When you are building, you are trying to close the gap between your taste and your execution. Only by building can you actually move toward where you want to be. Once you start, failing is inevitable. The struggle and the friction are not detours, they are where real learning happens.
AI shines at removing low value friction: repetition, formatting, and mechanical steps that add little insight once they are mastered. But building mental models, debugging when things break, decomposing systems, and making decisions under constraint are what create long term value. Those are the parts you still have to earn yourself.
Dan, I hope you don't mind, but I have some thoughts on your thesis here.
First, you're absolutely right about specific knowledge being the moat. That Naval quote is spot-on, and I think you've correctly identified the problem with commoditized information products.
But I think there's a dangerous assumption underlying this piece: that rapid iteration and short business model lifespans are inevitable rather than symptoms of competing in commoditized markets.
I've spent the last decade building precision timing infrastructure and AV control systems. Before that, 20 years in live sound engineering. The principles I learned mixing concerts in the '90s are still relevant today. The broadcast timecode fundamentals haven't changed in decades. When I'm debugging PTP synchronization issues, I'm applying knowledge that compounds over years, not iterates every quarter.
Here's my concern with the "build for tomorrow, not next year" mindset: if your business model expires in 2-3 years, you don't have a business—you have a job with extra steps.
The creator economy conflates disseminating information with creating value. You can wrap knowledge in AI chatbots, but you can't prompt-engineer the kind of specific knowledge that comes from 10,000 hours of actual practice. My brother is a retired rock drummer. You could feed every book on timing into an AI, create the world's best "drum coach" chatbot, and it still wouldn't capture what he knows from four decades behind a kit.
The real moat isn't speed of iteration. It's depth of expertise that takes years to develop and can't be commoditized by the next platform or tool.
What am I missing? Where do you see deep technical expertise fitting into this model?
Yes, true. I remember building a project with AI a few years ago, but I used my expertise to get what I wanted, and the results were excellent because I knew what I was talking about. If you don't have the expertise, you might do it, but it will likely fail because you don't have firsthand experience.
You captured a remarkable challenge to this piece, because what is proposed involves the conversion of expertise into information, which often requires taking embodied skills and making them explicit and transferable. Whether using an ai wrapper or selling a course, what one is really saying is that the impediment that needs to be solved for is access to knowledge about the skill and not the skill itself
Gee, that is weird for me to say! Don't encounter to many other Greg's. Sorry, at any rate if you are interested I lamented further on this sort of subject in one of my LinkedIn articles, check it out it's a good read:
Hi Greg. Yeah, I try to avoid LinkedIn. But I am interested in your thoughts. Also you could check out my substack where I am writing about the “difficulties” with LLMs, which are fascinating as they highlight (as you did) key human factors that the tech cannot address. Always happy to discuss not just “expound”
So many get caught up in the exhausting spin of creating timely, niche businesses that become obsolete in a matter of months. This will only get more exhausting as AI accelerates.
Better to zoom out and play a timeless game: design an entire category. A sandbox to play in and shape over time.
Create a new future for a group of people in this world who desperately need it.
You're right, Dan. Sure, the tools and formats for connecting and sharing knowledge are changing faster than a cat meme goes viral, but at the heart of it, the mission of sharing our genuine experiences remains unchanged. - Embracing that and leveraging AI is the key to not just surviving, but thriving in this changing world.
I had this thought at the begining of last year about student lead, ai assisted colleges. Where the courses were free and the content was infinite but specialized for each student. Imagine learning disibilities being healed in real time cuz the infinite need for special need teachers is AI implimented. And the currency for everything was creation. What did you create with all this information and direction? If learning felt less restricted and arranged by people we would never listen to, what would learning and art look like? So much of life is being told to learn and create in a specific way that rarely aligns with who we are individually. So your article definetly unlocked some pathways for me. Thank you.
This breakdown of the 6th stage of market sophistication is spot on. The part about specific knowledge being the moat really resonates - I've seen plenty of people launch "ChatGPT wrapper" apps that die within months because they just cloned something trendy without any unique perpective. Your point about building apprenticeship-style learning vs static courses is where the real opportunity sits. I tried building an interactive prompt system for a niche skillset last year and the engagment rate was 3x higher than anything I'd done with traditional video courses.
