Hi, my name's Andy. I'm new here.
I’ve come to share a research project that I believe will upgrade everything that uses data — including how we write, execute, store, and train on it.
I’ve developed a new symbolic language system called Tribit. It compresses natural language, code, and commands into deterministic 36-bit glyphs. Each glyph represents a word or instruction using a fixed 6×6-bit layout, rendered visually or used directly for symbolic processing.
What Tribit does:
-
Live compression of strings into fixed glyphs — no parsing or tokenization needed
-
Reduces CPU/GPU load by removing string evaluation at runtime
-
Works across languages, operating systems, command-line logic, and AI models
-
Achieves 7.5× compression over plain text using visual or binary
.tbit
format -
Enables deterministic AI systems that reason over symbolic sequences, not stochastic tokens
Applications:
-
AI memory engines: Tribit replaces long string context with compact symbol memory
-
CLI and execution systems: Shell instructions are compressed into 1-glyph command chains
-
Code execution & scripting: Language-agnostic logic sequences encoded in symbolic form
-
Dataset compression: Convert .jsonl or .txt into compact, printable
.tbit
-
Low-resource hardware: Enables logical computing without traditional inference or compute cycles
I’ve released the Tribit Dev Kit for free:
https://triskeldata.au/shop/tribit-live-compression-tier-1-developer-license/
The full spec, font set, encoder, and symbolic vocabulary are available there.
I’m here to answer questions, receive feedback, and explore integrations with anyone interested in symbolic reasoning, AI compression, or execution-layer innovation.
Thanks for having me.