This AI System Will Make You So Smart It’s Almost Unfair — Transcript

Full transcript

1. 0:00I built a fully AI-powered brain that
2. 0:02runs across every part of my life. It
3. 0:04helps me run dozens of companies, and it
4. 0:06gets days of my life back. So, today I'm
5. 0:08going step-by-step on how you can build
6. 0:11the same AI brain I use to become part
7. 0:13of the 0.0001%
8. 0:15of AI users in the world, starting with
9. 0:18going pro.
10. 0:20Simple, get a pro membership for your
11. 0:22AI. It doesn't matter which AI, invest
12. 0:25the money. Free versions of AI give you
13. 0:27the worst version of the tool, slow to
14. 0:29answer. They're older models. So, as
15. 0:32soon as you pay, you've now become part
16. 0:34of the 0.3% of all users in the world
17. 0:37that have a paid membership. But, we
18. 0:38want you in the 0.01%.
19. 0:40So, let's get into how we actually do
20. 0:42that. Install the brain.
21. 0:44Every AI conversation forgets the second
22. 0:47you close it. Sure, they have some tools
23. 0:49for memory, but it's basic things. The
24. 0:510.01%
25. 0:53give AI a permanent place to read from.
26. 0:56That place has to be your files, in your
27. 0:58structure that any AI can read your
28. 1:00digital brain. So, here are three
29. 1:02options you can pick from. The first is
30. 1:04a Google Drive. Most people have this
31. 1:06already, and it's kind of where it all
32. 1:07started. But, the problem is is that
33. 1:10it's pretty limited, and it's not
34. 1:11process. The second is Notion. This is a
35. 1:14cloud-based data structure system where
36. 1:17I believe that the way it's built is AI
37. 1:20native. Google is a basic digital
38. 1:22version of a folder system. Notion is a
39. 1:25modern AI-powered data store. So, yes,
40. 1:28it has the files, but it also has the
41. 1:30context, it also has the data. If you
42. 1:32want to go to the next level, and this
43. 1:34is what I use, I use Obsidian. The
44. 1:36challenge with a lot of these AI systems
45. 1:38is that you can't read what it's saved
46. 1:40in. You can't read the information. So,
47. 1:42Obsidian uses what's called markdown
48. 1:44files. It's free, and it's visual. So,
49. 1:47not only is it like a big wiki that sits
50. 1:49on your computer, but you can also see
51. 1:50how the brain forms and all the
52. 1:53connections and links to the information
53. 1:55within it. And the cool part is the
54. 1:56graph view within Obsidian is the only
55. 1:59one that actually feels like [music] a
56. 2:00brain. At the end of the day, it comes
57. 2:01down to personal preference. There's
58. 2:02literally hundreds of other tools you
59. 2:03can use. These are the ones I've used
60. 2:05and where I'm at today. So, if you're
61. 2:06not sure, just start with Obsidian,
62. 2:08download it, and just have the AI create
63. 2:10a second brain. It takes 5 minutes. Now
64. 2:12you have a brain that's installed, but
65. 2:14if AI reads it, there's nothing in it.
66. 2:15So, how do you make sure that actually
67. 2:16has the information in it sounds like
68. 2:18you? Give your brain an identity.
69. 2:21Without an identity, the AI is just
70. 2:23going to sound like some robot.
71. 2:25But to have the agent act like you, you
72. 2:27have to tell it how to act. That's the
73. 2:29identity. These three files will
74. 2:31literally turn your AI from a tool to a
75. 2:34digital clone that sounds exactly like
76. 2:36you. Number one, user.md.
