The AI Second Brain Complete System Walkthrough

The AI Second Brain: Complete System Walkthrough ![rw-book-cover](https://img.youtube.com/vi/-TAowrw97-4/maxresdefault.jpg) 🔗 **Source:** [The AI Second Brain: Complete System Walkthrough](https://www.youtube.com/watch?v=-TAowrw97-4&t=2149s) > [!note]+ Document Notes > tags: #ai #productivity #knowledge-management > author: Tiago Forte > source: YouTube > reference: https://www.youtube.com/watch?v=-TAowrw97-4&t=2149s > date created: 2024-12-28 12:00 > > # 🏷️ Title: The AI Second Brain: Complete System Walkthrough (Live with Tiago Forte) >  ## 🧰 Action Inbox - [H1] Create foundational master prompt with basic personal details — immediate priority for AI optimization #ai - [H2] Design platform-agnostic organizational system using PARA method — week 1 implementation #knowledge-management - [H3] Test context window limitations with current AI tools — validate 30-50K word practical limits #productivity ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - [Building a Second Brain](https://fortelabs.com/blog/basboverview/) - [PARA Method](https://fortelabs.com/blog/para/) ## 📝 Summary - Tiago Forte presents his AI Second Brain system that builds on pre-AI productivity methods (CODE and PARA) with AI-powered enhancements - The core bottleneck has shifted from time/attention/intelligence to collecting, curating, and providing context to AI tools - Master prompts serve as personal operating manuals for AI, containing curated context about the user's goals, values, and working style - Context windows have practical limits of 30-50K words despite claims of millions, requiring selective curation rather than data dumping - The system is taught in three progressive levels: foundational, intermediate, and advanced over a 3-week intensive program ## 📖 Main Content Forte argues that productivity bottlenecks have fundamentally shifted in the AI era. Rather than being constrained by time, attention, or intelligence, professionals now face limitations in Personal Context Management (PCM) - the skill of collecting, curating, and providing relevant context to AI systems. The AI Second Brain system builds on established methodologies. CODE (Capture, Organize, Distill, Express) remains the foundation for creative processes, while PARA (Projects, Areas, Resources, Archives) provides organizational structure. These pre-AI methods become more critical, not obsolete, because AI requires well-organized input to produce valuable output. Context windows present a crucial technical constraint. Despite vendor claims of million-token capacities, research shows effective context windows are limited to 30-50,000 words. Beyond this threshold, AI systems experience context poisoning (fixation on errors), context distraction (irrelevant information interference), and context confusion (conflicting information handling). The master prompt concept serves as a personal operating manual containing basic personal details, professional information, income sources, core values, communication preferences, working styles, and quality standards. This curated context enables AI to provide personalized, relevant responses rather than generic outputs. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/-TAowrw97-4?t=0) — [H1] New Productivity Bottleneck Identified - Key idea: Modern productivity constraint has shifted from time/attention to context management - Evidence: "There's a new bottleneck in modern productivity... collecting, curating, and providing context" #### ⏱️ [05:30](https://youtu.be/-TAowrw97-4?t=330) — [H2] Context as Competitive Advantage - Key idea: Personal curated context becomes irreplaceable competitive moat - Evidence: "That context is the only way that you escape competition, that you create a real competitive advantage" #### ⏱️ [08:45](https://youtu.be/-TAowrw97-4?t=525) — [H3] Context Window Reality Check - Key idea: Practical context limits are much smaller than advertised - Evidence: "Even if they say a million words, 2 million, 5 million words, if you exceed around 30 to 50,000 words... you start having these negative impacts" #### ⏱️ [12:15](https://youtu.be/-TAowrw97-4?t=735) — [H1] Pre-AI Skills More Important Than Ever - Key idea: Organization and curation skills become amplified, not obsolete - Evidence: "All those things are still important. They're even more important because you have to do them to take advantage of AI's power" #### ⏱️ [18:20](https://youtu.be/-TAowrw97-4?t=1100) — [H2] Four-Part System Architecture - Key idea: AI Second Brain consists of capture, organize, master prompt, and AI advisors - Evidence: "Capture notes... organize it with PARA... flows into what we call a master prompt... AI board of advisors" #### ⏱️ [22:30](https://youtu.be/-TAowrw97-4?t=1350) — [H1] Master Prompt as Operating Manual - Key idea: Single document containing comprehensive personal and professional context - Evidence: "Your personal operating manual for AI, for all AI... imagine if that AI knew your basic personal details" #### ⏱️ [35:45](https://youtu.be/-TAowrw97-4?t=2145) — [H2] Live Demonstration Results - Key idea: Master prompt enables sophisticated business analysis and recommendations - Evidence: "You have no scalable high-touch offering that doesn't require Tiago's direct involvement" #### ⏱️ [45:10](https://youtu.be/-TAowrw97-4?t=2710) — [H3] Three-Level Progressive System - Key idea: Build same system three times at increasing complexity levels - Evidence: "Week one... foundational level... week two... intermediate level... week three... advanced level" #### ⏱️ [52:20](https://youtu.be/-TAowrw97-4?t=3140) — [H2] Contrarian Opinion: AI Is Hard Work - Key idea: AI requires more cognitive effort and complexity than commonly portrayed - Evidence: "This stuff is hard... AI takes more cognitive effort... making these architectural decisions" #### ⏱️ [55:40](https://youtu.be/-TAowrw97-4?t=3340) — [H3] Note-Taking Still Essential - Key idea: Human discernment required to identify unique information not in AI training data - Evidence: "Note taking is more valuable than ever... still requires human discernment and human judgment" #### ⏱️ [58:15](https://youtu.be/-TAowrw97-4?t=3495) — [H2] Agents Are Not Ready - Key idea: AI automation tools frequently fail and make critical errors - Evidence: "Agents are largely a bunch of BS... they constantly fail... constantly making mistakes" #### ⏱️ [61:30](https://youtu.be/-TAowrw97-4?t=3690) — [H1] Human Enhancement Over Replacement - Key idea: Focus on cognitive augmentation rather than automation - Evidence: "AI as an enhancement and as an amplifier to you, the human... augmentation, not automation" ## 📚 References - [Building a Second Brain by Tiago Forte](https://fortelabs.com) - [PARA Method Documentation](https://fortelabs.com/blog/para/) - [Context Rot Research Paper](https://fortelabs.com/blog/context-rot/) ## 🗣️ Paraphrase - The fundamental constraint in modern productivity has shifted from time and attention management to the ability to curate and provide relevant context to AI systems - Personal context management becomes a crucial skill as AI can only work effectively with what users deliberately provide as curated information - While AI capabilities are impressive, they require well-organized human input and thoughtful design rather than simple data dumping to achieve meaningful results ## 📰 Newsletter **Transform Your Productivity with AI-Enhanced Knowledge Management** The landscape of professional productivity is undergoing a fundamental shift. Tiago Forte's latest insights reveal that our traditional bottlenecks—time, attention, and even intelligence—are being replaced by a new constraint: our ability to effectively curate and provide context to AI systems. This evolution has profound implications for how we approach knowledge management and productivity systems. The pre-AI methodologies like PARA (Projects, Areas, Resources, Archives) and CODE (Capture, Organize, Distill, Express) aren't becoming obsolete—they're becoming more critical than ever. **Key Practical Applications:** - Create master prompts that serve as personal operating manuals for AI interactions - Understand the practical limits of context windows (30-50K words) to avoid information overload - Build progressive systems starting with foundational levels before advancing to complex automation **Helpful Resources:** - Master Prompt Template: Start with basic personal details, professional info, and core values - PARA Organization System: Structure information for both human and AI comprehension - Context Curation Guidelines: Focus on unique insights not available in AI training data **Community Connection:** This approach emphasizes augmentation over automation, making humans more effective rather than replacing them. Join discussions about implementing these systems while maintaining the human element in creative and strategic work. Resources: [Learn WordPress](https://learn.wordpress.org/), [Developer Docs](https://developer.wordpress.org/), [WP-CLI](https://wp-cli.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create master prompts for technical context sharing in demos and talks - Education: Design progressive curriculum following foundational→intermediate→advanced model - Developer experience: Implement context curation guidelines for documentation systems - Developer marketing: Develop AI-enhanced content that maintains human expertise and perspective - Community management: Establish frameworks for collecting and organizing community insights #### OSPO - Policy & Governance: Draft AI usage policies that balance innovation with data privacy - License Compliance: Implement context management for complex license interaction scenarios - Security Management: Develop protocols for sensitive information in AI assistance workflows - Software Inventory: Create master prompts for project-specific technical contexts - Developer Support: Train teams on effective AI interaction through curated context provision - Community Engagement: Use structured context bundles for upstream collaboration - Strategic Alignment: Implement progressive measurement systems for AI integration impact #### Community - Maintainers: Develop master prompt templates for project-specific governance contexts - Contributors: Create onboarding systems that progressively build context and expertise - Sponsors: Implement transparent context sharing for funding decisions and impact reporting - Ecosystem: Design cross-project context exchange protocols and standards #### WordPress - Core: Implement context management for complex codebase navigation and decision-making - Core Performance: Use structured prompts for performance analysis and optimization planning - Core AI: Apply master prompt methodology for responsible AI integration guidelines - Core Program Team: Develop progressive coordination systems following foundational→intermediate→advanced model - Mobile: Create platform-specific master prompts for cross-platform development decisions - Playground: Build educational context bundles for hands-on learning experiences - Community: Design master prompt templates for local meetup and WordCamp organization - Training: Implement progressive curriculum design following three-level system architecture - Marketing: Develop AI-enhanced campaigns that maintain authentic WordPress community voice - Support: Create context-rich knowledge systems for complex support scenario resolution - Design: Use structured prompts for design system decisions and user experience research - Documentation: Implement context curation for technical writing and developer guidance - Photos: Develop master prompts for content curation and community moderation - TV: Create context management for educational content production and archiving - Polyglots: Design translation context systems for cultural and technical accuracy - Accessibility: Implement structured prompts for WCAG compliance and usability testing - Meta: Use master prompt methodology for .org infrastructure and API development decisions - CLI: Create context bundles for command development and documentation enhancement - Hosting: Develop cross-host context sharing for compatibility and performance optimization - Tide: Implement context management for plugin scoring and quality assessment - Openverse: Design master prompts for content curation and legal compliance workflows - Themes: Create context systems for review processes and developer guidance - Plugins: Implement security context bundles for vulnerability assessment and developer education #### The WPCC - Programs: Design progressive fellowship and grant systems following foundational→intermediate→advanced levels - Projects: Create master prompt templates for governance and documentation sprint coordination - Funding: Implement context-rich systems for transparent grant allocation and impact measurement #### FAIR - Technical independence: Develop context curation standards for metadata and provenance verification - Policies: Create master prompt frameworks for funding transparency and anti-capture protocols - Governance artifacts: Design progressive charter and agreement templates with structured context management ## 🧪 Hypotheses - Personal context curation creates sustainable competitive advantages that resist commoditization by general AI capabilities - Progressive system building (foundational→intermediate→advanced) produces higher adoption rates and better long-term outcomes than complex initial implementations - Context window limitations will drive innovation in selective information architecture rather than expanding memory capacities - Human-AI collaboration systems focused on augmentation outperform automation-focused approaches in knowledge work scenarios ### 📊 Methodology - Longitudinal study comparing professionals using structured context management vs. ad-hoc AI interactions over 6-month period - A/B testing of progressive vs. comprehensive system implementation approaches measuring completion rates and sustained usage - Analysis of context window utilization patterns across different professional domains and task types - Ethnographic research on human-AI collaboration workflows in creative and strategic work environments ### ❓ Questions - How do context curation skills transfer across different AI platforms and evolving model capabilities? - What organizational structures best support collective context management in team environments? - How does the value of personal context change as AI training data becomes more comprehensive and current? ### 🔮 Future Research - Development of automated context relevance scoring and optimization systems - Investigation