The Fission of Writing: From Static Text to Dynamic Prototypes
At this stage, we are witnessing many paradigm shifts brought about by generative AI: chat, drawing, programming, and more. I focus on the act of writing, whose definition is quietly undergoing fission. In the past, writing meant producing static text; now, writing is about designing interactive digital experience prototypes. In the past, authors delivered content; now, authors deliver the rules for training AI.
"We need a medium that can simultaneously accommodate the ambiguity of human intent and the precision of machine execution—a dual text that serves as both a draft and source code."
The differences will manifest in interaction methods and the imagination for resource integration. Chat dialog boxes compress thought into linear prompts, serving as the "command line" of the AI era—usable, but far from sufficient. We need a medium that can simultaneously accommodate the ambiguity of human intent and the precision of machine execution—a dual text that serves as both a draft and source code, written for both humans and AI.
This is why Markdown deserves attention. It can be seen as a community-negotiated symbolic intermediary layer that allows for blank spaces and expansion. Those #, *, ![]() are not rigid syntax rules but dynamic agreements between humans and machines on "how to jointly understand a piece of text." The formatting is for smooth reading and workflow orchestration; the embedded blocks represent associative thinking and can also serve as microservice containers.
Input
Markdown File
+
Logic
AI Prompt
→
Output
Interactive HTML
Driven by this idea, I developed an experimental project. It started with a blog because a blog is the simplest MVP.
One Markdown file + one prompt → a self-contained HTML with styles and interactions.
This is its simplest form; the rest is up to the user. This is fundamentally different from the current flood of AIGC. AI-generated content often lacks a creative lineage and is easily categorized as spam, rejected by the entire internet. Here, Markdown is the source of truth. AI does not replace creators but enhances them in a human-centric way.
NamBlog: A Document Experience Compiler
NamBlog is an open-source prototype I built based on this idea. Readers are encouraged to try what happens when an article is no longer just formatted and beautified but serves as an "intent instruction."
In NamBlog, the built-in AI Agent and hot-reloadable prompt system form a real-time compilation environment. When you submit a document, the system, based on your instructions—"generate an interactive chart," "design an immersive reading page," or "build an animated card layout"—calls AI in real-time to compile the Markdown into an HTML application with complete styles and scripts.
For developers, this is also a modern technological playground. The backend is built on .NET 10, adopting a DDD (Domain-Driven Design) layered architecture. The data layer abandons traditional RESTful interfaces, fully adopting GraphQL, allowing the frontend to query data on-demand without being constrained by predefined template structures.
The world may eventually unify language, but not in symbols or pronunciation—rather, in the paradigm of transformation. This unification will occur in an extremely efficient manner because the AI revolution is underway. Under this unification, all languages are merely dialects, automatically translated into forms acceptable to the target audience.
If this revolution is not led by humans, people can at least seize some initiative—by providing raw materials. Not organic matter in the food chain, but informational elements. Humans can maintain their unique advantage in converting sensory stimuli and mental ripples into expressive content. This project is the prototype of that "converter," allowing your thoughts (Markdown) to be losslessly and polymorphically transformed into the forms needed by readers (or computers in the future).
RAW MARKDOWN SOURCEUTF-8
# The Fission of Writing: From Static Text to Dynamic Prototypes
At this stage, we are witnessing many paradigm shifts brought about by generative AI: chat, drawing, programming, and more. I focus on the act of writing, whose definition is quietly undergoing **fission**. In the past, writing meant producing static text; now, writing is about designing interactive digital experience prototypes. In the past, authors delivered content; now, authors deliver the rules for training AI.
> "We need a medium that can simultaneously accommodate the ambiguity of human intent and the precision of machine execution—a dual text that serves as both a draft and source code."
The differences will manifest in interaction methods and the imagination for resource integration. Chat dialog boxes compress thought into linear prompts, serving as the "command line" of the AI era—usable, but far from sufficient. We need a medium that can simultaneously accommodate the ambiguity of human intent and the precision of machine execution—a dual text that serves as both a draft and source code, written for both humans and AI.
This is why **Markdown** deserves attention. It can be seen as a community-negotiated symbolic intermediary layer that allows for blank spaces and expansion. Those #, *, ![]() are not rigid syntax rules but dynamic agreements between humans and machines on "how to jointly understand a piece of text." The formatting is for smooth reading and workflow orchestration; the embedded blocks represent associative thinking and can also serve as microservice containers.
Driven by this idea, I developed an experimental project. It started with a blog because a blog is the simplest MVP.
```
One Markdown file + one prompt → a self-contained HTML with styles and interactions.
```
This is its simplest form; the rest is up to the user. This is fundamentally different from the current flood of AIGC. AI-generated content often lacks a creative lineage and is easily categorized as spam, rejected by the entire internet. Here, Markdown is the source of truth. AI does not replace creators but enhances them in a human-centric way.
## NamBlog: A Document Experience Compiler
NamBlog is an open-source prototype I built based on this idea. Readers are encouraged to try what happens when an article is no longer just formatted and beautified but serves as an "intent instruction."
In NamBlog, the built-in AI Agent and hot-reloadable prompt system form a real-time compilation environment. When you submit a document, the system, based on your instructions—"generate an interactive chart," "design an immersive reading page," or "build an animated card layout"—calls AI in real-time to compile the Markdown into an HTML application with complete styles and scripts.
For developers, this is also a modern technological playground. The backend is built on **.NET 10**, adopting a **DDD (Domain-Driven Design)** layered architecture. The data layer abandons traditional RESTful interfaces, fully adopting **GraphQL**, allowing the frontend to query data on-demand without being constrained by predefined template structures.
## A Glimpse into the Future
The world may eventually unify language, but not in symbols or pronunciation—rather, in the paradigm of transformation. This unification will occur in an extremely efficient manner because the AI revolution is underway. Under this unification, all languages are merely dialects, automatically translated into forms acceptable to the target audience.
If this revolution is not led by humans, people can at least seize some initiative—by providing raw materials. Not organic matter in the food chain, but informational elements. Humans can maintain their unique advantage in converting sensory stimuli and mental ripples into expressive content. This project is the prototype of that "converter," allowing your thoughts (Markdown) to be losslessly and polymorphically transformed into the forms needed by readers (or computers in the future).