What is the new Moat in the new "AI" Vibe Coding" world
Its no longer effort and i'm not sure its expertise either, it might still be experience..
The moats of old are disappearing and im ok with that.
The thing I love about experimentation is it helps me coalesce some divergent ideas that have been floating in my head for a while, into some semblance of order.
A bit like writing does.
I wanted to get a handle on the latest state of “vibe coding” so decided to experiment with building an app using Claude Code and Opus 4.6.
I picked vibe coding an app for the Information Product Canvas, its an app I have wanted built for a while, but the cost to build it the old way never matched the value I found people were willing to pay for it.
Vibe coding in theory reduced the cost, the experiment was would it?
You can see the results of a few iterations of the canvas:
Plus a side experiment into the world of the Context Plane (couldn’t help myself).
But that not the point of this post.
Nick Pinfold is experimenting with his teams Agile Data Ways of Working and freely sharing his journey via LinkedIn comments.
In this comment
He talks about how he is creating a Streamlit app that allows him to capture the IPC content as Context and use it to assist with the next steps in their Information Factory.
Anna Bergevin posted a comment on this Substack post:
On how she is seeing “datapacks” that provide the content of a book in a way it can be easily used in an LLM as having some value.
Both of these show that there is value in writing books and creating apps, but that the typical moat of both of those things has changed.
Nick can vibe code a IPC app as fast as I can, if not faster
Anna can take the ePUB version of my book and use it in a LLM as fast as I can. If she pays more for tokens and the latest models than I do, she can do it faster and better than I can.
So effort and expertise are no longer the moat.
Given I have always said I “Can’t Code, Don’t Code, Won’t Code” but now I can now create an App but just asking questions, Im pretty sure Effort and Expertise is not the moat it was anymore either.
But maybe experience is.
To create my app in a way that meant it was actually useful, I had to have experience:
experience using the canvas
experience working with multiple data teams on the problem the IPC solves
experience using apps to know what UX features were needed
experience to know that I needed to add Google Auth to provide an easy login
experience to know not to store any API keys in a way that were public
experience to know that I wanted to use Google Spanner as the backend data repository and Svelte as the front end language
experience to know I wanted a API layer between the backend data repository and the front end app
experience to know …..
I don’t have an answer to what the new moat is, but I have more clarity on what it isn’t after these experiments.