Dan, I've a question. What you're explaining sounds to me like a customized GPT with different instructions rather than an entire micro Saas like you discussed on the video and in the newsletter.
Correct me if I'm wrong, but it sounds like what you're describing is a GPT with a certain set of instructions based on the user's own personal experience and potentially some documents that the user has created based on their research and their experience, and then sold to the masses as an AI coach/mentor.
Would that not mean that you don't necessarily need to create an entire website or agent? It could just be a GPT that answers your questions and guides the user through a specific set of steps that the creator decides.
I think he means to create prompts for specific chatbots with a certain amount of questions/responses. I created one for self-reflection that I find very practical. You can talk about anything, but it has specific instructions that prompts it to have a concrete "personality" that I find especially useful when it comes to self-inquiry.
"Your years of reading, experimenting, failing, and figuring things out is the moat."
Exactly this. 💯
As a designer and builder working independently for 20 years, with a highly trained visual brain, experience creating brands and designing user experiences, including tons of failures, now I've perfect agent partners to develop my ideas, and turn them into one-person business.
I've started building tools to make fellow designers workflows easier. First one is a bookmarking app, that includes CanvasAI visual workspace, so they can easily capture knowledge and connect the dots in their projects. Build inspiration libraries with AI-powered organization.
For the last couple of months I've shipped more code than my whole career.
When you are building, you are trying to close the gap between your taste and your execution. Only by building can you actually move toward where you want to be. Once you start, failing is inevitable. The struggle and the friction are not detours, they are where real learning happens.
AI shines at removing low value friction: repetition, formatting, and mechanical steps that add little insight once they are mastered. But building mental models, debugging when things break, decomposing systems, and making decisions under constraint are what create long term value. Those are the parts you still have to earn yourself.
Great conclusion, I’m still finding my taste and knowing will evolve with time.
Dan, I hope you don't mind, but I have some thoughts on your thesis here.
First, you're absolutely right about specific knowledge being the moat. That Naval quote is spot-on, and I think you've correctly identified the problem with commoditized information products.
But I think there's a dangerous assumption underlying this piece: that rapid iteration and short business model lifespans are inevitable rather than symptoms of competing in commoditized markets.
I've spent the last decade building precision timing infrastructure and AV control systems. Before that, 20 years in live sound engineering. The principles I learned mixing concerts in the '90s are still relevant today. The broadcast timecode fundamentals haven't changed in decades. When I'm debugging PTP synchronization issues, I'm applying knowledge that compounds over years, not iterates every quarter.
Here's my concern with the "build for tomorrow, not next year" mindset: if your business model expires in 2-3 years, you don't have a business—you have a job with extra steps.
The creator economy conflates disseminating information with creating value. You can wrap knowledge in AI chatbots, but you can't prompt-engineer the kind of specific knowledge that comes from 10,000 hours of actual practice. My brother is a retired rock drummer. You could feed every book on timing into an AI, create the world's best "drum coach" chatbot, and it still wouldn't capture what he knows from four decades behind a kit.
The real moat isn't speed of iteration. It's depth of expertise that takes years to develop and can't be commoditized by the next platform or tool.
What am I missing? Where do you see deep technical expertise fitting into this model?
Yes, true. I remember building a project with AI a few years ago, but I used my expertise to get what I wanted, and the results were excellent because I knew what I was talking about. If you don't have the expertise, you might do it, but it will likely fail because you don't have firsthand experience.
You captured a remarkable challenge to this piece, because what is proposed involves the conversion of expertise into information, which often requires taking embodied skills and making them explicit and transferable. Whether using an ai wrapper or selling a course, what one is really saying is that the impediment that needs to be solved for is access to knowledge about the skill and not the skill itself
Hi Greg,
Gee, that is weird for me to say! Don't encounter to many other Greg's. Sorry, at any rate if you are interested I lamented further on this sort of subject in one of my LinkedIn articles, check it out it's a good read:
https://www.linkedin.com/pulse/performing-robots-when-experience-became-liability-gregory-meisel-x7eie
Hi Greg. Yeah, I try to avoid LinkedIn. But I am interested in your thoughts. Also you could check out my substack where I am writing about the “difficulties” with LLMs, which are fascinating as they highlight (as you did) key human factors that the tech cannot address. Always happy to discuss not just “expound”
I agree.