of collaborative context sharing protocols for team-based AI assistance - Exploration of context compression techniques that preserve essential information while respecting window limitations ### 💡 Implications - Practice: Knowledge workers must develop new competencies in information curation and AI interaction design - Policy: Organizations need guidelines for balancing AI assistance benefits with data privacy and intellectual property concerns - Security: Context management systems require robust safeguards against data leakage and unauthorized access - Education: Curriculum design should follow progressive complexity models rather than attempting comprehensive initial training - Funding: Investment in human-AI collaboration tools should prioritize augmentation capabilities over full automation features ### 🧠 Learning Styles Assessment Reading level estimate: 12-14 (Graduate level due to technical concepts and complex system descriptions) #### 👀 Visual learners - Present: System architecture diagrams, context window visualization, master prompt examples - Improve: Interactive flowcharts showing information flow, visual templates for context organization #### 🔊 Aural learners - Present: Live presentation format with verbal explanations and Q&A interaction - Improve: Audio-only versions of key concepts, podcast-style deep dives on specific techniques #### 🗣️ Verbal learners - Present: Clear definitions of technical terms, analogies (manufacturing line, water treatment plant), progressive explanations - Improve: Glossary of AI terminology, story-driven implementation examples #### 🧩 Kinesthetic learners - Present: Live demonstrations with Claude, step-by-step master prompt creation, hands-on cohort format - Improve: Interactive exercises, template customization workshops, sandbox environments #### 🧠 Logical learners - Present: Systematic three-level progression, clear cause-and-effect relationships, structured methodology - Improve: Decision trees for context selection, quantified metrics for system effectiveness #### 🧑‍🤝‍🧑 Social learners - Present: Community-centric learning approach, peer collaboration emphasis, Q&A interaction - Improve: Structured peer review protocols, collaborative context sharing exercises #### 🧘‍♀️ Solitary learners - Present: Individual master prompt creation, self-directed system building, personal context focus - Improve: Self-assessment checklists, reflection prompts for context curation decisions #### 🌿 Naturalistic learners - Present: Ecosystem thinking about AI tools and workflows, organic system evolution concepts - Improve: Metaphors connecting information management to natural systems and cycles ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading structure, descriptive content organization, logical flow - Improve: Alt text for system diagrams, detailed descriptions of visual demonstrations #### Keyboard navigation - Present: Standard web navigation patterns, accessible video controls - Improve: Skip links for long-form content, keyboard shortcuts for key concepts #### Contrast and readability - --- ## Full Transcript [00:00] Okay. [00:02] So, as I said, I want you to have a [00:04] roadmap, a high-level roadmap of what [00:07] the AI second brain system, the system [00:09] that we talk about looks like. [00:12] And how if if assuming you want to learn [00:14] alongside us, what that might look like, [00:16] how it works. [00:19] All right? [00:22] Um [00:23] I don't have that many slides. I think I [00:24] have like 25 or 26. [00:27] Uh there's going to be plenty of time [00:28] for us to interact. That's kind of one [00:31] of the main benefits of this format. Um [00:34] in the cohort itself, we do breakout [00:35] rooms, we do coaching, we do a lot of [00:37] different things. [00:39] In today's call, just because it's so [00:40] large, it will be mostly questions that [00:42] we will interact with via the chat. [00:45] So, I'll have two separate times for you [00:47] to to ask questions in the chat. You can [00:49] put questions anytime and we'll do our [00:51] best to answer them, but I will [00:52] specifically turn to questions in the [00:54] middle after about um 45 minutes or so, [00:58] and then at the end for the last 15 to [01:01] 20 or 30 minutes. [01:04] Okay? So, save your hardest questions [01:06] for me. [01:08] But first, let's get on the same page. [01:12] >> [clears throat] [01:12] >> So, [01:14] a few kind of observations [01:17] and perspectives here [01:19] uh [01:20] to kind of set the stage, set the [01:21] context. [01:24] I really believe that there's a new [01:27] bottleneck in modern productivity. [01:29] There's a new constraint. [01:33] It's no longer your time. [01:35] So, just just think about [01:37] >> [laughter] [01:38] >> first line. Think about this for a sec. [01:39] Think about your career. [01:42] Think about your life, [01:44] your life experiences, [01:46] and just imagine for a sec, even not [01:48] even to do with AI, just set AI aside. [01:50] Can you even imagine what it means to [01:53] have your time, [01:55] the time you have available not be a [01:57] bottleneck? [02:00] Just just like try to imagine. [02:02] Imagine if you had infinite time. [02:06] If time was not a factor, time was not a [02:08] limitation. [02:09] It's almost as if you could pause time [02:11] for as long as you needed [02:14] and spend as much time as was needed on [02:16] whatever you wanted to create or learn [02:18] or experience. [02:20] Every anywhere you wanted to travel, [02:21] anyone you wanted to meet. [02:24] It's like a really an imagination [02:26] challenge here. [02:29] Imagine what it would be like to have [02:31] your attention not be a bottleneck. [02:34] You know, your attention, your ability [02:35] to focus, [02:37] your attention span, [02:39] your ability to to pay attention to [02:41] things that are hard or complicated or [02:44] boring. [02:46] What would you do with that mind of [02:48] yours, with that creativity, [02:50] if attention was not a bottleneck? [02:55] And then intelligence. What if your [02:56] intelligence even was not a bottleneck [02:59] on what you could create, achieve, [03:01] experience, the impact you could have? [03:04] Not your your natural talents that you [03:07] were born with, your IQ, your EQ even, [03:11] all those natural traits. What if those [03:14] were just not a limitation anymore? [03:20] And imagine or just consider that the [03:23] bottleneck has moved. The bottleneck has [03:25] moved to collecting, [03:28] curating, and providing context, which [03:31] is really just information, [03:33] to increasingly powerful generations of [03:36] artificial intelligence tools. [03:39] Which is a skill. It's not it's not [03:42] obvious how to do this. It's a skill. [03:44] It's a it's a discipline that we're [03:46] calling PCM, personal context [03:48] management. [03:50] A brand new skill in the world that if [03:53] you can master, you can get around this [03:55] bottleneck. [03:59] So, thinking about you as the human in [04:02] all this. [04:03] Don't want to forget about that part. [04:06] Consider that [04:09] your AI, all your different AI tools can [04:11] only use, they can only make use of what [04:14] you give it. [04:16] What you give it. [04:18] Your personal curated context, your [04:21] notes, your knowledge, your expertise, [04:23] your experience, your memories, your [04:25] points of view, your perspectives, your [04:27] relationships, your network, all the [04:29] stuff that makes you you. [04:31] Is now [04:33] not just this, you know, a little weird [04:35] productivity habit you have. It's it's a [04:37] treasure trove. [04:38] It's just an incredibly valuable [04:40] treasure trove of data [04:42] that no one else has, [04:44] no one else can access, no one else can [04:47] compile, no one else can generate. [04:50] That context, I believe, is the only way [04:53] that you escape you escape, you know, [04:57] the disruptive changes that are coming, [04:59] that you escape uh competition, that you [05:02] create a a real competitive advantage [05:04] that doesn't, you know, get eliminated [05:07] with the next AI release. [05:10] Uh I call it the mediocrity trap. [05:13] Everything with AI, you you may have [05:14] noticed, kind of [05:15] pulls you towards the average. [05:18] The way that AI thinks, the way that it [05:19] writes, the way that it [05:22] uh it works through problems. It's [05:23] constantly sort of homogenizing your [05:25] thinking. [05:26] And that's because it's trained on the [05:27] entire internet, everyone else's ideas [05:30] and opinions and everything they've ever [05:31] said. [05:33] You can't stand out. You can't really [05:35] have a competitive advantage if you [05:37] allow yourself to be pulled into that [05:39] that average, that, you know, that that [05:43] that average level of quality, that [05:46] mediocrity. [05:51] I published a video a few months ago. Um [05:54] maybe someone on the team could drop [05:55] that in the chat. [05:56] Which was exploring uh some research [06:01] that has been done on something called [06:02] context rot. [06:04] Okay? So, imagine [06:09] Actually, before I go there, there's [06:10] something called a context window. [06:13] Okay, we're going to start we're going [06:13] to start here at the very, very [06:14] beginning. [06:15] >> [laughter] [06:16] >> I'm not going to assume that you know [06:18] you know anything about how these AI [06:20] platforms, which are known as LLMs, [06:22] large language models, work. [06:24] So, your context window, the AI's [06:26] context window, is how much information [06:29] it can work with. [06:30] You can think of it almost like it's [06:31] working memory or it's RAM. [06:35] Right? So, you can give [06:37] uh an LLM, you know, 10,000 words or [06:40] tokens, 50,000, 100,000, but at a [06:43] certain point, it it quits. At a certain [06:47] point, there's an error. I'm sure you've [06:48] come across this. You have exceeded the [06:49] context window. [06:52] So, [06:54] without getting too technical here, what [06:56] you really need to understand is context [06:58] windows have limits. [07:00] They are not infinite. Not even close. [07:03] Even with the latest AI models that say, [07:06] "Oh, a million token context window, 2 [07:09] million context window." Those are lies. [07:11] >> [laughter] [07:12] >> As this paper [07:13] uh really proves and explains, [07:16] the effective context window, how much [07:18] you can actually give to these AI [07:20] platforms, is really only maybe 25 to [07:23] 50% of what they claim. [07:26] That comes out to something around 30 to [07:28] 50,000 words. [07:31] Okay? So, even if they say a million [07:33] words, 2 million, 5 million words, [07:36] if you exceed around 30 to 50,000 words [07:39] or in AI speak tokens, you start having [07:42] these negative impacts, which is what [07:44] this paper documents. You start having [07:46] context poisoning. [07:48] Context poisoning is when one little [07:50] error, one little mistake, it poisons [07:53] the whole interaction with AI. It's like [07:56] it gets fixated, obsessed with this one [07:59] error, and it can't let it go. I'm sure [08:01] you've experienced this. [08:04] Or this context distraction. If there's [08:06] too much in the context that you give [08:08] the AI that is not relevant to the task [08:10] at hand, it gets distracted, just like [08:12] humans. [08:14] Right? It it gets it gets kind of [08:16] all over the place. It gets fragmented. [08:18] It's it's paying attention to the wrong [08:19] things. It's it's getting distracted, [08:21] just very much like humans do. [08:25] And then there's context confusion. [08:27] These are all kind of failure modes that [08:29] the research has documented. Context [08:31] confusion is when there's things that [08:32] conflict. [08:34] What is the AI supposed to do if you [08:36] have two pieces of information that are [08:39] either inconsistent or they're directly [08:41] contradictory or they're incompatible, [08:44] which I guarantee you will happen the [08:46] more context that you try to give it. [08:49] So, [08:50] setting aside all this this is the most [08:52] technical we're going to get right here. [08:53] >> [laughter] [08:55] >> So, setting aside all this new [08:56] terminology and these numbers, [08:58] that little warning sign down there at [09:00] the bottom right is what I want you to [09:02] take away from this. You cannot just [09:04] dump everything in. [09:06] I wish you could. [09:08] I wish you could just say, "Here, here's [09:09] my hard drives, my computer, my cloud [09:11] drives, here's all my logins. Just [09:13] access it all, make sense of it all, [09:15] organize it all." You really cannot. You [09:17] have to selectively curate the right I [09:20] think of them as bundles of context. [09:24] The minimum amount of context that you [09:26] need to accomplish the task at hand. [09:30] And that might seem like just a random, [09:32] you know, it's kind of technical detail, [09:34] but it's huge, you guys. [09:36] It's so important because it means that [09:39] all the stuff that we [09:42] that we had to do and that we learned [09:43] pre-AI, how to organize, how to curate, [09:45] how to select, how to take notes, how to [09:48] structure, all those things are still [09:50] important. [09:51] They're even more important because you [09:53] have to do them to take advantage of [09:55] AI's power. [10:00] So, zooming out for a second, what is [10:03] this AI second brain? It's the same as [10:05] it always was. [10:07] >> [laughter] [10:07] >> Going back with my work over 10 years [10:10] now, [10:11] and really if you look at the history of [10:12] what we call PKM, that's personal [10:15] knowledge management, goes back decades. [10:17] Uh decades to the [10:19] early years of the 20th century. [10:22] It's a system designed to amplify you. [10:26] It's a system that lives outside your [10:28] head, outside your body, that is [10:30] designed to amplify you. [10:33] To unlock your creative potential. [10:36] And ultimately, at the grandest scale, [10:40] to unlock your ability to lead a [10:42] thriving generative life. [10:45] So, we're using generative AI to lead a [10:47] generative life. [10:49] >> [laughter] [10:50] >> And [10:52] this graphic on the right [10:54] is kind of the the 50,000-ft view of how [10:57] we're going to do this. [10:58] It's been amazing to me. It's been so [11:00] surprising that [11:02] all the the the techniques and the [11:05] skills that came pre-AI, not only are [11:08] still relevant, like I said, they're [11:10] more important than ever. So, at the [11:11] bottom there, we have code. That is the [11:13] core method that I talked about in my [11:17] book Building a Second Brain. [11:19] Okay, so that really is centering your [11:21] creative process, your productive [11:24] process. And getting information in, [11:26] which is the C, capture, organizing it, [11:29] that's the O, [11:31] distilling it, and then expressing it. [11:33] That has to be the foundation. [11:35] Ultimately, you're trying to put [11:37] something out there. [11:38] To publish something, communicate [11:40] something, launch something, sell [11:42] something. There's an expression [11:44] at the end. [11:47] To do that, [11:48] we build this new layer on top, which is [11:51] PARA, [11:52] which is a way to organize your entire [11:54] digital life that was explained in my [11:56] book The PARA Method. [11:58] Because you need certain kinds of [11:59] information to express whatever you're [12:01] trying to express. [12:02] You have projects, that's the P. You [12:04] have areas of responsibility both at [12:06] work and in your personal life, that's [12:08] the A. [12:09] You have a wide variety of resources, [12:11] that's the R. [12:12] And you have archives, that's the second [12:14] A. [12:16] Okay. So, those two methodologies, CODE [12:19] and PARA, at this point are very very [12:22] very well distributed. [12:25] Um in fact, [12:26] we'll send you some links. Like there's [12:28] the books themselves, but if you don't [12:29] want to read a book, there's YouTube [12:31] videos, there's the podcast, there's [12:32] many blog posts. You can have at it. In [12:35] whatever format you like to learn, there [12:37] is a way for you to learn those two [12:38] methodologies. [12:41] But we're adding a new one. [12:43] Right? We're adding this new layer that [12:45] is building on. It is extending those [12:49] pre-AI capabilities, but with this new [12:51] generation of technology that we call [12:53] AI. [12:56] All right? [12:58] So, let's zoom in a bit. Let's get into [13:01] it. How how does this AI-powered second [13:03] brain work? [13:06] So, I always like to think of like a [13:08] system, like a like almost like a [13:10] mechanical system. Like if this was a a [13:12] manufacturing line, or if it was a, you [13:15] know, a water treatment plant. Like how [13:17] does the like the actual bits of [13:20] information flow from one place to [13:22] another, and how does that add value to [13:24] my life? [13:27] So, the four main parts that we're going [13:28] to talk about in the program that we're [13:30] teaching, and this is also going to find [13:32] its way into everything else that we do, [13:35] is in four parts. So, the first one, as [13:38] ever, as always, is capture. [13:42] You need a way to capture information. [13:45] Right? That's that's how you that that [13:48] that's where it all starts. [13:50] Capture notes, capture ideas, capture [13:52] insights, highlights from your reading. [13:55] Now, there's new ways to do that with [13:56] AI. You can have an AI note-taker now, [13:58] you can do voice transcription. There's [14:00] a lot of really powerful AI capture [14:03] tools, but you still need to think about [14:04] it. [14:05] There's still some decisions to be made. [14:07] How do I want to capture information? [14:10] What is the way of capturing that fits [14:12] me? [14:15] Once you've captured it, as before, this [14:18] part also hasn't changed, is you want to [14:20] organize it. [14:21] And that's where PARA comes in. [14:24] Okay? And the key thing with PARA that [14:26] I'll point out here, there's there's a [14:27] lot to say here, but [14:29] you need an organizational system that [14:32] makes sense for both humans and AI. This [14:35] is the key point, you guys. You can have [14:37] AI create a completely custom, bespoke [14:40] organizational system that perfectly [14:42] fits how it thinks, but if it doesn't [14:44] fit how you think, then you've just [14:46] created a system that you can't actually [14:48] make use of. [14:51] Right? So, so PARA not only is, you [14:53] know, really well proven and well [14:55] validated, it's simple enough and it's [14:58] straightforward enough that AIs can [15:01] understand it perfectly, and so can [15:03] humans. [15:04] That's the key the key criteria. [15:08] Now, this is where it changes. Okay? [15:10] Instead of going to distillation, which [15:12] is a which was quite a manual, you know, [15:16] labor-intensive process, [15:18] the information in PARA that you've [15:20] organized now flows into what we call a [15:22] master prompt. [15:25] Okay, I'm going to say a lot more about [15:27] that in just a minute. But it's [15:28] basically that curated context that you [15:31] provide to AI that I was mentioning [15:33] before. [15:36] Then [15:38] the that master prompt, or those master [15:40] prompts plural, [15:42] flow into and inform what I think of as [15:44] a set of AI advisors, like an AI board [15:47] of advisors. Imagine if you had a coach [15:50] and a business consultant and a [15:52] strategic analyst, and then also a [15:56] relationship coach and a health coach, [15:59] and, you know, all the all the people [16:01] that if you had unlimited money, you [16:02] would hire to help you, you can now have [16:05] with AI. But it requires some design. [16:08] Requires a little bit of of uh of, you [16:10] know, of designing. [16:15] All right? [16:17] So, I want to say a little bit of more [16:18] about the master prompt, because this is [16:20] really [16:22] this is really kind of the core thing. [16:25] What is a master prompt? [16:28] You can think of a master prompt as [16:31] your personal operating manual for AI, [16:34] for all AI. All the different You [16:36] probably have half a dozen or a dozen [16:38] different AI tools you work with. So, [16:40] this is a completely cross-platform, [16:43] platform-agnostic [16:45] way of working with AI. It's not limited [16:47] to any one platform. [16:50] Okay? So, imagine just a and I'll show [16:52] you mine in a second, but uh imagine [16:55] just a document. Like any document, a [16:58] TXT document, a markdown document, a [17:01] Word document if you want, a Google Doc. [17:03] Just a single long document that has [17:06] some of these little bubbles that you [17:08] see around the outside. [17:12] So, as I read each of these, imagine [17:14] your interactions, whether it's with [17:16] Claude or Gemini or ChatGPT or Notebook [17:20] LM or Perplexity or whatever it is. [17:23] Imagine if that AI knew, say, your basic [17:27] personal details. [17:29] Okay, I think this [17:31] You guys, this is this is this is so [17:33] fundamental. I feel like this needs to [17:35] be taught in schools. I I every single [17:38] person in the world should do this. In [17:39] fact, this is why we we we published a [17:42] video on the master prompt uh in the [17:44] summer. It was one of our most popular [17:46] videos of the last year. In fact, that [17:47] became a whole series. Someone in the in [17:50] the uh from the team, if you could share [17:52] that in the chat, that would be great. I [17:53] decided to just give that content away [17:55] freely to the world, because I think [17:57] it's just it's like basic fluency for [18:00] modern life at this point. But imagine [18:02] if it knew your basic personal details. [18:03] Who you are? What's your name? How old [18:06] are you? Where do you live? [18:08] Who's in your family? [18:10] Imagine if it had your basic [18:11] professional info. What's your job? [18:14] What's the title? What's the role? What [18:16] team are you on? Who do you report to? [18:17] Who do who who reports to you? [18:21] Imagine if it had go just going around [18:23] clockwise around the circle, your income [18:25] sources. Do you ever ask AI for any [18:27] advice or have any questions for it [18:30] related to money? [18:31] Related to your finances or related to [18:34] getting a promotion or related to how [18:36] you invest or how you save? Well, [18:38] wouldn't it be useful if it knew how [18:40] much money you make every year? [18:43] Or if it knew what your investments are [18:45] currently? [18:46] Or if it knew your risk tolerance for [18:48] your finances? Or if it knew how much [18:49] savings you have? You have to provide [18:52] this information. [18:55] And I'll just go quickly. Like [18:57] AI needs to know your core values. I'm [18:59] not talking about company. I mean you as [19:00] an individual. What are your core values [19:03] in life? What is your vision? What is [19:05] your mission? [19:07] What are your intentions for AI usage? [19:09] How What are you trying to accomplish in [19:11] your usage of AI? [19:13] What are your communication preferences? [19:14] How do you like to communicate and [19:16] collaborate? How do you not like to? [19:18] Your working style, your your tool [19:20] preferences, your workflows, your rules, [19:21] your boundaries, your goals and [19:23] intentions. Goals and intentions are [19:24] twice, cuz they're so important. [19:27] Your personal SWOT, your strengths, your [19:29] weaknesses, your opportunities, your [19:30] threats, your skills, what quality looks [19:32] like to you. Okay? I know this might [19:34] seem like a lot. [19:37] Like when I talk to people, even people [19:39] who are very deep in AI, who have very [19:41] sophisticated systems, I 99% of the time [19:44] when I ask them if they have this this [19:47] this kind of information provided to [19:49] their AI tool of choice, the answer's is [19:51] I I kind of can't believe that after all [19:54] this time, we still largely don't do [19:56] this. [20:00] All right. [20:02] So, let me show you. [20:22] All right. This is my Claude. [20:26] This is the desktop app for Mac. [20:31] And what you're looking at, let's just [20:33] make this very simple. So, I'm clicking [20:35] over here on the left where it says [20:38] projects. See that? [20:40] A project is simply a little [20:42] self-contained, almost think of it like [20:44] a folder, [20:46] that is kind of self-contained. It's [20:47] limited. Everything that happens within [20:49] that little container is limited to that [20:51] container. [20:52] Okay, so I have one for the book I'm [20:53] writing, for the upcoming cohort, for [20:56] the company as a whole, for my writing, [20:57] for a different program that we run. [21:00] Each little self-contained box. [21:04] So, the one I want to go into is Tiago's [21:06] work master prompt. I have a different [21:08] one for my personal life cuz I don't [21:09] want them to bleed over into each other. [21:11] But if I click into that, this is the [21:13] project interface. I'll go ahead and [21:17] close the sidebar. [21:19] Okay? [21:20] And the main thing I want you to notice [21:22] is this little box over here that says [21:24] files. This is that curated context that [21:28] I was talking about. [21:30] They're also known as just project [21:31] files. [21:32] Okay? [21:34] And there's only one file. You can have [21:36] any number of things. You can see this [21:38] one document is only using 1% of my [21:40] project capacity, so that's good to [21:42] know. I could add lots of other things. [21:45] But if you click on this, [21:47] you'll see my work master prompt. [21:51] All right, so you have [21:53] basic personal details about me, where I [21:55] live, company information, our AI [21:57] strategy, our ideal customer profile, [22:01] our core products, um [22:04] that we offer, the company mission, the [22:06] company vision, the core values of the [22:08] company, the business strengths. A lot [22:10] of this is very business-oriented, but I [22:12] have something similar with my personal [22:14] master prompt. I have the weaknesses of [22:17] the business. [22:18] Uh examples of our brand voice, [22:21] uh the org chart, [22:24] uh some of our internal processes, our [22:26] strategic and operational process [22:29] architecture, our core value chain, how [22:30] we serve customers, etc., etc., etc. [22:34] All right? So, [22:36] with that master prompt, that's what [22:38] this is here as context, [22:43] I can ask it something like, [22:47] "With everything that you know about me, [22:49] what is the biggest single gap in our [22:51] product offerings, and what would you [22:53] propose in order to address it?" [22:57] So, the reason I could speak that, by [22:59] the way, is I use an app called [23:00] WhisperFlow, which allows me to just [23:02] hold down a key on my keyboard and [23:04] speak. Well, that's one of the capture [23:06] tools that we'll show you how to use. [23:28] And notice, if you look along the top [23:29] here, there's three modes that you can [23:31] use Claude in, chat, co-work, and code. [23:34] I'm using chat. [23:35] So, this is the this is the most basic, [23:37] the most beginner level, the most [23:40] elementary [23:42] um version of Claude. [23:45] If you want more power or you need more [23:47] capabilities, you can use the others. [23:50] Okay, interesting. So, it has a strong [23:52] opinion. [23:53] Notice that there's no way for it to [23:55] have a strong opinion or for its strong [23:56] opinion to be of any use to me if it [23:58] doesn't have a tremendous amount of [24:00] context. Think how much context you need [24:01] to answer this question. [24:03] You got to be able to holistically see [24:05] the entire business plus my role in that [24:08] business. [24:09] And here's the gap that it sees. You [24:11] have no scalable high-touch offering [24:12] that doesn't require Tiago's direct [24:14] involvement. And I think that I mean, [24:16] that's exactly right. It's so true. [24:19] And then it makes a proposal here. [24:22] Okay, so notice that it knows my product [24:24] ladder, how people kind of ascend [24:26] through our ecosystem. It knows that I'm [24:28] involved, highly involved in pretty much [24:30] all of them, especially the high-touch [24:32] ones. [24:34] Um it can it knows that this is [24:36] capitalizing on our three biggest [24:37] strengths. [24:39] Uh it knows that we previously tried to [24:41] move into B2B offerings and it didn't [24:43] work, and it even knows why. [24:46] Right? And then it has follow-up [24:47] questions. And they're not just general [24:50] vague ones, they're really specific. For [24:51] example, "What's your wife up to?" The [24:53] fact that it knows I'm married, knows [24:55] her name, knows what she's up to, knows [24:58] that she's not currently involved in the [24:59] business. This isn't a long response. [25:02] Look, this