispatch the swarm. Instead of
95. 4:12searching for a single lead, give
96. 4:14ChatGPT a command. [music] Find 50 leads
97. 4:17in the SaaS space. Verify their emails.
98. 4:20Rank them by revenue. That's it. Let it
99. 4:22run. You go get coffee. But there is one
100. 4:25[music] catch. ChatGPT is still trapped
101. 4:27inside a text box. It can find leads,
102. 4:30but it can't actually put them into your
103. 4:33CRM or build a process around them.
104. 4:35Number two, your eyes.
105. 4:38Stop describing your problems. ChatGPT
106. 4:41is very visual. You can take a photo of
107. 4:43a whiteboard sketch, a broken car part,
108. 4:46or a complex analytics dashboard. Just
109. 4:49upload it and then say, "Analyze this
110. 4:52picture. Tell me what's wrong with it
111. 4:53and tell me one move that I can make to
112. 4:55fix it." You're using ChatGPT as your
113. 4:58eyes, not just your keyboard. [music]
114. 5:00And the great thing is ChatGPT can see
115. 5:02things human eyes can't. And finally,
116. 5:04number three, your research assistant.
117. 5:07You can do deep research that was once
118. 5:09the domain of an analyst who you would
119. 5:11pay $50,000 for. Recently, I was in a
120. 5:14meeting with a chief technology officer
121. 5:16of a large firm and he said something
122. 5:18very interesting. He said, "I had to
123. 5:21evaluate six [music] presentations
124. 5:23delivered by a very large IT consulting
125. 5:25firm on the technology infrastructure of
126. 5:28a massive data [music] center." And
127. 5:30there were 300 pages of dense
128. 5:32information. But what did he do? He
129. 5:35uploaded them all in ChatGPT, hired a
130. 5:37swarm of AI analysts one afternoon.
131. 5:41That's all it took. Now, that would have
132. 5:43taken weeks before.
133. 5:45>> [music]
134. 5:45>> That's the world we're heading into. In
135. 5:47the age of AI, you won't be billing for
136. 5:50your time anymore. You'll only bill for
137. 5:53your taste. [music] So, this swarm gives
138. 5:55you an army of powerful interns. But,
139. 5:58your next strategy isn't about hiring AI
140. 6:01to take orders. [music] It's there to
141. 6:03challenge you. Your third strategy,
142. 6:05then, is the devil's advocate. Around
143. 6:08the 1500s, the Catholic Church was the
144. 6:11most powerful organization on Earth.
145. 6:14But, it was also under a lot of attack.
146. 6:17By 1587, the Pope had a massive problem.
147. 6:20The Church was canonizing saints left
148. 6:23and right. So, the Pope did something
149. 6:25pretty radical.
150. 6:27He created an office called the
151. 6:29Advocatus Diaboli, [music] the devil's
152. 6:32advocate. This person's entire job was
153. 6:35to argue against the canonization
154. 6:37[music] of any saint, to find the dirt,
155. 6:40to poke holes, to be ruthless. The
156. 6:43Church [music] realized this 400 years
157. 6:45ago. When the stakes are high, good
158. 6:47decisions are rarely made through
159. 6:50agreements. They come from conflict.
160. 6:52Now, why is this necessary? Well, it's
161. 6:54called the IKEA effect. We fall in love
162. 6:57with anything we build ourselves.
163. 6:59Research shows that smart people are
164. 7:01great at spotting [music] flaws in other
165. 7:03people's thinking, physically incapable
166. 7:06of seeing them in [music] their own. So,
167. 7:08the key insight? Most people use ChatGPT
168. 7:12for validation.
169. 7:13But, you should use it for violation.
170. 7:17But, here's the problem. ChatGPT's
171. 7:18[music] programmed to keep you happy.
172. 7:21So, you have to force it to behave
173. 7:23differently. So, here are the three
174. 7:25tricks to break the AI's people-pleaser
175. 7:28programming. First, use [music] magic
176. 7:30triggers. Force adversarial thinking.
177. 7:33Instead of saying, "What do you think?"
178. 7:35be specific. Prompt it and say, "Audit
179. 7:38this using first [music] principles."
180. 7:40Or, "Act as my fiercest competitor." You
181. 7:43may not get what you want, but you will
182. 7:45get what [music] you need. Second,
183. 7:47shadowboxing.
184. 7:49Don't step into the ring unprepared.
