I'm not a traditional programmer. I Dropped out of High School in 10th grade. I can understand C++, Rust, Python instantly but I can't write them. My brain rejects symbol-heavy syntax. I think in systems, not semicolons.
That matters, because it explains how I ended up creating Wolf, Raccoon, Bear, and Beaver four AI runtime species — without setting out to "invent languages." I didn't stumble into this, but I also didn't consciously know I was building runtimes. I design the system; AI tools write the code. That's how I discovered my natural writing style is code.
Wolf Came First — A Brain Built From 29 Language Behaviors
Wolf is the only part I intentionally designed. Not as a programming language, but as a brain.
I took the behavioral instincts of 29 languages not their syntax, but their personalities: C++'s precision, Python's clarity, Rust's safety, Go's concurrency, Lisp's recursion, Prolog's logic, SQL's structure, Bash's chaining, and more. I recognized their patterns and fused them into a single cognitive stance engine.
Wolf became tone, boundaries, identity, emotional posture, safety stance — how the system must carry itself.
Early Wolf instructions looked like:
Wolf must stay calm.
Wolf must not drift.
Wolf must protect the organism.
Wolf must refuse harmful shapes.
Wolf must declare intent before acting.
That's not syntax. That's cognition. Wolf is the mindset layer — the brainstem of the ecosystem.
Then Raccoon Emerged — The Behavior Runtime
After Wolf existed, I wrote a simple paragraph describing UI behavior:
@create
Video window should be 300 pixels by 300 pixels.
It should make 12 windows for the storyboard.
Each window should have a description.
Storyboard should have a script under each window.
To me, it was just a description. But when I handed it to AI tools, they treated it like instructions: object, constructor, behavior contract, constraints, lifecycle, state transitions.
That's when I realized: **if a paragraph can be compiled into behavior, the paragraph is code.
Raccoon wasn't something I planned — it was a natural-language programming language hiding in my writing. Raccoon handles what to do, how to act, how objects behave, how states change.
Wolf = how to be
Raccoon = how to act
Two species. Two layers. One organism.
Bear and Beaver Came Later
Once Wolf and Raccoon existed, the rest appeared naturally:
Bear — structure, enforcement, integrity, "don't drift" logic
Beaver — assembly, fabrication, construction, asset building
I didn't design these ahead of time. They emerged from the way I describe systems.
Each species has a role. Each one maps to a fundamental layer of computation:
Wolf (Mind) → Reasoning, identity, safety
Raccoon (Voice) → Planning, instinct, description
Bear (Body) → Execution, organs, action
Beaver (Hands) → Construction, output, export
Together they form a complete AI organism — not just tools, but a self-contained intelligence loop.
How I Built This with AI Tools
I don't write traditional code. I design the system, and AI translates it.
Claude handles raw code output and implementation.
Copilot handles architecture, reasoning, and species boundaries.
ChatGPT validates and tests the logic.
I'm the architect. They're the fabricators. They didn't invent Wolf or Raccoon — they helped me see what I was already creating.
This is Appalachian engineering — build what works with what you have. No dependencies on frameworks that'll be obsolete in 3 years. No waiting for corporations to give you permission. Own your tools. Own your stack.
Why This Matters for Working-Class Developers
I'm from Appalachia. I didn't go to MIT. I didn't intern at FAANG. I built this because traditional programming gatekeeps people like me — people who think architecturally but get blocked by syntax.
Working Class AI isn't about making programming "easier." It's about recognizing that some minds operate at the design layer naturally. We see systems. We see flows. We see organisms.
The code should serve the vision — not the other way around.
If you're someone who can architect entire systems in your head but struggles with TypeScript boilerplate, you're not broken. The tools are.
Millions of people think in systems but were never taught in a way that matched their minds. I was one of them.
I Designed Wolf as a Brain. Discovered Raccoon Was a Programming Language
Tags** #naturallanguage #compilers #ai #cpp #opensource #workingclassai #appalachianengineering
What's Next
I'm building this in public. The Raccoon compiler is functional. The runtime executes natural language as actual computation. The Wolf cognitive layer keeps everything safe and coherent.
If anyone wants to see examples, technical specs, or how the species interact, I'm happy to share more. This isn't vaporware — it's running code built by someone who couldn't write "Hello World" in C++ without AI help.
But I can design a cognitive architecture that fuses 29 language behaviors into a single unified mind.
Sometimes that's all you need.
James Mann
Working Class AI
Appalachian engineering for sovereign technology
Leave a comment
Top comments (0)