isn't one of those like mega [25:04] five, you know, a hundred or a thousand [25:07] word things that I have to pour through. [25:09] It's just a very succinct and very [25:11] targeted kind of observation, the way [25:13] that a coach would do, pointing out [25:15] what's missing in my work currently. [25:18] And I think it's quite insightful and [25:20] quite accurate. [25:24] Okay? [25:25] And the other thing I want to show you [25:27] before we move on is [25:31] uh let's see. [25:35] Notice that this was updated today. See [25:38] 326? [25:39] I updated it this morning. [25:41] And the way that I updated that, [25:43] I'm not going to get into this now, [25:46] but every once in a while, every week or [25:47] two weeks or three weeks or whenever, I [25:49] give it access to my notes and I say, [25:52] "Based on all of my latest notes, across [25:54] all of my current projects, all the [25:56] areas of responsibility, all the [25:58] resources, if there's any new ones or [26:00] any new archives, update my master [26:02] prompt." [26:04] And it did that. [26:05] In a few minutes. I think I told it to [26:07] do that. I went and heated up my coffee, [26:09] came back, and it was done. [26:13] Do you guys understand what is going on [26:16] here? [26:17] >> [laughter] [26:18] >> Like this is so insane. This is so [26:20] insane that I can just take notes in my [26:22] casual, free-form, kind of messy way, [26:25] and then once a week or every two weeks [26:26] say, "Hey, Claude code." In this case, I [26:28] used Claude code. [26:30] "Update my master prompt." And it went [26:31] ahead and did that. [26:34] And and that's just like the most basic [26:36] level. That's just like kind of beginner [26:38] or what we call the foundational level [26:40] of what this looks like. [26:45] All right? [26:47] I don't know about you, but my my mind [26:48] is still blown by this. [26:51] >> [sighs and gasps] [26:52] >> Okay, I can sense those questions. [26:55] I can feel the the the tidal wave of [26:58] doubts and worries and follow-ups, and I [27:01] want to get to those. [27:03] Um I think there's a couple more things [27:05] I want to share with you, and then we'll [27:07] pause. [27:10] So, [27:12] how do I recommend you go about creating [27:14] this thing called a master prompt? [27:17] I have a very strong opinion on this. [27:20] Okay? [27:21] And this really speaks to what I think [27:24] I've done so much reflection on [27:26] the, you know, the decade that we were [27:28] teaching Building a Second Brain, the [27:30] the previous, you know, the previous [27:32] ecosystem, the previous method. [27:35] Uh we taught it in so many different [27:36] formats and so many different ways. [27:38] And just reflecting on all that [27:40] experience, I think there was a big [27:41] mistake that I made. [27:44] There was like there was a variety of [27:45] mistakes, but this was the biggest one. [27:47] Which is [27:49] I sort of imagine the ideal system. [27:52] What is the ultimate way this looks? [27:54] Which for me took years and years and [27:56] many iterations, many versions. And then [27:59] I said, "Let me teach this ultimate [28:01] system to everyone." [28:04] And that had, you know, some benefits. [28:06] It was it was impressive. It was [28:08] inspiring. Kind of got you excited. [28:11] But I think the problem with that [28:12] approach is that the fully mature [28:14] system, [28:16] as cool as it is, is not how you start. [28:19] That's that that can't be how you start. [28:21] That's the worst way to start. That's [28:24] trying to, you know, to swallow [28:26] something that you're not ready for. [28:28] It's trying to build something that is, [28:30] you know, on version 15 when you really [28:33] need to need to start with version one. [28:36] So, in this cohort, in this program, I'm [28:39] going to try to rectify that mistake. [28:43] And the way we're going to do that is [28:45] we're going to build the same system [28:47] three times. [28:50] Okay? So, week one, we're going to spend [28:53] week one at the foundational level, just [28:55] creating the basic master prompt. [28:58] The the version that I really believe is [28:59] like the MVP, the minimum viable [29:02] product, minimum viable prompt. [29:05] Like what every single professional, no [29:07] matter what you do, needs to have. We're [29:09] going to do that in week one. So, if you [29:11] just do week one and then you get busy [29:13] or you drop off or that's enough for [29:14] you, great. You got the the 80/20 of [29:17] what I have to offer. [29:19] Okay? [29:20] And the [29:21] the other thing that's important here is [29:24] even for those of you who are more [29:26] advanced, and many of you are, if you [29:28] think about it, most of your AI [29:29] interactions still reside at at kind of [29:32] a basic level. [29:34] Right? You don't need to fire up the [29:35] most advanced system. You don't need to [29:37] fire up Claude code for probably, I [29:40] don't know, 50, 60, 70% of your AI [29:42] interactions. So, even for advanced [29:44] users, you want almost like a fallback. [29:47] You want something that is that is [29:49] small, simple, and minimal, and [29:51] streamlined that you can fall back to [29:53] for those kind of simple everyday [29:55] interactions, and that's the [29:56] foundational level. [29:59] Then, we're going to [30:01] do it again. [30:02] We're going to [30:04] sort of rebuild the system [30:07] in week two at an intermediate level. [30:10] Because what you're going to find is for [30:11] many tasks, that foundational level is [30:14] not enough. And And it's it's funny. I [30:16] actually want you I want you to run into [30:18] those barriers yourself. [30:20] I want you to have the experience of [30:22] what it's like having a more basic [30:23] master prompt and then running up [30:26] against the limits of it, the places [30:27] where it doesn't work. [30:29] And there's so many. [30:30] >> [laughter] [30:31] >> So then, you feel the need like [30:33] viscerally, like in your gut, you feel [30:36] the need for what we're going to do at [30:37] the intermediate level. [30:39] Which is basically to break apart your [30:41] master prompt into little modular [30:43] bundles. [30:46] Okay, I know that sounds kind of [30:47] you know, complicated, but it just [30:49] basically means imagine your master [30:51] prompt as a whole whole collection of [30:54] different documents. [30:56] And you only load up the ones that you [30:58] need when you need them. [31:00] So imagine you have a master prompt just [31:01] for one project. [31:03] If you're working on that project, you [31:05] only load up that master prompt and [31:06] nothing else. You have a master prompt [31:08] for your finances. You have a master [31:10] prompt for your relationship. You have [31:12] one for your health. You have one for [31:15] your vacations. You have like for each [31:17] little part of your life and your work, [31:19] you have like almost like a little [31:20] specialized master prompt so that you [31:22] avoid kind of overloading it and [31:24] confusing it with too much random [31:26] context. [31:28] Okay? [31:29] Then, [31:30] in week three, we're going to do it [31:31] again. [31:33] Because the intermediate level is also [31:36] it's not that it's not enough, it's that [31:37] it has limitations. There's always [31:39] limitations. [31:41] So, [31:41] the important point here is it is only [31:44] at the advanced level that we're going [31:46] to layer in all the stuff, [31:49] the crazy, complicated, impressive, [31:52] sophisticated stuff that makes up 90% of [31:55] what you see on social media. [31:57] Okay, the long-term memory where the the [32:00] AI can actually build and keep track of [32:02] its own memory itself, [32:04] agents and sub-agents, integrations with [32:07] other platforms, advanced harnesses, all [32:09] this stuff that's going on. I really [32:12] strongly believe should only come at the [32:14] advanced level when you have the the [32:17] previous layers, the more foundational [32:20] levels in place. [32:24] All right. [32:27] So, [32:30] I think I'm going to pause here. [32:34] Okay, that was that was a lot already. [32:36] >> [laughter] [32:37] [gasps] [32:38] >> Um [32:40] I'm just opening the chat now. [32:42] Team, do you have any [32:44] questions or comments you want to [32:45] surface for me? [32:47] Yeah, for sure. Definitely seeing a [32:49] theme around privacy and security when [32:52] it comes to the master prompt. This is a [32:54] question that that popped up a lot. So, [32:57] how do you think about that? Like giving [32:59] more sort of like sensitive personal [33:01] data to the LLM or putting it into your [33:03] master prompt? [33:06] This is a whole conversation. [33:09] >> [laughter] [33:10] [sighs and gasps] [33:10] >> It's a whole conversation. We're We [33:12] We're going to have not an entire live [33:14] session, but a big chunk of one towards [33:17] the end, I believe, about this because [33:18] it's very nuanced, you guys. [33:21] It's very nuanced. Don't don't think [33:23] there's a single like the single answers [33:26] that you get these days are very black [33:27] and white. They're very all or nothing. [33:30] Give AI everything. It's no problem. [33:33] You know, YOLO. [33:35] And the other extreme is no, AI is [33:37] inherently evil. It's stealing all your [33:39] data. [33:40] Pretty sure there's some middle ground [33:42] there. [33:42] >> [laughter] [33:43] >> Um and it it it has implications. Has [33:45] implications for which tools you use in [33:47] the first place, for how you use those [33:49] tools, for what you provide to those [33:50] tools, for the permissions you give at [33:53] different levels. There There's This is [33:54] a world. [33:55] And what we will talk about it for sure, [33:57] but it's nuanced. [33:59] I would say it's If I had If I had to [34:01] say one answer, it's just a trade-off [34:03] like everything. It's a trade-off [34:04] between getting the power and the [34:06] leverage of AI without getting the risk [34:10] and the pitfalls and the dangers of of [34:14] AI. [34:18] Yeah, I like how Muhammad just put it in [34:20] the in the chat. Security in the AI era [34:22] is a calculated risk. [34:25] Yeah. [34:26] Like everything. I mean, everything is a [34:28] risk. Getting in your car, driving down [34:29] the freeway. [34:30] Probably, you know, one of the biggest [34:32] risks you take. They're They're all [34:33] calculated. They're mitigated. They're [34:36] They're counterbalanced. [34:42] Yeah, anything else? [34:49] Hey, should we get to some questions [34:50] from the Q&A? [34:54] Yeah, would you mind surfacing those for [34:56] me cuz the [34:57] the chat is pretty crazy right now. [35:01] >> [gasps and sighs] [35:02] >> All right. [35:05] Let's pick this one from from Julie. I'm [35:08] building my second brain in Notion and [35:10] Claude. [35:11] Um I bought the Notion power template. [35:14] Personally working with Claude directly. [35:16] Do you have Claude skills you're [35:17] building or endorsing? How will the [35:19] cohort maybe cover skills? [35:22] Yes, we will talk about skills. That's [35:24] one of those I'd say intermediate to [35:26] advanced things. So what you're going to [35:28] notice is that when you build your [35:29] foundational master prompt, you're going [35:32] to put stuff in there that later on I'm [35:34] going to have you look look through it [35:36] or have your AI look through it, and [35:37] you're going to notice some of the stuff [35:38] you're putting in your master prompt [35:39] shouldn't be there. It should be skills. [35:42] If it's like something like in in this [35:45] case, do that. [35:47] Or in these situations, apply this this [35:50] lens. Those things that that are sort of [35:52] triggered. [35:53] Uh I mean, you can you can create [35:55] something else that's called a hook. [35:57] This is This is advanced, but you can [35:58] create a hook or a skill. It's It's [36:00] almost like what was originally just the [36:02] single monolithic master prompt is now [36:04] being pulled apart into different [36:07] specialized, purpose-built [36:10] tools. [36:12] But I want you to start with the master [36:13] prompt. Just get it all in one place [36:14] because AI can do a lot of that for you. [36:17] It can tell you what things should be [36:19] and where they go and what format they [36:20] should be in. [36:22] Yeah, skill skills are [36:24] of of the advanced things we're going to [36:26] look at, I think is So, one of the [36:28] criteria that I have is just what's [36:29] mature. [36:31] I'm only going to teach you guys things [36:32] that are mature, that are proven. [36:34] Uh because there's this wild crazy [36:36] frontier right now, like like open [36:38] Claude, for example. We're not going to [36:39] do anything related to open Claude [36:41] because it is a circus. It is a wild [36:43] west. I have poured so many hours with [36:46] zero return on that thing. [36:48] >> [laughter] [36:49] >> It's not ready. You know, maybe later it [36:51] will be ready, but we we're sticking [36:53] just to Claude code, which has been out [36:54] for more than a year now. That is [36:56] mature. [36:57] And then we'll revisit what's ready to [36:59] to look at in the in the fall. [37:03] Let's go to a comment from Andy who is [37:06] on answering [37:08] asking from YouTube [37:09] about your experience on letting Claude [37:12] update the master prompt on its own. [37:15] And how it might understand your [37:17] intentions, update itself. Um how are [37:20] you making sure that the updates are [37:21] happening actually happening? Would you [37:23] like read the whole thing again? [37:26] Yes, that's that's probably the very [37:28] first thing that will break in your [37:29] master prompt. It will go outdated [37:31] within like 48 hours. [37:33] >> [laughter] [37:34] >> Because that's the pace of modern life. [37:36] That's That's how fast things move. [37:39] And then we're going to have a [37:39] conversation cuz there's there's [37:41] different approaches to this, you guys. [37:43] There's no This is one of our guiding [37:45] principles, in fact. There's never just [37:47] one right way. I I I wish there was. I [37:49] wish I could just tell you the best way. [37:51] But even something as simple as updating [37:53] your master prompt, [37:54] there's sort of a spectrum, which I [37:56] think this person was kind of alluding [37:57] to, from more automatic ways. [38:00] Like you could really say, just go ahead [38:01] and update it. I don't even need to know [38:02] about it. Just do that every hour. But [38:04] then, there's that that risk trade-off. [38:07] It could put something in there that is [38:08] not correct or is not accurate. And then [38:12] on the other end of the spectrum is full [38:14] human updating, which again has certain [38:16] advantages and disadvantages. I tend to [38:18] be in the middle, which is go ahead and [38:20] propose a set of changes, flag the most [38:23] high-stakes one for me so that I can say [38:25] yes or no or provide input input, and [38:28] then once I do, go ahead and do them. [38:34] Uh similar question from Elise and Megan [38:37] regarding work versus personal. Would [38:40] you recommend different, for example, [38:42] Claude accounts for each? Do you put [38:45] them together? How should How do you [38:47] think about the separation? [38:50] Yeah, again, this is this going to be my [38:52] stock answer for everything, but this is [38:53] a conversation. I think you need some [38:56] sort of separation. [38:58] Some sort of separation, but you can do [39:00] it. So, we're going to look at something [39:02] which is the context stack. Imagine a [39:04] pyramid [39:05] where there's different layers of [39:08] context. The ones that are more [39:09] foundational influence all of your AI [39:11] interactions. [39:13] So think just just think for a moment, [39:15] what is so inherent to you and universal [39:18] to to everything that you do that you [39:20] want it to influence every single AI [39:22] interaction that you have. It's actually [39:24] not that many things. [39:26] Right? That's That's quite a low level, [39:28] but then there's higher levels where it [39:30] starts to become something that you can [39:31] turn on or off. [39:33] Um I typically choose to separate it at [39:37] the level of the master prompt. I have a [39:38] master prompt for work and a different [39:40] one for personal. [39:41] And that's a pretty clean separation. [39:44] Um but then I have both of those in the [39:46] same cloud account which is tied to my [39:48] work email. So, there's always a little [39:50] bit of bleeding in and out or some [39:53] overlap. [39:58] Okay, I think I'm going to keep going. [40:01] Um [40:07] Okay, so that was that was a little bit [40:08] in the weeds, right? [40:11] Uh I kind of like to zoom in and out. [40:12] Always keep keep in mind the big picture [40:15] and not get lost. [40:17] And so I want to zoom out. [40:20] >> [clears throat] [40:20] >> I want to zoom out because [40:24] it's sort of like one of the the [40:26] questions that I I would love for you to [40:28] answer out of this this this call [40:32] is whether we're the right people for [40:33] you to learn from. [40:35] Right? I don't assume by any means that [40:37] I am the right person, that our team is [40:40] the right person for everyone. Right? [40:42] You You really want to find the teacher [40:44] or the community or the method or the [40:47] philosophy that fits you and that you're [40:49] going to be able to have confidence in. [40:51] So, to help you do that, I want to just [40:53] give you I think there's five of them. [40:55] Like my five most contrarian opinions. [40:59] They're contrarian in the sense that [41:01] they're the most different from [41:02] everything that I see online. Everything [41:04] I see on X, on Instagram, on LinkedIn, [41:06] on Substack. [41:08] Right? So, to the extent that you either [41:10] agree with these contrarian opinions or [41:12] not even necessarily agree but are [41:14] willing to entertain them, willing to [41:16] consider them, then I'm probably the [41:18] right teacher for you. [41:20] All right? [41:22] So, the first one is this stuff is hard. [41:26] This is like maybe the most contrarian [41:27] thing. Every day I see [41:30] AI influencer after AI influencer just [41:33] either say or imply that it's easy, it's [41:35] seamless, it takes seconds, it's it's [41:38] you know, you could do this if you're [41:39] brain dead, it takes no effort. And then [41:42] every time I take what they're telling [41:44] me and I go and just try it [41:46] virtually all the time I find in my [41:49] experience and I'm you know, fairly [41:51] competent [41:53] uh it just requires so much more [41:55] complexity than they say. [41:57] So much more subtlety is involved in the [41:59] decisions. There's so much more setup [42:01] than they, you know, set the [42:03] expectations for and there's so much [42:05] more maintenance. [42:06] So, I just want to be very honest with [42:08] this that this isn't this like magic [42:10] button that you press and life becomes [42:13] easier, your work becomes easy. It's [42:15] it's hard work [42:17] and it takes it it's in in a weird way [42:19] AI takes more cognitive effort. Like [42:22] like it really burns through your energy [42:25] to to work with AI because you're making [42:27] these these kind of like architectural [42:29] decisions that have a lot of, you know, [42:30] downstream implications. [42:33] So, I'm never going to pretend or say [42:35] that this is easy. [42:38] Okay? [42:42] Hm. [42:46] Number two, um I see a lot of commentary [42:49] out there that oh, you never need to [42:51] take notes. Just have your AI note taker [42:53] do everything for you and just, you [42:55] know, dump everything on all your drives [42:57] and all of your your place. No. Note [42:59] taking is more valuable than ever, you [43:00] guys. You still it still requires human [43:03] discernment and human judgment [43:05] to pick out the signal in the noise. [43:08] What of all this stuff matters? [43:11] What is the important factor, the [43:14] important priority? What's the point of [43:16] view that I'm taking? [43:18] At the same time the [43:20] the value and the purpose of note taking [43:22] is shifting. So, [43:23] there's this idea from from finance, [43:25] from investing called alpha. [43:28] Alpha is a kind of information that [43:30] other people don't have. Right? So, if [43:32] you were a Wall Street investor, you you [43:34] know, maybe you you travel out there to [43:36] like Minnesota and you find out that [43:38] there's like a [43:39] you know, a pest that is eating the the [43:41] corn fields and you have alpha. You have [43:43] this unique piece of information that is [43:45] not out there in the public and you can [43:48] use that alpha to invest, to make [43:50] decisions that give you a better return. [43:52] That same idea is coming to all [43:55] knowledge work. [43:56] There are [43:58] pieces of information that are not known [44:00] to AI. [44:01] Information that is only, you know, in [44:04] certain places at certain times that is [44:06] in certain people's heads, that is in [44:08] old books that have not been digitized, [44:10] that is in certain formats such as video [44:13] and audio that are harder for the AI to [44:16] ingest. There's all and there's hundreds [44:18] of them, hundreds of little pockets of [44:20] information [44:21] of alpha that if you can find them and [44:24] capture them in your notes, you now [44:25] actually have kind of a disproportionate [44:28] advantage over everyone else who is only [44:31] using what the AI already knows and what [44:32] it already has in its training data. [44:35] Okay, so we are not leaving note taking [44:37] behind. [44:40] Number three, I I am of the opinion [44:43] that [44:45] agents are largely a bunch of BS. [44:48] >> [laughter] [44:49] [gasps] [44:49] >> Um they're so not ready. They're so not [44:52] ready. I've I've done so many tests and [44:55] so many experiments [44:56] and I think basically that the the [44:58] online hype cycle is putting just far [45:01] too much attention on automation and [45:03] execution. [45:06] Those are the buzz buzzwords right now. [45:08] AI to automate your life, AI to execute [45:10] for you, to, you know, do your to-do [45:13] list for you. [45:14] And I think that's just so misguided. [45:15] It's it's not true. They they actually [45:18] have tons of bugs. They constantly fail. [45:21] They're constantly making mistakes. And [45:23] in a way it's it's not even the right [45:25] place to focus because [45:27] the simplest bits of execution AI cannot [45:30] do. The other day I asked it to just [45:32] look at my calendar. [45:33] Like the simplest thing. Hey, go and [45:35] look at my calendar calendar and tell me [45:37] what I'm doing tomorrow. Well, it [45:38] immediately deleted a very important [45:40] meeting that had 30 people invited to it [45:43] and then it caught it and said, oh, [45:45] sorry, I deleted that meeting. Let me [45:46] undo it. And it said, I undid it and it [45:48] didn't. [45:49] It's like execution is not the right [45:51] place for AI to focus. What AI can do, [45:54] what it's more than ready to do is [45:56] judgment [45:57] or rather helping you with your [45:59] judgment. It's decision-making, making [46:02] better decisions. It's researching. It's [46:04] it's synthesizing research. It's finding [46:07] the gaps in research. [46:10] Right? That's the realm of pure [46:11] information. [46:13] The realm of pure information where it [46:14] doesn't have to interact with anyone or [46:16] anything outside of your computer [46:19] which is where it starts to go wrong. [46:21] Okay, so we are not going to be talking [46:23] about any kind of agent. We are talking [46:25] about a cognitive exoskeleton, AI as an [46:28] enhancement and as an amplifier to you, [46:31] the human. [46:34] Contrarian opinion number four is I [46:36] really think [46:38] you know, we talk about the frontier of [46:39] AI. I really think the true frontier is [46:41] inside our minds and bodies. [46:45] Um and we're going to have a series of [46:46] guest instructors that are all related [46:48] to this. [46:49] Uh this is it, you guys. It's not about [46:51] the tools. It's not about the features. [46:52] It's not about the technology. It's [46:54] about do you have self-awareness? [46:58] Do you know yourself? [47:00] Do you know how your mind works? [47:03] Do you have emotional fluidity? [47:05] Do you have the ability to shift between [47:06] emotions? The ability to regulate [47:09] yourself? The ability to rest [47:11] and recover? [47:13] Okay, this [47:14] over the long term, maybe not on the [47:16] order of weeks and months but on the [47:17] order of months and years, this is [47:19] what's going to determine [47:21] if AI is a force for good in your life [47:24] or if it's a force [47:26] or if it's a negative force. [47:31] >> [clears throat] [47:31] >> And fifth, this might be the most [47:33] controversial but I absolutely refuse [47:35] the urgency. [47:37] I just think it's so it's so misguided. [47:41] All this thinking, this apocalyptic [47:43] thinking that you have 3 months, you [47:46] have 6 months, you have a year, you have [47:47] 18 months to, you know, make all the [47:50] money you can before our world is over [47:52] and you join the permanent underclass. [47:55] I've been alive long enough that I have [47:56] seen multiple technology waves and this [47:58] is what they always say. [48:00] Every single new technology wave they [48:02] say it's all over and you have to do A, [48:05] B, and C right now. I I think typically [48:07] when they do that [48:08] it's it's to get you to do something. [48:11] Right? It's to get you to panic, give in [48:13] to fear, buy something, do something, [48:16] buy into something, an idea and I just [48:19] won't do it. Okay, this this is going to [48:21] take decades. [48:23] This is going to take decades. [48:25] Um [48:26] it's going to play out for the rest of [48:27] our lives. When we are, you know, at the [48:30] end, [48:31] on our deathbed, [48:33] they're still going to be arguing about [48:34] AI's impact on society. People will [48:36] still have jobs. We'll still send our [48:38] kids to school. We'll still read books. [48:41] I'm not giving in to the panic and the [48:43] hype. We are being more thoughtful and [48:45] more present [48:46] uh in the face of what AI is doing [48:48] rather than less. [48:53] Okay. [48:54] So, we already did that. We're going to [48:56] have more time at the end but let me [48:57] just keep going. [48:59] So, [49:02] let me transition now from sort of [49:04] talking in broad strokes about [49:07] you know, my point of view, my method, [49:09] my principles, etc. that I'm that I'm [49:12] following to the program that we're [49:14] launching. [49:15] Okay, this this is [49:17] the next action. This is the concrete [49:19] thing that you can do like now, like [49:22] starting today if you have decided you [49:25] want to keep learning with me and keep [49:26] learning with us. [49:28] Okay? It's the first time that we're [49:30] teaching uh a live cohort. At least we [49:34] we did one focus on businesses last year [49:35] but this is the first time we're doing [49:37] one for individuals, for professionals [49:39] in 3 years. [49:41] It's taken 3 years of research and [49:43] thinking and we've done many, many [49:45] experiments, many, many, you know, trial [49:48] and error things and this is what we've [49:50] come up with as as really the new [49:51] direction of our entire company and my [49:53] work. [49:55] It's an intensive live 3-week program. [49:57] Okay, so for 3 weeks we are going to [49:59] meet three times a week. [50:01] I will be there every minute of it [50:03] personally to lead you through [50:04] everything. [50:06] And what you're going to walk away with [50:07] here at the bottom is a working AI [50:09] second brain at some level of [50:11] capability. I don't know if it's going [50:12] to be at the foundational level or the [50:14] intermediate level or the advanced [50:16] level. That's kind of up to you. That's [50:18] up to what you're ready for, what you [50:19] have time for, what you're even [50:21] interested in. [50:22] Okay, but what that is is a [50:24] personalized, organized, context-rich [50:27] digital exoskeleton. [50:30] And what it's going to do is make every [50:32] single conversation [50:34] and collaboration [50:36] not just something that anyone else in [50:38] the world would get from working with [50:39] AI, but something that's informed by [50:41] your expertise, your context and your [50:43] goals [50:44] and many other things as well. [50:48] Okay. [50:51] So, I want to kind of zoom into that. [50:53] What is the way what what are the kind [50:55] of the principles that we are using to [50:56] deliver this? So, [50:59] the first one is [51:01] compression. [51:03] Uh [51:04] the first assumption