185. 7:51Let's say if you're going out to raise
186. 7:53money for your venture. Tell the AI
187. 7:55simulate a debate between a ruthless
188. 7:58skeptical VC and a visionary product
189. 8:01designer. Let the VC push back on the
190. 8:04idea. I want to see that fight and learn
191. 8:07from it. This way you let AI expose the
192. 8:10weak points so you don't fumble on them
193. 8:13>> [music]
194. 8:13>> in the real world meeting. Third, the
195. 8:15blind spot check. Feed it your strategy
196. 8:18and ask one question. Based on what you
197. 8:20know about me, what are the gaps in my
198. 8:23reasoning? When we get successful, we
199. 8:26develop two bad habits. We trap
200. 8:29ourselves
201. 8:30>> [music]
202. 8:30>> in this bubble of our ego and we
203. 8:33surround ourselves with people who just
204. 8:36say yes. That's why this strategy is so
205. 8:38critical to save you from yourself. But
206. 8:41knowing where you're wrong or why you're
207. 8:44wrong is not enough. You still have to
208. 8:46learn fast about what's right. That's
209. 8:49where our next strategy comes in, the
210. 8:50neural link. And it breaks the rule most
211. 8:53of us grew up believing that mastery
212. 8:56takes 10,000 hours.
213. 8:58You know, there is a famous story about
214. 8:59a physicist, Max Planck. After he won
215. 9:02the Nobel Prize, he went on a tour
216. 9:05giving the same speech about quantum
217. 9:07mechanics every night. Now his chauffeur
218. 9:09sat in the front row every time and
219. 9:11literally memorized every word. One
220. 9:14night the chauffeur said, "Professor,
221. 9:15I'm getting bored just listening to you.
222. 9:17Let's switch. I'll give the lecture on
223. 9:19stage and you sit in the audience
224. 9:21wearing my cap." Planck of course
225. 9:23agreed. Same words, same structure, same
226. 9:26cadence. The speech was flawless, but at
227. 9:29the end a physicist in the audience
228. 9:32stood up and asked a deep technical
229. 9:34question. But the chauffeur didn't panic
230. 9:37at all. He smiled and said, "I am
231. 9:39surprised to hear such a simple question
232. 9:42in a city as advanced as Munich. To
233. 9:44prove how simple it is, I'm going to let
234. 9:46my chauffeur answer it." Of course, that
235. 9:48was Max Planck. But here's the point.
236. 9:50Most people use ChatGPT like that
237. 9:53chauffeur. They copy-paste
238. 9:55smart-sounding phrases and answers
239. 9:57[music] to sound intelligent. They have
240. 9:59fluency without foundation. So here are
241. 10:02three ways to get that Neuralink true
242. 10:04understanding without becoming that
243. 10:06chauffeur. First, the jargon translator.
244. 10:09For instance, tell it, "Explain the
245. 10:11concept of API using the restaurant
246. 10:14metaphors only." Suddenly, the menu
247. 10:17becomes the documentation, the kitchen
248. 10:19is the server, and the waiter is the
249. 10:22API, the messenger that takes your
250. 10:24order, delivers it in the kitchen,
251. 10:26brings back the results. You aren't
252. 10:28learning from scratch, you're
253. 10:30translating that jargon into your
254. 10:33jargon. Second, the progressive
255. 10:35professor. Richard Feynman said that if
256. 10:38you can't explain it simply, you don't
257. 10:40understand it. So don't ask ChatGPT to
258. 10:42explain quantum computing, but instead
259. 10:45say, "Explain quantum computing to me
260. 10:48three ways. One, like I am a
261. 10:51five-year-old. Give me the simplest
262. 10:53explanation. Number two, like I am a
263. 10:5515-year-old. Explain it with logic and
264. 10:58examples. And three, like I'm a college
265. 11:00student. Share the underlying theory."
266. 11:02When you do that in that particular
267. 11:05order, you will feel any concept click
268. 11:08into place in your brain instantly. And
269. 11:10third, the degree download. You teach
270. 11:13it, you keep it. So use AI to be your
271. 11:16target audience or a peer. Prompt it to
272. 11:19say something like, "I want to build a
273. 11:21cool AI commerce product. I need to
274. 11:24learn the foundation of how machine
275. 11:26learning works. Skip the theory, only
276. 11:28give me actionable skills, no fluff, and
277. 11:30then test me so I can teach it back to
278. 11:32you. You just replaced a boot camp worth
279. 11:35$50,000
280. 11:37with a PDF. So, here you are. You have
281. 11:39built a high-performance ChatGPT
282. 11:42machine. But, such machines tend to
283. 11:44overheat. The fifth strategy is called
284. 11:47the executive coach, and without it none
285. 11:50of the other strategies would work.
286. 11:51>> [music]
287. 11:52>> 2,000 years ago, Marcus Aurelius was the
288. 11:55emperor of Rome, but he was also the
289. 11:58loneliest. He had no one to talk,
290. 12:00[music] so he hired a coach. That coach
291. 12:04was a blank scroll. Every night in his
292. 12:06tent, he wrote down his anxieties,
293. 12:09[music]
294. 12:09his fears, his advice to himself. He
295. 12:12titled that manuscript to himself. If
296. 12:16you are a successful executive [music]
297. 12:18or on the path of being one, you need an
298. 12:21executive coach. I had two, and they
299. 12:24both changed my trajectory dramatically.
