KCC is a formally specified C17 compiler implemented entirely in ARM64 assembly and expressed as a single literate program using weft. It parses and checks C against the Standard, performs optimisations and register allocation, emits native machine code, and is self-hosting - producing a byte-identical recompiled binary. KCC cross-compiles to x86-64 on Linux and macOS, builds nontrivial software such as Lua, SQLite, and DOOM, and is capable of booting a Linux kernel. The entire implementation and explanation is published as a roughly 1,000-page typeset book: the .weft literate sources are the sole source of truth, every assembly instruction is preceded by prose, generated build artifacts are never edited directly, and the committed kcc.pdf provides a readable snapshot of the current text.
The build and verification workflow is deterministic and reproducible: tangling the weft sources produces a generated tree (gen/) and Makefile, then make assembles gen/kcc and runs extensive test targets (difftest, diagtest, codegen/e2e/interop, self-host, meta-c and a full verification CI). make pdf weaves and typesets the book with LuaLaTeX; optional steps use rsvg-convert for covers, Docker and qemu-user for kernel reproduction. Prerequisites include weft ≥2.0, a C toolchain’s as/ld, and a TeX distribution. The project is MIT-licensed and authored by Cuauhtémoc Pacheco with an AI collaborator (Claude), advancing a literate-driven approach to verified compiler construction.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
AstaBrief is an open-source model designed for fast scientific report generation, capable of turning research questions and literature into cited reports. It reduces generation time by nearly three and a half times compared to proprietary models, making it a useful tool for researchers. (allenai.org)
Shrivu Shankar argues that SaaS companies are evolving into 'harnesses,' where core tasks are performed by AI agents rather than humans. This shift transforms organizational structure, with humans overseeing and reviewing AI-driven processes within the company. (blog.sshh.io)
Three AI agents - Meta’s Muse, Anthropic’s Claude, and OpenAI’s GPT - were tested on multilingual research tasks involving U.S. and Iranian data. The experiment focused on how language and context influence agent performance across reasoning, sourcing, and artifact creation. (royapakzad.substack.com)
An open-source project named LDRAW-NOVA allows users to generate Lego models using AI. The repository contains tools, examples, and documentation for creating Lego designs with artificial intelligence. (github.com)
DwarfStar 4 (ds4) is a local inference engine that runs large language and vision models on high-memory Mac, CUDA, and ROCm machines. It supports models like DeepSeek V4 and Qwen3.8, offering CLI, APIs, and persistent cache features for efficient local deployment. (dwarfstar.sh)
Rai is a CPU-only large language model inference engine written entirely in Rust. It aims to provide efficient and pure Rust solutions for LLM deployment without relying on GPUs. (github.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.