So many get caught up in the exhausting spin of creating timely, niche businesses that become obsolete in a matter of months. This will only get more exhausting as AI accelerates.
Better to zoom out and play a timeless game: design an entire category. A sandbox to play in and shape over time.
Create a new future for a group of people in this world who desperately need it.
You're right, Dan. Sure, the tools and formats for connecting and sharing knowledge are changing faster than a cat meme goes viral, but at the heart of it, the mission of sharing our genuine experiences remains unchanged. - Embracing that and leveraging AI is the key to not just surviving, but thriving in this changing world.
Agreed!
Mission, mission, mission. The brand isn't enough anymore. What does the brand stand for? What does it mean? I got so much out of this. Thank you.
True!
I had this thought at the begining of last year about student lead, ai assisted colleges. Where the courses were free and the content was infinite but specialized for each student. Imagine learning disibilities being healed in real time cuz the infinite need for special need teachers is AI implimented. And the currency for everything was creation. What did you create with all this information and direction? If learning felt less restricted and arranged by people we would never listen to, what would learning and art look like? So much of life is being told to learn and create in a specific way that rarely aligns with who we are individually. So your article definetly unlocked some pathways for me. Thank you.
This breakdown of the 6th stage of market sophistication is spot on. The part about specific knowledge being the moat really resonates - I've seen plenty of people launch "ChatGPT wrapper" apps that die within months because they just cloned something trendy without any unique perpective. Your point about building apprenticeship-style learning vs static courses is where the real opportunity sits. I tried building an interactive prompt system for a niche skillset last year and the engagment rate was 3x higher than anything I'd done with traditional video courses.
You are so generous. Thank you, Dan!
Tldr: build a micro saas using the help of AI and your content as the feeder
Si?
Brand build when you are on mission. Take one mission and win ✌️
Dan, I've a question. What you're explaining sounds to me like a customized GPT with different instructions rather than an entire micro Saas like you discussed on the video and in the newsletter.
Correct me if I'm wrong, but it sounds like what you're describing is a GPT with a certain set of instructions based on the user's own personal experience and potentially some documents that the user has created based on their research and their experience, and then sold to the masses as an AI coach/mentor.
Would that not mean that you don't necessarily need to create an entire website or agent? It could just be a GPT that answers your questions and guides the user through a specific set of steps that the creator decides.
I think he means to create prompts for specific chatbots with a certain amount of questions/responses. I created one for self-reflection that I find very practical. You can talk about anything, but it has specific instructions that prompts it to have a concrete "personality" that I find especially useful when it comes to self-inquiry.
A balanced article . Content material is carefully crafted. Caution is a must for a starter who is keen to build his career .
"Your years of reading, experimenting, failing, and figuring things out is the moat."
Exactly this. 💯
As a designer and builder working independently for 20 years, with a highly trained visual brain, experience creating brands and designing user experiences, including tons of failures, now I've perfect agent partners to develop my ideas, and turn them into one-person business.
I've started building tools to make fellow designers workflows easier. First one is a bookmarking app, that includes CanvasAI visual workspace, so they can easily capture knowledge and connect the dots in their projects. Build inspiration libraries with AI-powered organization.
For the last couple of months I've shipped more code than my whole career.
2,718 lines changed
1,856 additions / 880 deletions
50+ files touched
This is just in ~4 weeks.
Here is the result you can watch if interested:
https://youtu.be/p5JF93kAfrM
Thanks Dan, for the inspiration and motivation.
From one corner of the world to the other, words still have the greatest impact.
Progress comes from extracting the right lessons from life. Not all of them are useful. Some quietly cap your growth. Here’s how to tell which is which: https://open.substack.com/pub/hutchykc/p/how-to-know-if-youre-actually-growing?r=5uzq6q&utm_medium=ios&shareImageVariant=overlay
very insightful
The metric is "Time to First Result." If the AI helps a user write a draft in 10 minutes, that is the value. Track the output. Not the login time.