300. 12:27But, sometimes executive coaches and
301. 12:29mentors aren't going to be available to
302. 12:31you at midnight when you need them the
303. 12:33most, when you can't sleep because of a
304. 12:35decision you have to make. And that's
305. 12:37when you can use ChatGPT to show you a
306. 12:39mirror with zero emotional baggage. You
307. 12:42can even turn the voice mode, that's
308. 12:45what I do.
309. 12:45>> [music]
310. 12:46>> Here are three tricks to turn ChatGPT
311. 12:48into your temporary executive coach.
312. 12:51First, the vomit mirror.
313. 12:53When you are overwhelmed, your brain is
314. 12:57going to be a tangled mess. It happens
315. 12:59to me. Open voice mode and say, "I'm
316. 13:02going to rant for 5 minutes about
317. 13:05everything stressing me out. Don't
318. 13:06interrupt, just listen. When I'm done,
319. 13:09organize my chaos into three lists:
320. 13:12problems I can solve today, problems I
321. 13:14[music] need to delegate, noise I need
322. 13:16to ignore."
323. 13:17That's it. And then,
324. 13:19>> [music]
325. 13:19>> go ahead. Just rant as long as you want.
326. 13:22Second, the stoic mirror. It's It's
327. 13:25that you may have just lost a deal or
328. 13:29you got fired or something terrible
329. 13:31happened. You can say to ChatGPT at that
330. 13:33point, I just took a massive loss. I
331. 13:36feel like a failure. Reframe the
332. 13:38situation for me using [music] the stoic
333. 13:41principles and tell me why this is
334. 13:44actually an opportunity. Have that
335. 13:46conversation until you are convinced.
336. 13:48Push back, let [music] it push back to
337. 13:50you.
338. 13:51Even if you're not convinced, it'll slow
339. 13:54down the emotional spiral that you're on
340. 13:56instantly.
341. 13:57>> [music]
342. 13:57>> And the third and the final mirror, the
343. 13:59decision mirror. Before you make a
344. 14:01massive decision like quitting a job or
345. 14:04starting a new company, talk it out. I
346. 14:06am about to do X. Here's my logic. Let's
347. 14:09have a conversation. You don't have to
348. 14:10put faith in ChatGPT's advice. In
349. 14:13[music] fact, please don't. But talking
350. 14:16it out clears your own head about your
351. 14:18own judgment. We covered five strategies
352. 14:21and 15 ways to optimize your
353. 14:23relationship with ChatGPT. But here's
354. 14:26the truth. All of these skills are
355. 14:28essentially alpha. In finance, when
356. 14:32traders find an edge, they make a
357. 14:33fortune and it's called the alpha. But
358. 14:35eventually, everyone figures it out.
359. 14:38What's rare today becomes required
360. 14:41tomorrow. That's how life works. And
361. 14:43what happens when it all becomes a
362. 14:46commodity? [music]
363. 14:46How will you maintain your edge then?
364. 14:49I'll share my personal story that
365. 14:50changed my idea about my edge. A week
366. 14:53before I took my CEO job, my executive
367. 14:56coach pulled me aside
368. 14:58>> [music]
369. 14:58>> and he said, remember, your title is
370. 15:01CEO, chief executive officer, [music]
371. 15:03but that E does not just stand for
372. 15:06executive. It also stands for emotions.
373. 15:09You're the chief emotions officer now as
374. 15:11well. You have to manage the emotions of
375. 15:15your own people.
376. 15:16>> [music]
377. 15:16>> That conversation stayed with me for
378. 15:19years because for a long time, we have
379. 15:22forgotten that message. You know, we
380. 15:24spend the entire industrial age trying
381. 15:26to turn ourselves into machines. We
382. 15:28measured our worth by our output, our
383. 15:31productivity. If we didn't produce or we
384. 15:33didn't consume, we didn't matter to the
385. 15:35market. So, what did we do? We tried to
386. 15:38out-robot the robots. But now, the
387. 15:41actual machines have arrived to take
388. 15:44their job back.
389. 15:45>> [music]
390. 15:45>> And that's not a threat. That's actually
391. 15:47a permission.
392. 15:48Permission to us to stop pretending like
393. 15:51we're machines. Permission to feel,
394. 15:55to show compassion, to connect to those
395. 15:57mysteries, to connect to others, to
396. 15:59connect to this mystery called life. In
397. 16:01a way, I feel that AI and machines are
398. 16:04coming to free us from labor
399. 16:06>> [music]
400. 16:06>> so we can return to our essential
401. 16:09nature.
402. 16:10AI's ultimate gift might very well be
403. 16:13then
404. 16:14that it will make us more human.
405. 16:17If you like this video, here's another
406. 16:19one that I think you will like. Thank
407. 16:21you.
408. 16:22And I love you.