Dan Koe shares his unique approach to using AI, emphasizing that most people treat it like a slot machine instead of programming it for specific tasks. He aims to provide an actionable guide on leveraging AI, specifically LLMs like ChatGPT or Claude, through text-based interactions. Dan introduces his "secret" method to impose one's own taste on AI outputs, viewing AI as a digital employee that needs detailed instructions to avoid mediocre results. He outlines four options for teaching AI, including writing detailed instructions, asking AI to create a guide, finding expert sources, and emulating examples. Dan also introduces his software, Eden, designed to enhance AI workflows, and provides examples of using AI as an intellectual sparring partner, a creative thought partner, and for various business-related tasks like content creation, product development, and copywriting, advocating for a shift from passively receiving AI-generated content to actively orchestrating and documenting processes with AI.
Part 1: Introduction and Core Concept
00:00Introduction to a Unique Approach to Using AI
Introduction to a Unique Approach to Using AI
The speaker introduces a unique method of using AI that unlocks new capabilities for users, enabling them to accelerate business growth, skill acquisition, and topic comprehension. This approach contrasts with the common perception of AI as a mere "slot machine" or an advanced search engine, emphasizing the untapped potential beneath the surface. The speaker argues that the initial hype surrounding AI has waned, with many failing to realize its transformative potential. The core idea is that most people don't see the power under the hood.
02:32AI as a Digital Employee: The Key to Effective AI Utilization
AI as a Digital Employee: The Key to Effective AI Utilization
The discussion shifts to viewing AI as a digital employee that requires precise instructions to perform effectively. The speaker stresses that if the user lacks expertise in the task they're assigning to the AI, the results will likely be disappointing. The speaker highlights the importance of providing specific instructions to avoid relying on the AI's default programming, which is tailored for the average user. The speaker argues that consumer products dumb down AI so that it can be useful and sycophantic and give you your cheap little dopamine hits so you keep coming back to the slot machine.
03:47The Pitfalls of Generic AI Prompts and the Need for Customization
The Pitfalls of Generic AI Prompts and the Need for Customization
The speaker illustrates the ineffectiveness of generic AI prompts with the example of asking ChatGPT to generate a viral YouTube script on productivity. The output is deemed mediocre and lacking the unique style and expertise of established YouTubers like Ali Abdaal or Alex Hormozy. The speaker emphasizes that AI lacks the specific context and instructions needed to produce high-quality content. The speaker argues that the AI doesn't have any of your specific context or instructions on what to do. The speaker stresses the need to teach the AI how to create content in a specific style, moving beyond randomly generated outputs. This involves writing detailed prompts, ranging from 500 to 2,000 words, to provide the AI with sufficient guidance.
Part 2: Teaching AI - Options and Examples
06:17Overcoming the Knowledge Gap: Teaching AI How to Perform Tasks
Overcoming the Knowledge Gap: Teaching AI How to Perform Tasks
The discussion addresses the challenge of using AI when the user lacks expertise in the task they're trying to accomplish. The speaker outlines four options for teaching AI how to perform tasks effectively, emphasizing the importance of creating detailed instructions. The speaker argues that this isn't the magic step yet, but this is the prerequisite to getting to the magic step.
07:14Option 1: Writing Detailed Instructions for AI
Option 1: Writing Detailed Instructions for AI
The first option for teaching AI how to perform tasks is to write out detailed instructions. The speaker provides an example of a prompt created to replicate their tweets, including requirements, post examples, and output format. The speaker notes that while this approach yields good results, it doesn't fully capture the nuances of their thinking process when writing content. The speaker argues that most people, when they do this, they're going to give one or two examples and then all of their tweets are just going to look the same.
08:57Option 2: Asking AI to Create a Detailed Guide
Option 2: Asking AI to Create a Detailed Guide
The second option involves asking AI to create a detailed guide on a relatively well-known topic with minimal variation. The speaker uses the example of creating a customer avatar, where AI can provide a comprehensive guide. This guide can then be used to create a prompt that interviews the user to gather information and generate a detailed customer avatar. The speaker argues that since I'm doing this with ai, I'm not just staring at this blank template where it's like a customer avatar template and I'm expected to fill in.
10:46Options 3 & 4: Expert Sources and Emulating Examples
Options 3 & 4: Expert Sources and Emulating Examples
The third option involves finding an expert source of information, such as Alex Hormozy's PDF on offer creation, and using it to guide the AI. The fourth option is to emulate an example the user likes, such as a landing page with compelling copy. The speaker suggests copying the content from the website into a chat and asking the AI to break down the structure and psychological tactics used. The speaker argues that I want to find something unique and this page in specific, this anti-metal page, has a really cool storytelling structure that is very attention-grabbing and compelling.