77. 2:39Who are you? What is your role? How do
78. 2:41you communicate? What are the frameworks
79. 2:43you live by? How you present problems to
80. 2:46you? Two is the soul.md file. This is
81. 2:49for the AI. It's kind of wild, but you
82. 2:51can define the voice, the tone, the
83. 2:54values that essentially the AI will
84. 2:57adopt so that it can interact and
85. 2:59operate like you. So, the user.md file
86. 3:02is how you are, the soul is how it acts,
87. 3:06and then the third is the identity.md
88. 3:08file. That is who the AI is and what
89. 3:11they do. For example, I named my Kai. I
90. 3:13actually didn't even name him. I used
91. 3:15him for 2 weeks and I had him name
92. 3:16himself. So, for example, in my user.md
93. 3:19file, I included who I am. You know, I'm
94. 3:21a founder, I'm a CEO, I'm an author, I'm
95. 3:23an investor. My communication style,
96. 3:26things like being direct, fast, concise,
97. 3:28frameworks over fluff, and things I care
98. 3:30about, like freedom, execution, systems,
99. 3:33leverage, some of my operating
100. 3:35principles, theory of constraints,
101. 3:37minimum effective dose, the 1 3 1 rules,
102. 3:39a lot of different frameworks that I've
103. 3:40taught you, I put those in there so it
104. 3:42can operate from those principles by
105. 3:44default. Then my soul.md includes
106. 3:46instructions for my AI. So, for example,
107. 3:49talk like Dick. I'm direct, high
108. 3:50conviction, founder to founder,
109. 3:52challenge me with love. No hedgy words.
110. 3:54I don't like try, I don't like hope, I
111. 3:56don't like maybe. Get rid of those. And
112. 3:58then be resourceful. Try to solve the
113. 3:59problem before you ask me. These are
114. 4:01things that are in the soul file. And
115. 4:02then my identity file. Kai is part
116. 4:04coach, part chief of staff, part
117. 4:06accountability partner. These are things
118. 4:07that we define so that he can operate
119. 4:09from that place. Now, here's a pro tip.
120. 4:12Don't write these yourself. You can get
121. 4:14any AI to draft them for you. It's kind
122. 4:16of the coolest thing in the world. You
123. 4:17can ask it, "Hey, I need to create this
124. 4:18new AI, so can you interview me about
125. 4:20how I work, what I value, how I like to
126. 4:22communicate, and what I need from an AI
127. 4:24assistant?" Then have it draft the agent
128. 4:27files, and it'll know user, soul, and
129. 4:29identity files. And once you have those,
130. 4:31now we can just save them and then point
131. 4:33our AI to it so we can reference back to
132. 4:35it every single time. Now, look, this is
133. 4:36great for you, but what about the rest
134. 4:38of your team? That's why I built a whole
135. 4:40AI company operating system playbook
136. 4:42around how to integrate AI into every
137. 4:44department of your business. It's got
138. 4:46all the tools, the workflows, the
139. 4:48templates, everything your team needs to
140. 4:50use AI to get 10x the results. If you
141. 4:52want it, just find me on Instagram and
142. 4:53DM me the words YouTube OS, and I'll
143. 4:55send it right over. So, the brain knows
144. 4:57who you are now. Cool. Now, you probably
145. 5:00want to start feeding it information.
146. 5:02Not yet. Before you do that, you have to
147. 5:04make sure it's organized first. Wire the
148. 5:07brain. AI needs a way to clearly
149. 5:09identify the context of the brain. No
150. 5:12structure means AI just drowns in noise
151. 5:14and eventually hallucinates. I went from
152. 5:1660% accuracy in asking it questions to
153. 5:1985% just by setting up the right
154. 5:21folders. So, create these seven folders
155. 5:24inside your vault to get your brain
156. 5:26organized. The first one is people. You
157. 5:28should have a place where every person
158. 5:30in your life has a people file and has
159. 5:31all the context on those people. Two is
160. 5:34projects. What's one place where all of
161. 5:36your projects, timestamped, are listed
162. 5:39and located with all the context that it
163. 5:40needs. Third is decisions. What are the
164. 5:43decisions you've made in the past? How
165. 5:45did you decide? What the alternatives
166. 5:47are? Four is companies. This is any
167. 5:49information on the company I'm dealing
168. 5:51with, the research, the competitors,
169. 5:53etc. Essentially, it's like another
170. 5:55index of companies that I might be
171. 5:56interacting with. Number five is
172. 5:58meetings, cuz that's where a lot of the
173. 5:59decisions are made, that's a lot of
174. 6:01preferences are made, that's where team
175. 6:02and information about people is made.