that I have is that [51:07] you don't have time. [51:09] Okay, I'm assuming you don't have time [51:11] or at least not as much time as you want [51:13] or as you need to uh explore the [51:16] frontier of AI. I mean, I don't have [51:18] time and it's my full-time job. Takes [51:20] our entire team spending most of our [51:22] time, hundreds, really probably [51:25] thousands hours, thousands of hours of [51:27] research and experimentation and [51:29] hands-on trial and error compressed into [51:32] 3 weeks. That's the main value [51:33] proposition. That's why you should join. [51:36] In 3 weeks we're going to take you from [51:38] wherever you are, even if you've just [51:40] barely opened ChatGPT, all the way to [51:43] not completely on the frontier cuz as I [51:45] said, it's not ready, it's not mature, [51:47] but pretty close to the frontier. [51:51] Everything we do in this cohort is [51:53] community-centric. Okay, so [51:56] this is hard to understand unless you've [51:57] experienced it, but [51:59] the way we do education, it's not like I [52:02] have all the answers and I am just [52:03] delivering them on to you. [52:06] Okay, that that's not how it works. I [52:08] really see this cohort more as like it's [52:10] all it's a learning community. It's a [52:11] community of practice. I really want and [52:14] I and I expect I'm going to I'm going to [52:15] ask those of you in the cohort. I I I [52:18] hesitate to even call you students. I [52:20] like participants or or peers more [52:23] because we're going to be constantly [52:25] sharing what we're figuring out, [52:27] constantly sharing our experiments, our [52:29] mistakes, our failures, our wins and [52:31] successes. I would say from our our past [52:34] years of doing cohorts that at least 50% [52:37] of what you learn and a lot of the most [52:39] impactful things come not from me, but [52:41] from the other participants. [52:43] It's a it's a collective, it's a it's a [52:46] crowdsourced learning experience that [52:48] we're doing here. [52:51] Everything we do is platform-agnostic, [52:53] right? We're So, I should say all the [52:56] demonstrations will be on Claude. Okay, [52:58] cuz that's the one I know the best, but [53:00] everything that we're building, the [53:01] master prompt, the little context [53:03] bundles, the AI board of advisors, my [53:06] this is like a life principle for me. I [53:07] don't do anything that is limited to one [53:09] platform. I just don't. That's that's [53:11] not a wise thing to do. So, everything [53:13] we're creating is completely [53:14] platform-agnostic. You take it with you. [53:19] And I think this is the last one, but [53:20] and I've I've kind of mentioned it a few [53:22] times, but I'm really not interested in [53:24] replacing humans, even if humans can be [53:26] replaced. [53:28] Um that's just not my preference. [53:30] >> [laughter] [53:31] >> I like humans. I value humans and my [53:34] time and attention and effort is going [53:36] to augmenting them, to making them [53:38] smarter, faster, wiser, more embodied, [53:43] more connected to their intuition [53:46] to their intuition, more creative, with [53:47] better judgment. [53:49] Right? So, we we are really focused on [53:52] augmentation, not automation. [53:56] And this is the price. So, regular price [53:59] is 3,000 bucks. [54:01] We have payment plans available via two [54:04] two different services. Klarna Klarna is [54:06] one of them and then Afterpay. [54:09] And there's an early bird [54:11] uh option. So, this is only for the [54:12] first 50 people, which is $2,000. It's a [54:15] $1,000 discount. I would really [54:17] encourage you if you are interested in [54:19] that to do it now. Like that will be [54:21] gone in hours. That will be gone [54:23] definitely before the end of today. [54:26] And what do you get for that investment? [54:29] A few things. First one is the option to [54:32] rejoin the next cohort. [54:36] Okay, so the first one, cohort one, the [54:38] founding cohort is next month in April. [54:41] The next one will be in the fall. We're [54:43] not sure exactly when, probably around [54:44] September. But you're essentially [54:47] getting two for the price of one. You [54:49] have the option to join a whole second [54:51] cohort in the fall [54:53] which will have a whole new curriculum. [54:54] Like I'm going to I'm going to update it [54:56] probably massively with a whole new [54:58] round of updates with everything that [54:59] has happened between April and [55:01] September. [55:04] Okay. [55:05] You also get access in the in-between [55:07] time to our second brain membership. So, [55:10] this is this is the cool thing is we're [55:11] not creating a community from scratch. [55:14] This is a community that has been around [55:15] for many years. Uh it's called the [55:17] second brain membership. It's on a [55:18] platform called Circle. Uh we have our [55:20] own custom app on iOS and Android, so [55:23] you can access it on the go. [55:25] Um there's polls, there's a discussion [55:27] forum, there's different spaces you can [55:29] join to kind of get into different [55:30] niches. You can look up people's [55:32] profiles. So, you can you can DM them. [55:35] There's a map. You can see like where in [55:36] the world do people live? There's a [55:38] whole world in there that you get access [55:40] to automatically um when you join the [55:42] cohort. I think that's like a [55:44] I believe it's a $650 per year value [55:47] that you're going to get with that. [55:49] Uh and then everything, all the course [55:50] materials, slides, curriculum, [55:52] everything will be available for 1 year. [55:58] Not going to get into this too much cuz [55:59] I want time for questions, but we have [56:02] so far confirmed four guest instructors. [56:05] This is something [clears throat] I'm so [56:06] excited about because this is almost [56:08] like a it's like a parallel curriculum [56:10] because I really think these are some of [56:12] the most important topics in in on the [56:15] AI landscape. My great friend Chloe is [56:18] going to talk about attachment, how your [56:19] attachment style [56:21] um shapes how you attach, how you relate [56:24] to AI itself. [56:26] Uh Joe Hudson, who is I believe one of [56:28] the most talented coaches in the world. [56:30] He's the personal coach of Sam Altman, [56:32] has worked with half a dozen teams at [56:34] OpenAI uh on the topic of fear, fear [56:38] when working with and within AI. [56:41] Uh my friend Raya is going to talk about [56:44] AI as co-creator. [56:46] She has some really cool projects and [56:48] companies that she's started around [56:49] that. [56:50] And then these are all my friends. These [56:52] are just basically my buddies that I [56:54] the people that I've learned the most [56:55] from that I am, you know, calling in to [56:57] teach you as well. Uh Johnny's a dear [56:59] friend and he's going to talk about [57:01] nervous system regulation. How do you [57:03] regulate yourself with the just [57:06] unbelievable pace of change that AI is [57:08] unleashing upon the world? [57:12] And just in case it wasn't clear, [57:15] this is not really a 3-week [57:17] arc, it's a year-long arc. [57:20] Okay, I'm not leaving you to your own [57:21] devices on May 1st when the 3-week [57:24] cohort ends. That's really just kind of [57:26] the kickoff. We have a 3-week kickoff. [57:29] Then we have five or six months of live [57:31] events in the membership. Most of those [57:33] live events are just going to unpack and [57:35] go deeper on what we covered in the [57:37] cohort. [57:38] Right? So, those happen at least weekly, [57:40] sometimes multiple times a week. We're [57:42] going to have all the time in the world [57:44] to just get into examples, get into case [57:46] studies, get into, you know, [57:47] step-by-step walk-throughs, answer your [57:49] questions, do coaching sessions. [57:52] That all happens within the membership, [57:53] which is also in the process of pivoting [57:55] toward AI. [57:57] Then in the fall we have cohort number [57:59] two with a updated curriculum and then [58:01] there's another 6 months. There's a [58:02] second half of that year membership to [58:05] then follow up on everything that you [58:06] learn there. [58:08] So, what I'm really asking you to do is [58:10] to go on a learning journey with us. Go [58:12] on a year-long learning journey. [58:14] I can't even imagine the things that are [58:16] going to be unveiled over the next year [58:19] in the world of AI, but we're going to [58:21] do it together. [58:23] That's the bottom line here. [58:25] That's my concluding thought is [58:27] when I think about the scale of years [58:29] and decades, like think about what what [58:32] is most important if this is something [58:34] that is going to unfold over many years [58:36] and even decades. [58:38] I think the single most important thing [58:40] for you to decide now is who are you [58:42] going to learn with? [58:45] You know, any book, any single book or [58:48] video or or course is going to pass away [58:50] within a matter of months. [58:52] Uh any method, any technique, any set of [58:55] principles, they're all just being [58:58] revolutionized on a monthly basis. What [59:01] will persist, I think, what's going to [59:03] stand the test of time is your [59:05] community. [59:06] The people, the [59:08] the other human bodies [59:10] that you decide you want to influence [59:12] and be influenced by. [59:15] So, I guess [clears throat] if there's [59:16] one decision that I would make, you [59:18] know, today, whether you choose to join [59:19] us or not, really is you know, that's [59:21] completely your choice, just really [59:23] thoughtfully consider who do you want to [59:26] learn with. [59:27] Right? Don't make the [59:29] the feed on Twitter your community. [59:32] Don't make the feed of endless hype that [59:35] community. Choose a real community. [59:37] Choose a community that is that is [59:39] thoughtful, that is wise, that is [59:41] diverse, that has a lot of different [59:42] viewpoints, and really invest in it. [59:45] Become a member of that community. [59:48] And if what I've said today resonates [59:50] with you, if it feels like it fits the [59:52] way you see the world, then I I invite [59:55] you to join ours. [60:01] All right. [60:07] Okay. [60:09] Woo, that was a lot of talking. [60:11] We have 25 minutes left. [60:21] And I'm here for you. Give me your [60:23] hardest [60:25] most complex question. [60:28] My answer may be it depends, but I'll do [60:30] my best to give you some clarity. [60:35] All right. So, why not start with the [60:37] question from Sherilyn? Can you talk [60:39] more about how emotional regulation, [60:41] nervous system, inner awareness are [60:43] critical for this process? Just want to [60:46] understand more about how you understand [60:48] this, and who are some of the guest [60:50] instructors? Uh [60:52] We actually saw that, so. [60:54] >> [laughter] [60:56] >> Yeah, I would say um I just want to say [60:58] one thing about that. Uh this is [61:01] a few weeks ago, about 3 weeks ago, I [61:02] did a week-long uh course with Joe [61:05] Hudson, one of the guest instructors, [61:08] uh in Sonoma in California. [61:10] I'm going to be writing [clears throat] [61:10] a blog post on this. It's the the draft [61:13] is currently 7,000 words, and I think by [61:16] the time I'm finished will be like [61:17] 10,000 words. So, I'm sorry for [61:19] >> [laughter] [61:20] >> the massive amount of text that I'm [61:22] about to unleash. Um but the the big [61:25] breakthrough, the big takeaway was [61:26] around fear. [61:28] I would not have said that I was afraid [61:30] of AI or feeling fear toward AI. I would [61:32] have said I was excited. [61:34] Right? That I was optimistic and [61:35] positive. [61:37] But through this process, and we're [61:38] going to we're going to do some of the [61:40] exercises and some of the practices that [61:41] we did in my course, I've asked Joe to [61:44] lead us through those like live on Zoom. [61:47] Because it goes so deep, I realized that [61:49] I had so much fear. I had multiple [61:51] levels. Fear on the personal level, [61:52] what's going to happen to me, my work, [61:54] my career, my income. Fear on the family [61:57] level, what are my kids' futures going [61:59] to be like? What is the right way to [62:01] parent? [62:02] What do my relationships look like at [62:04] the level of the team? [62:06] The business, the business model, the [62:08] industry, all the way to like the fate [62:10] of the world and the future of [62:11] civilization. There's no way that all [62:14] those levels of fear are not really [62:16] really powerfully shaping [62:18] the way you think and the decisions you [62:20] make. And so, the important thing is not [62:22] to deny that, it's to bring it into the [62:24] light. [62:25] To feel through it. [62:27] Cuz each one of those kinds of fear are [62:29] telling you something. They have a [62:31] message. They have a signal that you [62:33] need to hear, and the key is just to [62:34] hear them. [62:36] Which is [62:37] I'm making it sound like an intellectual [62:38] process. It's not. It's a somatic, [62:41] embodied, experiential process. [62:43] And we will we will try some of that [62:46] in that guest session. [62:54] What else? [62:55] Anything especially that we like we [62:58] didn't cover on the the website or in [63:00] the in the emails that I've been sending [63:02] out, something that has not been [63:03] addressed elsewhere? [63:05] Yeah, I'm I'm hearing um that some [63:07] people are a little bit concerned if [63:09] they are not really fully up to speed [63:11] with maybe code or PARA yet, that they [63:13] might be starting the cohort already on [63:16] the behind. So, what could you say about [63:18] that? [63:20] Yeah, so a couple things. [63:23] Um [63:24] the first thing you will see the minute [63:26] you sign up, [63:27] like even today, [63:28] um is we have a little I I believe it's [63:31] already in there, right Julia? The [63:32] Claude course? [63:35] Is that in the curriculum? [63:36] >> Yeah, yes. We have a beginner's guide to [63:37] Claude that you will get immediate [63:39] access to when you sign up. [63:42] So, we're really not assuming any [63:43] knowledge. Like it's really helpful if [63:45] you've opened Chat GPT once or twice, [63:48] uh or Claude, but um the first thing you [63:51] will see in the curriculum, like right [63:53] now when you sign up for the cohort, is [63:55] a pretty It's it's not extensive. It's [63:57] maybe like five or six modules or [63:59] something like that. Just giving you a [64:01] beginner's introduction to Claude. [64:03] Right? So, we're giving that to you up [64:04] front. So, like when I refer to a [64:06] certain feature or I talk about [64:08] something that is specific to Claude, [64:09] you can already know what I'm talking [64:11] about. That's the first thing. [64:13] The other thing is we have so much [64:14] existing content, and I I agree it's [64:16] almost a problem because, you know, I'm [64:18] not going to have you go read the PARA [64:19] book. That would take a long time. Just [64:21] watch the PARA video. [64:23] Okay, that's 10 minutes, and you can get [64:25] most of what you need to know. [64:27] Uh if you need to know about code, we [64:30] have, I believe, a video on that. [64:32] There's practically everything that I'm [64:34] going to talk about that is sort of like [64:35] foundational knowledge, there is either [64:38] an existing piece of content that is [64:39] quite succinct that you can consume in a [64:41] matter of minutes or maybe half an hour [64:43] or something. And if not, then we're [64:45] going to provide it as part of like the [64:47] pre-work, the pre-curriculum inside [64:49] circle. [64:51] Um [64:52] My my principle here is I want the floor [64:54] to be very low. [64:56] I mean, I'm going to have like my family [64:57] members, my my parents. [64:59] >> [laughter] [65:00] >> Sorry for using you guys as an example. [65:02] I think they're on the call right now. [65:03] But my parents are going to be there, [65:04] okay? They're the ultimate [65:07] >> [laughter] [65:08] >> the ultimate judge of whether we've made [65:10] it, you know, easy to understand. But [65:12] then I want to make this the the the [65:13] ceiling very high. I want those of you [65:15] that are like all into this, that are [65:17] obsessed, that are just like going full [65:19] speed ahead by the time the cohort is [65:21] finished to feel like you've really [65:24] raised the ceiling of capability. Which [65:26] does mean we're going to move fast. [65:28] The only way to have those two things is [65:30] for there to be a a steep learning [65:31] curve. [65:32] And so, we have you know, we have three [65:34] 90-minute calls a week. It's it's [65:36] intensive. I put that word intensive in [65:38] there. It's not an accident. Like you [65:40] really need to like set aside the time. [65:43] Honestly. [65:44] Like set aside the time for it. [65:50] Question from Megan. How much time will [65:52] we spend on personal development work [65:55] versus sort of like this AI and thinking [65:59] brain training? [66:01] Yeah, it's a great question. So, [66:05] >> [clears throat] [66:05] >> excuse me. [66:10] Um [66:11] so, part part of my answer to all these [66:14] kinds of questions is [66:16] we are going to very much adapt and flex [66:21] what we're teaching in response to what [66:23] people are experiencing. Like like this [66:25] this is actually a very important kind [66:27] of disclaimer. [66:28] Um [66:29] the whole point of the live cohort is [66:31] not to like take you on a forced march [66:34] through this curriculum exactly as [66:36] designed. It's to feel you out. [66:39] It's to feel you out. I have so many [66:40] ways of doing that. I have the little [66:42] feedback surveys, I have your circle [66:43] comments, I have the expression on your [66:45] face on Zoom, because we'll be in in [66:48] meeting mode, so I'll be able to see the [66:49] gallery of all your faces. I have your [66:51] comments in the chat. If I sense that [66:53] people are struggling or that we're [66:55] moving too fast or that a you know, a a [66:58] previous concept didn't land, then we [67:00] stop. [67:01] We don't we don't just like charge [67:02] ahead. We flex, we adapt, we move [67:04] faster, we slower move slower. We often [67:06] like will create sessions like brand new [67:09] sessions. We'll be like, "Hey, we need a [67:11] breakout call, a whole other call to [67:13] just talk about that thing." cuz there's [67:14] like say a small sub [67:16] small subset of people who are [67:18] interested in that. [67:19] It's a it's an interactive experience. [67:22] Okay? Um I always say that each and [67:24] every person in the cohort very like [67:26] very much shapes the cohort. [67:28] Like individual people change. They have [67:31] an impact on how we go and where we go, [67:33] and that's so important, right? [67:35] Um and so, it's it partially depends on [67:38] the the interest in that personal [67:41] development side that I sense. [67:43] If we have like five people interested [67:44] in it, I'll kind of be like, uh let's [67:46] just kind of like move that conversation [67:48] to the the guest sessions, which is [67:50] almost like a parallel curriculum. [67:52] If there's a lot of interest in it, we [67:54] may we may emphasize it quite a bit. [67:58] Depends. [68:04] All right. What else? [68:06] Um I I just want to mention that we have [68:09] already sold out of of our 50 seats [68:12] at the $2,000 level. Okay. So, that that [68:16] wasn't planned. That wasn't like one of [68:17] these like crazy urgency-creating [68:20] tactics. That was much faster than I [68:22] expected, but uh early bird is over. [68:25] Early bird is finished. [68:30] All right. Okay, I want to ask [68:32] ask a question from Preston. How do you [68:35] find your alpha with AI? [68:38] >> [clears throat] [68:39] >> Yes, such a great one. Such a great one. [68:43] Oh my gosh, we're going to talk about [68:44] this. We have a whole session dedicated [68:46] to note-taking and capture. [68:48] Because it's subtle, you guys. It's so [68:50] subtle. Like [68:52] I'll just give you [clears throat] like [68:53] one random example. [68:55] Um [68:57] I'm I'm really getting into local [68:58] history. [69:00] This is this is the last thing I ever [69:01] thought I would be interested in, but uh [69:03] we live in this Mexican town right now. [69:05] It's a it's a small town. I mean, [69:07] smallish, 100,000 people. Small compared [69:09] to Mexico City, right? And I'm noticing [69:12] that it has a really rich history that [69:14] goes back almost 500 years. [69:17] And none of it is digital. [69:19] You can't find it online. I've tried [69:20] asking every AI platform, they're all [69:22] like, "We know nothing about this town." [69:23] right? [69:25] And at first I was like annoyed. I'm [69:26] like, "I just want to, you know, I want [69:27] a quick, easy, convenient, seamless way [69:30] of learning this. [69:31] But as I thought about it, I'm like, oh [69:32] my god, this is an incredible [69:34] opportunity. And I'm now like I'm kind [69:36] of like in the Indiana Jones. I'm like [69:38] going to the local library and taking [69:40] out old books and literally like blowing [69:42] on them with like the dust, you know? [69:44] >> [laughter] [69:45] >> And it's like I get to go on this little [69:46] adventure of finding these these rare [69:49] physical printed documents and like [69:52] reading it reading through them in some [69:53] cases and like old Spanish, using AI to [69:55] translate the old Spanish into English. [69:59] Um I'm working on a blog post where I'm [70:00] going to talk about the 500-year history [70:02] of this town. And so that's one example [70:04] that I never would have imagined, local [70:06] history. Think about the the local [70:08] history that exists all over the world. [70:10] Most of it is not digitized. [70:12] Most of it is is not in the AI training [70:14] data and therefore that's alpha. [70:17] Now, how you take advantage of that [70:18] alpha, that's a separate question. [70:20] You know, maybe there's not like a [70:21] direct way to say make money off that. [70:24] But like in my case, I'm using it for [70:25] community building. I'm starting a [70:26] nonprofit with a friend uh to promote [70:29] like sustainability and innovation in [70:31] this town and I'm going to like share [70:33] this research with this little community [70:36] that we're building as a way of saying [70:37] like you guys, there's a really rich uh [70:39] tradition here. There's a rich legacy. [70:42] And so like the the return on investment [70:44] for me is not dollars. It's like [70:45] connection, interestingness, satisfying [70:49] my curiosity, having conversation fodder [70:52] for, you know, these local [70:54] friends that I have in my town. So you [70:56] you got to think beyond finances. Think [70:58] beyond money. [71:00] Um but there's so many examples. Like [71:01] things that happen in conversation are [71:04] alpha. [71:05] Conversation with friends, things that [71:08] uh happened [71:09] in the past like during your childhood. [71:11] I'm doing another project now to [71:13] digitize all my home videos from my [71:14] childhood. [71:16] Cuz I realized, holy crap, those grainy, [71:19] low-resolution home videos from when I [71:21] was five, [71:22] AI can will soon is pretty close to be [71:25] able extract tremendous meaning from [71:27] that or at least just edit it into like [71:29] a highlight reel. [71:30] Uh it's everywhere. Alpha's everywhere. [71:34] Uh but you got to you got to you got to [71:35] think, you got to pay attention, you got [71:37] to think like what kinds of alpha are [71:38] most relevant to you. And then you have [71:40] to find an efficient means of capture [71:43] for getting that alpha into your own [71:45] context. [71:50] Okay, so this is a question from a [71:53] little earlier and I don't remember who [71:55] asked it, but it was around like [71:57] specific like use cases. Like how have [71:59] you specifically seen this sort of like [72:02] help you? Like you know, the master [72:03] prompt, what you will be teaching, [72:06] what what results? Talking about the [72:09] results you've gotten. [72:10] Yeah, sometimes we nerd out on the the [72:13] the process so much that we forget, you [72:15] know, the results. [72:17] Um there was another comment too I saw [72:19] in the chat like is this just for [72:20] business? [72:21] Or even just for work? And I would say [72:23] absolutely not. [72:25] Uh those tend to be I tend to use [72:27] work-related examples cuz the return on [72:29] investment the investment of your time [72:31] is so clear. But I would say the impacts [72:34] have been just as big in my personal [72:35] life. [72:37] I'll give you a couple examples and this [72:38] is what we're going to tackle in the AI [72:39] board of advisors. Those those advisors [72:42] are not just for work life. I have [72:43] actually found them more useful in my [72:45] personal life because in my personal [72:46] life I don't have as good of advice. [72:49] I I I like can't justify, you know, the [72:52] the consulting and the coaching in my [72:53] personal life. [72:55] So one example the two examples are my [72:57] health and my finances. So with my [72:59] health, [73:00] I'm going to write [clears throat] this [73:01] up uh soon, but [73:03] you guys, I just [73:05] I've just started a basically assembling [73:08] all the context I can find about my [73:10] health, [73:11] providing it to my AI health coach, my [73:13] health advisor. And I I would say I've [73:16] learned more about health and my own [73:18] personal health in the past couple [73:21] months than than I ever have before. [73:23] Like the its ability to cross-reference, [73:25] oh, you're taking that supplement, but [73:27] you're also do you're also having that [73:29] protein after your workouts. Well, that [73:31] this ingredient in that protein, which I [73:33] shared with it, is counteracting and [73:36] making it so you can't absorb that [73:37] supplement. Or oh, you're trying to get [73:39] more protein in your diet, but you're [73:41] you [73:41] eating too much steak. That's too much [73:44] red meat, which is increasing your [73:45] cholesterol, which I know you're trying [73:47] to keep an eye on. So substitute 50% of [73:49] that that red meat intake with fish and [73:51] chicken. Like things like you would need [73:54] a team of not even doctors cuz doctors [73:56] aren't good at nutrition. You would need [73:57] a team of nutritionists [74:00] around the table, you know, going [74:02] through your health documents and which [74:04] are of course are in a dozen different [74:06] formats. It's crazy. The number of [74:08] formats you have to work with with [74:09] health. And I talk to it almost every [74:12] day. [74:13] Uh and now I'm doing the same for my [74:14] family. I'm creating health advisors for [74:16] different members of my family because [74:18] I'm just like my parents, you know, [74:19] they're they're getting old and they [74:21] don't have anywhere near this good of [74:23] advice for their health. [74:25] Um the other one is [74:27] um finances. [74:30] Um [74:31] we've worked with a financial advisor [74:33] before, but do you know what financial [74:34] advisors charge? [74:37] Like they charge a percentage of fees [74:39] under management. Have you seen the math [74:41] on like if you work with a financial [74:43] advisor like over 5, 10, 20 years, [74:45] you're paying them like a sizeable [74:47] percentage of all the money that you [74:49] make from your investing. It's absurd. [74:51] Right? Even the the advisors that you [74:53] can pay like by the hour are expensive. [74:55] Um and so I've given it uh our financial [74:58] context and [75:00] there's just all these all these like [75:02] mistaken assumptions I was making about [75:04] my finances. There were some things I [75:06] just flat out did not understand about [75:09] how retirement account works [75:11] accounts work, how 529 education [75:13] accounts work. Like I had certain ideas [75:15] of what risk is or even means that were [75:17] actually completely misguided. Uh and [75:20] we've significantly changed how we save, [75:22] how we invest, how we give, how we how [75:26] we uh donate uh based on our work with [75:29] our AI financial advisor. Those are a [75:30] couple examples. [75:39] So the reason you need your uh we have [75:42] 10 minutes left. Your board of advisors [75:43] is So this is another example of where [75:46] the master prompt fails. Your master [75:48] prompt is very like hor Think of it as [75:50] horizontal. [75:51] Right? It's across, like I said, [75:53] everything. It's across every AI [75:56] interaction. [75:57] Right? [75:58] But if you have a certain domain of your [76:02] work or life, like health, like [76:04] finances, at work it might be like [76:06] looking