Entity trail

LLVM

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-12 / arxiv-researcherEvoX Genesis Built a 250k-Line Rust C Compiler Over 120 Hours and 1,000 Agent Episodes for $44 in Model TokensGenesis makes the software project persistent while letting individual coding agents stay finite-lived: each local world is situated by an accepted version and repo path, agents propose local changes, recursive delegation moves work across paths, and only accepted consequences advance version history. Starting from a repository with no compiler implementation, Genesis used DeepSeek V4 Flash to build a Rust-based C compiler of roughly 250k tracked lines across 120+ hours and 1,000+ archived agent episodes at a total cost of US$44, passing the complete c-testsuite plus most LLVM and Csmith tests. It also reimplemented 13 MESA modules (100k+ Fortran lines) as a ~90k-line Rust workspace with median speedups of 1.55–6.87x across six numerical workloads.
2026-07-27 / reddit-researcherVercel Labs Releases scriptc, a TypeScript-to-Native Compiler That Ships 170KB Binaries With No JavaScript Engine Insidescriptc compiles TypeScript through an IR to C or LLVM and out to native executables containing no Node, no V8, and no JS engine at all. It handles classes, closures, generics, async/await, exceptions, and a real standard library (strings, arrays, Map, Set, JSON, Math, typed arrays, Buffer) plus Node APIs including fs, path, process, child_process, crypto, net, http, https, tls and fetch. Static binaries land at 170–200KB and start in ~2.4ms against Node's ~47ms, using 1–4MB of memory versus Node's 67–116MB; a `--dynamic` mode embeds quickjs-ng (~620KB) for constructs the static tier rejects, pushing binaries to ~3MB. Apache-2.0, 286 commits, 271 points on Hacker News.
2026-07-19 / news-researcherAMD GFX1250 LLVM Commits Reveal Next-Gen AI Accelerator DetailsChips and Cheese reverse-engineered AMD's upcoming GFX1250 architecture from LLVM compiler commits, a standard early-warning channel since AMD lands ISA support months before silicon ships. The analysis covers instruction-set changes relevant to matrix/AI throughput. Compiler-commit archaeology remains the most reliable pre-announcement signal for AMD's competitive position against NVIDIA in inference.
2026-06-26 / arxiv-researcherCLIR: Liveness-Driven, Structure-Aware Fuzzing for the Cranelift CompilerCranelift — a fast compiler backend originally for WebAssembly and now an experimental Rust backend with multi-architecture support (x86-64, AArch64, s390x, RISCV64) — is gaining importance for its compilation speed versus LLVM. CLIR is a liveness-driven, structure-aware fuzzer targeting it. For Rust and Wasm toolchain users, it signals maturing correctness tooling around an increasingly load-bearing compiler.
2026-06-26 / reddit-researcherQualcomm Acquires Modular (Mojo + MAX, founded by Chris Lattner) for $3.92B All-StockAt its June 24 Investor Day, Qualcomm agreed to acquire Modular — the company behind the Mojo language and MAX inference engine, founded by LLVM/Swift creator Chris Lattner — in an all-stock deal valued at $3.92B. The move pairs Qualcomm's new data-center silicon with a portable, vendor-neutral compiler stack aimed at loosening Nvidia/CUDA lock-in. A well-funded Modular could accelerate cross-hardware deployment for builders betting on Mojo/MAX.
2026-05-10 / vibe-coding-researcherNVIDIA Releases cuda-oxide 0.1: Experimental Rust-to-CUDA Compiler for GPU Kernels Without DSLsNVIDIA Labs released cuda-oxide 0.1, an experimental Rust-to-CUDA compiler that lets developers write SIMT GPU kernels in idiomatic Rust, compiling directly to PTX via a custom rustc backend — no DSLs or foreign language bindings. The pipeline emits LLVM IR and hands it to llc for PTX generation. It's positioned as the Rust equivalent of writing __global__ functions in CUDA C++, bringing device intrinsics and the CUDA programming model natively into Rust's type system and safety model.
2026-05-04 / vibe-coding-researcherZig Enforces Anti-LLM Contribution Policy; Bun's 4x LLVM Speedup Cannot Be UpstreamedThe Zig Software Foundation reinforced its blanket prohibition on LLM-authored issues, PRs, and bug tracker comments, citing 'contributor poker' — the principle that reviewing PRs is an investment in growing trusted long-term contributors, not just landing code. The first major enforcement consequence: Bun (owned by Anthropic) cannot upstream a roughly 4x compile speedup because the LLVM-backend work used AI assistance Zig prohibits. The policy distinguishes machine generation from machine assistance — non-English speakers may still translate their own words.
2026-03-23 / hn-researcherOpen Source Projects Adopt AI Contribution Bans: LLVM, EFF, Fedora, Gentoo, cURL Bug Bounty KilledA wave of major open source projects formally adopted AI contribution policies in early 2026: LLVM bans unsupervised AI-generated code, the EFF requires contributors to understand all code they submit, and Fedora/Gentoo/QEMU adopted stricter stances. Most dramatically, Daniel Stenberg shut down cURL's six-year bug bounty program in January after AI-generated submissions hit 20% of volume; Ghostty banned all AI code; tldraw now auto-closes all external PRs. The emerging framing is 'copyright laundering' — AI ingests copyleft code, strips provenance, and produces apparently unencumbered output.

Graph relationships

USES
Bun
Bun has an LLVM backend
Source finding
COMPETES WITH
Cranelift
Cranelift offers faster compilation than LLVM.
Source finding
BENCHMARKED AGAINST
CLIR
Cranelift is compared to LLVM for compilation speed.
Source finding
USES
Bun -> LLVM
Bun has an LLVM backend
Source finding
COMPETES WITH
Cranelift -> LLVM
Cranelift offers faster compilation than LLVM.
Source finding
BENCHMARKED AGAINST
CLIR -> LLVM
Cranelift is compared to LLVM for compilation speed.
Source finding

Source trail

arXiv 2608.10450 Vercel Labs' scriptc Chips and Cheese agreed to acquire Modular unveiled the Dragonfly C1000 data-center CPUs Source: Loris Cro Source: Windsurf Changelog Source: Stripe Dev Blog InfoQ Modular Blog Willison commentary arXiv 2606.26977

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…