176. 6:04So, we need a meetings folder. Six is
177. 6:06daily. I use this as a daily dump to
178. 6:09analyze and lock in every day daily
179. 6:12decisions that are three to five lines
180. 6:14that explain what happened throughout
181. 6:15that day. Then seven is knowledge. This
182. 6:17is where I'm extracting insights,
183. 6:20frameworks, quotes, anything that kind
184. 6:22of makes me me that I could reuse in the
185. 6:24future, I want to save it in this
186. 6:26knowledge folder. Now, here's a pro tip.
187. 6:28I have an eighth folder. It's called an
188. 6:30MOC folder, a maps of content. It's a
189. 6:33folder that includes files that is
190. 6:35summaries of bodies of work. So, for
191. 6:37example, I could have a youtube.md MOC
192. 6:40linking all my scripting frameworks, all
193. 6:42my hook libraries, all my thumbnail
194. 6:44task, all living in different folders,
195. 6:46but the MOC pulls it together. It's kind
196. 6:48of like consolidating information into
197. 6:51one simple reference file. Now, you
198. 6:53don't have to build an MOC from day one,
199. 6:55that's why it's a pro tip, but you build
200. 6:56them when a topic seems to get messy.
201. 6:58Think of it like this, your brain has
202. 7:00different parts of your brain that's
203. 7:01responsible for different things, motor
204. 7:03functions, smells, vision. That's kind
205. 7:06of what we're doing for a AI brain
206. 7:09that's going to have to do things on our
207. 7:11behalf. That's why we've created these
208. 7:12folder structures. Now, you can create a
209. 7:13bunch more over time, but this is where
210. 7:16we start. Now, the brain is organized
211. 7:18and it's ready to be fed information,
212. 7:20but how do you actually feed it? Then
213. 7:22the next step, you have to feed the
214. 7:24brain.
215. 7:25So, what you do is you connect all your
216. 7:27system, then you have your AI extract
217. 7:30all that information into the folder
218. 7:33structure that your brain is structured
219. 7:35in, because the brain only needs the
220. 7:36information it needs to use to make
221. 7:38decisions. There is a hundred thousand
222. 7:41data points that are being sent to me
223. 7:43right now from the light to the
224. 7:44temperature to everything. And my brain
225. 7:46is filtering out all the noise, and it's
226. 7:48only capturing the stuff I need to know.
227. 7:51This system works the exact same way
228. 7:53because it doesn't need the raw stuff
229. 7:56it's pointed back to it, but it does
230. 7:57need to synthesize and extract the
231. 8:00people, the decision, the knowledge that
232. 8:02happened inside those meetings. You just
233. 8:04have your AI do it. You have your raw
234. 8:06files where you point it to with your
235. 8:07connectors, and then you just have your
236. 8:09AI extract every day all of the
237. 8:12information based on the folder
238. 8:14structure within Obsidian. You just
239. 8:16improve your brain at the speed of
240. 8:18links. That's where we're at right now.
241. 8:19Now, I use Grainola to auto transcribe
242. 8:21all my meetings and then extract
243. 8:23decisions, commitments, and actions. And
244. 8:25the cool part is in Grainola settings,
245. 8:26you can set a custom template for this
246. 8:28exact prompt to do the extraction. Just
247. 8:30write this down. Extract from this
248. 8:32meeting: one, decisions, what did I
249. 8:34decide today, by whom, and why? Two,
250. 8:36commitments, who promised what, by when?
251. 8:38Three, preferences, how people work and
252. 8:41communicate. And four, key insights,
253. 8:43frameworks, strategic shifts,
254. 8:45non-obvious observations. Output as a
255. 8:47markdown, skip small talk. Now, Grainola
256. 8:50drops those outputs into the meetings
257. 8:53folders as, again, the date, meeting
258. 8:55name.md. Now, this is where things get
259. 8:58cool. The brain is installed. It knows
260. 9:00who you are. It's organized and it's
261. 9:02starting to feed itself. But the real
262. 9:04brain doesn't just hold things, it
263. 9:06rewires itself overnight. But how do you
264. 9:08get the brain to get smarter overnight?