Part 3: Implementation and Process
13:17Introduction to Eden Software
Introduction to Eden Software
The speaker introduces Eden, a software with a canvas feature for file storage, AI chats, and prompt items. Eden allows users to paste various links, downloads, transcribes, and auto-tags them for easy searching. The speaker argues that this is the next iteration of Cortex.
14:46Step 2: Turning Detailed Instructions into a Prompt with Personal Context
Step 2: Turning Detailed Instructions into a Prompt with Personal Context
The discussion transitions to step two: turning detailed instructions into a prompt by adding personal context. The speaker emphasizes the importance of tailoring the AI's output to the user's specific company, product, and customer avatar. The speaker introduces a "Metaprompt," a prompt that helps users create prompts, streamlining the process and saving time. The speaker argues that this alone will change how you use AI as a whole, so please just save this somewhere safe.
16:03Using Claude Opus and Providing Instructions for a Personal Brand Coach
Using Claude Opus and Providing Instructions for a Personal Brand Coach
The speaker shares a preference for using Claude Opus 4.1 for this process and demonstrates how to use the Metaprompt within a canvas environment. The speaker provides instructions to the AI to create a detailed step-by-step guide on building a personal brand in 30 days, emphasizing the AI's role as an expert. The speaker argues that If you're trying to do this with a YouTube video, you can't really do that in any other app.
17:55Creating a Prompt for a Personal Brand Coach
Creating a Prompt for a Personal Brand Coach
The speaker details the process of creating a prompt for a personal brand coach, breaking it down into three phases: context gathering, action plan, and coaching. The speaker emphasizes the importance of gathering context to enable the AI to provide the best coaching. The speaker argues that I know of softwares out there that are literally personal brand coaches, right? They take an AI chat and they put a prompt like this in there that coaches you on how to build a personal brand and they charge 30 to $50 a month.
20:28Review of the AI Utilization Process
Review of the AI Utilization Process
The speaker reviews the process of using AI effectively, emphasizing the importance of creating or extracting detailed expert-level instructions and avoiding AI guesswork. The speaker argues that in essence, when you use AI this way, you are using AI to both learn and build at the same time.
Part 4: Advanced Applications and Business Building
22:26Documenting Processes with AI and Intellectual Sparring Partner
Documenting Processes with AI and Intellectual Sparring Partner
The speaker encourages users to document their processes with AI, building a library of prompts for specific tasks. The speaker argues that by doing this, I think that you bring yourself to a higher level of thinking rather than a lower level. The speaker introduces the concept of an intellectual sparring partner, using AI to acquire deep knowledge by learning from high-level thinkers like Naval Ravikant and Daniel Schmachtenberger.
24:01Intellectual Sparring Partner Example
Intellectual Sparring Partner Example
The speaker provides an example of using AI as an intellectual sparring partner by breaking down the worldviews of Naval Ravikant, Daniel Schmachtenberger, Krishnamurti, and Mihaly Csikszentmihalyi. The speaker argues that I'm struggling with how I should best manage projects for a software company with a small team. Can you give me perspective on how I can best do this?
25:19Creative Thought Partner Prompt
Creative Thought Partner Prompt
The speaker introduces the concept of a creative thought partner prompt, emphasizing that thinking is not a random process and that successful individuals have specific processes for how they think best. The speaker argues that without using someone else's advice, what is an effective step-by-step process to overcoming the problem and moving toward the ideal life?
27:21Creating a Prompt for First Principles Thinking
Creating a Prompt for First Principles Thinking
The speaker details the process of creating a prompt that helps users arrive at clear, novel insights through first principles questioning. The speaker argues that the more you practice this, because habit formation comes through practice, the more you form the habit of first principles thinking.
28:42Building a Business with AI: Content, Products, Offers, and Copywriting
Building a Business with AI: Content, Products, Offers, and Copywriting
The speaker discusses using AI to build a business, emphasizing that it involves the same tasks as building a business without AI, but with the added benefit of a library of prompts to assist with various aspects like content creation, product development, offer creation, and copywriting. The speaker argues that you don't need to spend so much time learning the skills before you actually start building the business.
31:13YouTube Workflow with AI
YouTube Workflow with AI
The speaker outlines a YouTube workflow using AI, focusing on creating prompts for titles, key points, introductions, scripts, b-roll ideas, and video descriptions. The speaker argues that when you think of using AI for YouTube, you're kind of thinking like, OK, how do I have it create the entire video for me?
Part 5: Conclusion
33:37Conclusion and Call to Action
Conclusion and Call to Action
The speaker concludes by summarizing the key takeaways and encouraging viewers to provide feedback. The speaker argues that this is what I wish I knew when I had first started learning AI.
Sign in to continue reading, translating and more.
Open full episode in Podwise