How about the fact that, in the new loop, all documents become machine-computable because natural language processing is moving to the apex of computing's UI/UX stack?
That's the fundamental shift brought about by AI.
For decades, the GUI sat at the apex: humans had to navigate screens, menus, forms, buttons, and application-specific workflows to tell software what to do.
Now, natural language becomes the primary interface. The GUI is demoted to a secondary, purpose-specific visualization and interaction layer that can be generated as needed.
Yes, you still need domain expertise to deliver quality products. But the cognitive space gained through AI shifts human effort away from operating software and toward what actually matters: verification, judgment, and taste.
AI is making product judgment more important, not less. I really agree that the core of product management hasn’t fundamentally changed. As building becomes faster and easier, the harder question becomes: “Should we build this?” Understanding users and solving the right problem may matter more than ever.
Build by Tony Fadell who worked with Steve Jobs to build the iPod, devotes a substantial portion of its text to the importance of storytelling in product building, and how to do it well. Incredible read
But the following statement is too broad and betrays the Coder's Conceit:
"There’s no question that the cost of making stuff has collapsed. "
No, the cost of making stuff in the world of bits has collapsed. The cost of making stuff in the world of atoms and organisms has dropped in many areas, but not nearly as dramatically.
I really lived this article and even shared it on my substack! Do you feel this closely related more to product also becoming about brand strategy? it feels in line of the trend where the frontier labs are hiring content creators
Don't measure how many people finished the flow, measure how many came back and took the core action. That's the same mistake I see in commercial teams, just wearing different clothes. Demos booked, pilots signed, LOIs collected. Every one of those is a completion metric gathered at the exact point where nothing has been proven yet, and every one of them looks like progress on a board slide.
The pilot is the onboarding flow. A customer who finishes one has told you they were curious. Whether they came back and did the thing again, with their own budget, is the only number that means anything.
AI can make building faster, but it cannot tell you what’s worth building, that still comes down to judgment, and probably matters even more now that the cost of making things is so low.
This is honestly one of the most inspiring articles I’ve read in a long time.
One of the things we’ve struggled with the most is how to convey our vision clearly to the end user. And when users don’t immediately “get” the product, the natural reaction is to think: maybe we just need more features.
So you start building more and more. It feels productive because you’re doing something, but in reality you can easily end up like a hamster on a wheel, running faster without actually getting anywhere.
This article really helped me see that more clearly. The problem isn’t necessarily that we need more features. We need to communicate the vision much better.
Thank you for writing this. Genuinely super inspiring, and I’m going to start applying these recommendations immediately.
"Product management is about telling the story of people who are going to use the product, and why it will matter in their lives"
Telling stories is the core job of product manager? Really?
At the end of the day, any product is a set of choices - some emphasized, some de-emphasized, now, and in the future. Isn't that at the core the job of a product manager?
How about the fact that, in the new loop, all documents become machine-computable because natural language processing is moving to the apex of computing's UI/UX stack?
That's the fundamental shift brought about by AI.
For decades, the GUI sat at the apex: humans had to navigate screens, menus, forms, buttons, and application-specific workflows to tell software what to do.
Now, natural language becomes the primary interface. The GUI is demoted to a secondary, purpose-specific visualization and interaction layer that can be generated as needed.
Yes, you still need domain expertise to deliver quality products. But the cognitive space gained through AI shifts human effort away from operating software and toward what actually matters: verification, judgment, and taste.
It will always be about telling stories.
10000% yes
AI is making product judgment more important, not less. I really agree that the core of product management hasn’t fundamentally changed. As building becomes faster and easier, the harder question becomes: “Should we build this?” Understanding users and solving the right problem may matter more than ever.
Build by Tony Fadell who worked with Steve Jobs to build the iPod, devotes a substantial portion of its text to the importance of storytelling in product building, and how to do it well. Incredible read
Nice article.
But the following statement is too broad and betrays the Coder's Conceit:
"There’s no question that the cost of making stuff has collapsed. "
No, the cost of making stuff in the world of bits has collapsed. The cost of making stuff in the world of atoms and organisms has dropped in many areas, but not nearly as dramatically.
Informative..Nice..
Informative..Nice.
I really lived this article and even shared it on my substack! Do you feel this closely related more to product also becoming about brand strategy? it feels in line of the trend where the frontier labs are hiring content creators
Six or eight turns a year made specs the filter for scarce engineering time. Now judgment is scarce, and the story moves to your eval set.
Don't measure how many people finished the flow, measure how many came back and took the core action. That's the same mistake I see in commercial teams, just wearing different clothes. Demos booked, pilots signed, LOIs collected. Every one of those is a completion metric gathered at the exact point where nothing has been proven yet, and every one of them looks like progress on a board slide.
The pilot is the onboarding flow. A customer who finishes one has told you they were curious. Whether they came back and did the thing again, with their own budget, is the only number that means anything.
AI can make building faster, but it cannot tell you what’s worth building, that still comes down to judgment, and probably matters even more now that the cost of making things is so low.
This is honestly one of the most inspiring articles I’ve read in a long time.
One of the things we’ve struggled with the most is how to convey our vision clearly to the end user. And when users don’t immediately “get” the product, the natural reaction is to think: maybe we just need more features.
So you start building more and more. It feels productive because you’re doing something, but in reality you can easily end up like a hamster on a wheel, running faster without actually getting anywhere.
This article really helped me see that more clearly. The problem isn’t necessarily that we need more features. We need to communicate the vision much better.
Thank you for writing this. Genuinely super inspiring, and I’m going to start applying these recommendations immediately.
The balance needed is between telling stories and asking questions!
Couldnt agree more. Here is what we just shared : https://substack.com/@mohree/note/c-335929988?utm_source=notes-share-action&r=937wah
I am so with you, because look...
at the evolution of our species!
In a sense, we became homo sapiens we are, because we started to make tools.
Tools -> nearly products!
Tools: getting from point A to point B with a way that does not equal the materials the tools are made of.
Products: get people to the abstract result with a matter that is quite abstract too (code) in 2026.
To make people believe they will get there, we need stories. Don't you think?
"Product management is about telling the story of people who are going to use the product, and why it will matter in their lives"
Telling stories is the core job of product manager? Really?
At the end of the day, any product is a set of choices - some emphasized, some de-emphasized, now, and in the future. Isn't that at the core the job of a product manager?