
LOOM Day 13 — from "what the code does" to "how much"
LOOM is a tiny effect-typed language: a trust layer for AI-written code. Every function must honestly...
Tag archive

LOOM is a tiny effect-typed language: a trust layer for AI-written code. Every function must honestly...
Verbose is a small proof-carrying language whose compiler is written in Verbose itself. Its self-hosted emitter, fed its own 855 KB source, produced a binary bit-identical to the one that built it — gen1 == gen2, same sha256, same 1,302,980 bytes. The classic self-hosting fixed point, its strongest form. Here's what that means, and the failure — misdiagnosed, then actually understood — that almost stopped it.
Building a programming language with an AI coding agent. On June 10, 2026, driven by my...

Originally published at norvik.tech Introduction Deep dive into Ideaware's impact on...
During my telecommunications degree I took a course on signals and noise, I spent a lot of evenings...

Originally published at norvik.tech Introduction A deep dive into the emerging...

Originally published at norvik.tech Introduction An in-depth analysis of how simple...
Technical Reconstruction of the System Mechanisms The system under analysis...
Verbose is a small experimental language whose compiler proves properties (like termination) and emits tiny, readable x86-64 — no runtime, no GC, no libc. On a working branch, a rule written in Verbose itself now emits a standalone ELF that runs and prints its result. Here's factorial(5) from source to binary, plus an honest benchmark vs gcc/rustc/go.
Before a compiler can understand a program, it has to turn text into a tree (an AST). Here's how a small proof-carrying language represents a tree — recursive sum types in an index arena — and why it's the brick that makes everything else possible.

What happens when executable code reads like poetry? From Java verses to CSS narratives, this post breaks down the rise of code poetry, how to analyse it, and why programmers are writing poems that actually compile. Based on a 2015 research paper by Boris Orekhov
When a problem is complex enough, the right move is to build a language for that problem. SICP's most powerful idea.