at your HR policies or looking [76:09] at your insurance needs or looking at [76:12] your marketing strategy, like one [76:14] vertical slice, your master prompt won't [76:17] have anywhere near enough context. [76:20] Right? How could it? How could it? So [76:22] for So for those vertical slices, you [76:24] have to go really deep on just that [76:26] slice. Where is the context that you'll [76:28] need? And often like you got to you got [76:31] to go spelunking. [76:32] You got to go log into random platforms, [76:34] download things, convert them from one [76:36] format to We're going to talk all about [76:38] this, but convert them in into a format [76:40] that is AI friendly such as markdown. [76:42] You have to um you have to do a little [76:44] dedicated like curation effort to get [76:48] that context in one place because it [76:50] tends to either not exist at all or not [76:53] exist in digital form or not exist in a [76:55] digital form that AI can make use of. [77:04] Um if you folks have asked if this is [77:07] suited for people with ADHD and maybe [77:11] the broader question is who is really [77:13] best uh would best benefit [77:15] from the course? [77:17] >> [gasps and sighs] [77:17] >> Yeah, there's a few I think we have this [77:19] on the website, but there's a few people [77:20] that come to mind. Um [77:23] I mean the the the main one as I said is [77:26] if time is a constraint. [77:28] Because if time is not a constraint, if [77:31] you can afford to spend I I feel like [77:33] realistically it takes at least one or [77:36] two hours a day to kind of tread water. [77:39] Like that's just that's like [77:40] maintenance. That's like basic [77:42] maintenance. It it really does, you [77:43] guys. Um like those one or two hours [77:46] would just be spent sort of like [77:48] maintaining what you already have. [77:50] You know, as I said before, it takes [77:52] maintenance. Don't let anyone [77:54] convince you that no maintenance is [77:55] required. [77:57] Uh and then I think you would need at [77:58] least another one or two hours to do [77:59] anything new. To learn anything new, try [78:02] anything new, adopt anything new. So [78:03] like three to four hours I feel like is [78:05] what it really takes. If you have three [78:07] to four hours a day, go for it. Like do [78:10] trial and error. There's no there's [78:12] nothing like direct experience. Right? [78:15] Um but if you don't, then allow us to [78:17] compress. [78:19] You know what I noticed that was funny? [78:20] I I went through [78:21] months or even like years of of notes [78:23] and Notion docs and experiments and [78:26] different things designing this [78:27] curriculum. There was like a six-month [78:29] window. [78:30] Only things about six months old. So uh [78:34] up to like the beginning of the fourth [78:35] quarter of last year were relevant and [78:37] then it like dropped off. [78:39] Everything that was more than six months [78:40] older was like zero relevant, zero [78:42] usage. It was actually like the wrong [78:44] thing to pay attention to. So I kind of [78:46] think in AI right now there's like a [78:47] moving six-month window of relevance. [78:50] And so this cohort, this three-week [78:51] cohort, we're getting really the last [78:53] six months and we're compressing it into [78:56] I think about 15 hours is what we have [78:58] in terms of live sessions. [78:59] So six months [79:01] of time compressed into 15 hours. That's [79:03] the main value. [79:05] And then besides that, I would say [79:06] professionals. There will there will be [79:08] I would say more of an emphasis on work. [79:10] That's tends to be the arena where [79:12] people use AI more. [79:14] Uh people who [79:16] um [79:17] are neurodivergent and have ADHD. I kind [79:20] of think that this is a golden age. [79:22] This This is AI is ushering in a golden [79:25] age for people with ADHD and people who [79:27] are neurodivergent. [79:29] Because finally multitasking [79:32] >> [laughter] [79:33] >> such as between multiple cloud code [79:35] sessions that are all happening at the [79:37] same time or multiple agents that are [79:38] all working. That's that's now going to [79:41] become one of the most valuable skills. [79:43] It's going to become so incredibly [79:45] important to be able to do that and it [79:46] is a skill. It's a skill. It's like the [79:49] skill of middle management. Those of you [79:51] who are are middle managers and have, [79:53] you know, eight direct reports. [79:55] That's people talk about, "Oh, middle [79:57] management is going away. That's going [79:58] to be the first people to lose their [80:00] jobs." Maybe so, but you're just we're [80:03] all going to become middle managers now. [80:05] >> [laughter] [80:05] >> We're all middle managers of our own [80:07] team of AI agents. [80:09] And as any manager will tell you, I mean [80:11] people people go through an incredible [80:13] like transformation to become managers. [80:16] To be able to lead, to be able to [80:17] delegate, to be able to coach, to be [80:19] able to guide, to see things from [80:21] different perspectives and let go of [80:23] your perspective to take on someone [80:24] else's. It's like a whole skill set that [80:27] we kind of need to learn now and people [80:28] with ADHD [80:30] I I I've seen some examples that they [80:32] tend to thrive in that environment. [80:37] Who else? Um any Anyone else come to [80:40] mind, Julia, as a as a good fit for [80:42] this? [80:45] I was answering questions and not [80:47] listening to you. [80:48] >> [laughter] [80:49] >> Perfect example. [80:53] >> [gasps] [80:54] >> Okay. [80:58] Uh question from Maria. She's in the [81:01] middle of reading second building [81:03] building a second brain. What's your [81:04] advice on starting to capture at this [81:07] point? [81:12] I think we've already talked quite a bit [81:13] about capture. [81:14] I would just say it continues to be [81:16] important as important as ever. [81:18] But the way it happens and what it's [81:20] focused on needs to change. [81:23] Needs to change. [81:25] Needs to become more about [81:27] just more about finding whatever is not [81:29] in AI's training data or what AI cannot [81:32] find through a web search. [81:34] It's as simple as that. [81:46] Anything else? 5 minutes left. [81:50] I wish I could see you all. I can I can [81:51] get a sense of you from the chat over [81:53] here on the right, but it's kind of it's [81:55] a lot going on, but [81:57] I love I love that we do meeting mode in [81:59] the cohort because it's it's it's almost [82:00] like we're just like it's like a work [82:02] meeting. We're getting on with our [82:03] colleagues, our collaborators, our peers [82:06] and we are [82:08] like yes, I'm teaching, but we're also [82:10] kind of knowledge sharing or [82:12] giving feedback on different things. [82:14] We're sharing things we've tried. No No [82:16] one No single person has enough time to [82:18] try everything. It's impossible. [82:20] So I I really in in a funny way I'm [82:22] trying I'm [82:23] >> [laughter] [82:24] >> I'm very much trying to assemble the [82:25] group of people that I personally want [82:28] to learn with. [82:30] Honestly, like that's it. I'm trying to [82:32] bring together a little crew [82:34] uh that, you know, shares some [82:37] principles, but then has other, you [82:39] know, divergent and diverse points of [82:41] view because I want to spend the next [82:43] few years learning learning with those [82:45] people and I can't do it all myself. [82:47] Even our team of six cannot do it all [82:49] ourselves. [82:57] Oh, the t-shirt. Yeah, this is These are [82:58] the previous t-shirts. [83:00] I guess they still kind of apply. [83:02] Um I've long wanted to make the t-shirts [83:04] available on the website. Maybe we'll do [83:06] that. [83:09] Uh there's no attendance [83:10] >> make an attempt. There's no attendance [83:12] limit. This is something that I that has [83:14] always been important to me. [83:16] Um that to me feels like an artificial [83:17] constraint. Like why as long as Zoom can [83:19] support it, why put a cap on it? [83:21] Uh which also means we don't actually we [83:23] really have no idea how many people are [83:25] going to [83:26] are going to end up in this. My goal was [83:27] at least 50. 50's kind of a nice number. [83:30] Seems like it's going to be more than [83:31] that, but [83:32] um [83:33] we don't know for sure how many people [83:34] will be in there. [83:41] And I just want to point out if anyone [83:42] has questions, um you can reach out to [83:45] us at [email protected]. [83:48] We'll be right there [83:51] answering your questions. [83:55] Thanks, everyone. [83:56] >> probably many still have some. [83:58] Yes, [clears throat] please send us all [83:59] your questions. We want to hear them [84:00] all. [84:01] And [84:02] >> And enrollment is open until April 14th. [84:06] That's one day before the first live [84:08] session, which is on April 15th. [84:13] Awesome. [84:15] Thank you. [84:16] Thank you all so much for being here. I [84:17] appreciate you. [84:19] And I hope to see you soon somewhere. [84:23] >> [clears throat] [84:27] >> Bye, everyone. Thank you, everyone. [84:31] Adios.  📄 **Full text:** ![[Full Document Contents/Articles/The AI Second Brain Complete System Walkthrough.md|The AI Second Brain: Complete System Walkthrough by YouTube]] ## Highlights - Okay. So, as I said, I want you to have a roadmap, a high-level roadmap of what the AI second brain system, the system that we talk about looks like. And how if if assuming you want to learn alongside us, what that might look like, how it works. All right? Um I don't have that many slides. I think I have like 25 or 26. Uh there's going to be plenty of time for us to interact. That's kind of one ([View Highlight](https://read.readwise.io/read/01kpm4m2s8vr6gw2sz5h0mh873)) - Tags: 👻 ai highlighted - is really just information, to increasingly powerful generations of artificial intelligence tools. Which is a skill. It's not it's not obvious how to do this. It's a skill. It's a it's a discipline that we're calling PCM, personal context management. A brand new skill in the world that if you can master, you can get around this bottleneck. So, thinking about you as the human in ([View Highlight](https://read.readwise.io/read/01kpm4m2sh0cwks37avdfgcn3q)) - Tags: 👻 ai highlighted - all this. Don't want to forget about that part. Consider that your AI, all your different AI tools can only use, they can only make use of what you give it. What you give it. Your personal curated context, your notes, your knowledge, your expertise, your experience, your memories, your points of view, your perspectives, your relationships, your network, all the stuff that makes you you. ([View Highlight](https://read.readwise.io/read/01kpm4m2sx0gmv6mma7zhbfbcn)) - Tags: 👻 ai highlighted - And that might seem like just a random, you know, it's kind of technical detail, but it's huge, you guys. It's so important because it means that all the stuff that we that we had to do and that we learned pre-AI, how to organize, how to curate, how to select, how to take notes, how to structure, all those things are still important. They're even more important because you have to do them to take advantage of AI's power. ([View Highlight](https://read.readwise.io/read/01kpm4m2tpzx26d0yvnhj26aew)) - Tags: 👻 ai highlighted - videos, there's the podcast, there's many blog posts. You can have at it. In whatever format you like to learn, there is a way for you to learn those two methodologies. But we're adding a new one. Right? We're adding this new layer that is building on. It is extending those pre-AI capabilities, but with this new generation of technology that we call AI. All right? So, let's zoom in a bit. Let's get into ([View Highlight](https://read.readwise.io/read/01kpm4m2vn72zd8wgmzx3vm300)) - Tags: 👻 ai highlighted - that if you had unlimited money, you would hire to help you, you can now have with AI. But it requires some design. Requires a little bit of of uh of, you know, of designing. All right? So, I want to say a little bit of more about the master prompt, because this is really this is really kind of the core thing. What is a master prompt? You can think of a master prompt as ([View Highlight](https://read.readwise.io/read/01kpm4m2yanrfeqbymewvz1qy2)) - Tags: 👻 ai highlighted - modern life at this point. But imagine if it knew your basic personal details. Who you are? What's your name? How old are you? Where do you live? Who's in your family? Imagine if it had your basic professional info. What's your job? What's the title? What's the role? What team are you on? Who do you report to? Who do who who reports to you? Imagine if it had go just going around clockwise around the circle, your income sources. Do you ever ask AI for any advice or have any questions for it ([View Highlight](https://read.readwise.io/read/01kpm4m303wjmphqbm845cq6va)) - Tags: 👻 ai highlighted - want to get to those. Um I think there's a couple more things I want to share with you, and then we'll pause. So, how do I recommend you go about creating this thing called a master prompt? I have a very strong opinion on this. Okay? And this really speaks to what I think I've done so much reflection on the, you know, the decade that we were teaching Building a Second Brain, the ([View Highlight](https://read.readwise.io/read/01kpm4m33n70cyek3vy5h630a9)) - Tags: 👻 ai highlighted - product, minimum viable prompt. Like what every single professional, no matter what you do, needs to have. We're going to do that in week one. So, if you just do week one and then you get busy or you drop off or that's enough for you, great. You got the the 80/20 of what I have to offer. Okay? And the the other thing that's important here is even for those of you who are more advanced, and many of you are, if you think about it, most of your AI interactions still reside at at kind of ([View Highlight](https://read.readwise.io/read/01kpm4m36xb89qjpje8bt0ate9)) - Tags: 👻 ai highlighted - And what it's going to do is make every single conversation and collaboration not just something that anyone else in the world would get from working with AI, but something that's informed by your expertise, your context and your goals and many other things as well. Okay. So, I want to kind of zoom into that. What is the way what what are the kind of the principles that we are using to deliver this? So, the first one is ([View Highlight](https://read.readwise.io/read/01kpm4m3dafherc3ba4xje4mbj)) - Tags: 👻 ai highlighted