265. 9:11Compound the brain.
266. 9:14Now, I'm going to process the brain the
267. 9:16way my real brain works. See, when we go
268. 9:18to sleep every night, it takes
269. 9:20everything that you were exposed to,
270. 9:22>> [music]
271. 9:22>> and it organizes, synthesizes, it
272. 9:25consolidates, it summarizes, it looks at
273. 9:27things that it needs to connect, and
274. 9:29then it even prunes things that it needs
275. 9:30to forget. And an AI brain works the
276. 9:32exact same way. And this is what
277. 9:35separates everything else we talked
278. 9:36about in the past, [music] storage, into
279. 9:39an actual brain. Here's two ways to
280. 9:41refine your brain. The first way is
281. 9:44manual. At the end of the day, if you
282. 9:45want, you can type in your AI a prompt
283. 9:48that's going to tell it to go into your
284. 9:50brain and run this process. Essentially,
285. 9:52you just say, "Read everything that
286. 9:53you've added to my vault today, and I
287. 9:55need you to find orphan notes, mentions
288. 9:57of people, projects, companies that
289. 9:58don't have their own files. You got to
290. 10:00create those files, you have to
291. 10:01consolidate duplicates, you have to
292. 10:03update the relevant map of contents with
293. 10:05new links. You have to flag anything
294. 10:07strategic for me to review for
295. 10:08tomorrow." It'll take 60 seconds, but
296. 10:10you'll wake up tomorrow to a cleaner,
297. 10:12more connected brain. The second, and
298. 10:14this is what I do, is automate it. I
299. 10:16actually use Claude's scheduled task to
300. 10:19run a cron job that runs the same prompt
301. 10:22I just gave you every night at 11:00.
302. 10:25And then, once you set this up, your
303. 10:26brain will constantly improve itself.
304. 10:29Every morning, I wake up and I click the
305. 10:30graph view within my Obsidian brain, and
306. 10:32I see it morph and evolve every night.
307. 10:36It's pruning, it's cleaning up, it's
308. 10:37essentially getting smart because the
309. 10:39more links that it creates, the higher
310. 10:41the signal is for context. So, when I
311. 10:43get my AI to talk to my brain, it comes
312. 10:46back with better answers. So, let's do a
313. 10:47quick recap. Number one, we went pro. We
314. 10:50decided to pay for real power in AI.
315. 10:53Then, we installed the brain. We figured
316. 10:56out where we're going to save it, and
317. 10:57then for most of you, you probably use
318. 10:58Obsidian. Then, we gave it an identity.
319. 11:00We created the files it needed to know
320. 11:03how to react and respond as an AI. Then,
321. 11:05we wired it to a brain, so it knew the
322. 11:08folder structure of how we're going to
323. 11:09save information. Then, we fed the
324. 11:11brain. And now, it gets smarter
325. 11:13overnight. You can't have an autonomous
326. 11:16agent unless it has context. Using the
327. 11:19brain is how we build that. If it knows
328. 11:22the meetings you were in, the decisions
329. 11:23you made, the people in your life, you
330. 11:25can say things like, "Send the invite to
331. 11:27John," and it knows who John is because
332. 11:29John has the most links in the brain for
333. 11:32the word John, and it knows John's
334. 11:34email, and it knows his cell number.
335. 11:36It's crazy how all this information is
336. 11:38available, but it's never reused in our
337. 11:40AI prompts. If you build the AI-powered
338. 11:43brain, you will start to buy back days
339. 11:45of your life because the AIs can
340. 11:48actually run your life. And remember, if
341. 11:50you want the AI company OS playbook for
342. 11:52your whole team, just DM me YouTube OS
343. 11:54on Instagram, and I'll send it right
344. 11:55over. And if you want to learn how to
345. 11:57build a $10 million solo AI business,
346. 12:00click here, and I'll see you on the
347. 12:01other side.