Porting Karpathy's Autoresearch to JavaScript: A Study in Voluntary SufferingAt this point the pattern should be familiar. Someone publishes a clean, well-documented ML project in Python. I look at it. I think: "I have a GPU library. I have tensors. How hard could it be?" ThenMar 27 · 12 min readfredrikpaulin.hashnode.dev
Whisper in JavaScript: Speech Recognition on Apple Silicon, For ShameThe project that made me do this is called insanely-fast-whisper. It transcribes 150 minutes of audio in under 98 seconds using OpenAI's Whisper Large v3. The implementation is four lines of Python wrMar 26 · 11 min readfredrikpaulin.hashnode.dev
Smith: Building a GPU Acceleration Library in JavaScriptI had written a JavaScript GPT implementation called TinyFormer. It trained transformers from scratch using nothing but Float32Array loops. It ran on CPU. It was correct. It was also, by any reasonablMar 25 · 12 min readfredrikpaulin.hashnode.dev
WEBSCAN: Rewriting a Metasearch Engine From C to BunFredrik Paulin — March 2026 I had a metasearch engine written in C. It worked. It used libcurl for parallel fetching, libxml2 for HTML parsing, pthreads for concurrency, and OpenSSL for cache key hasMar 18 · 8 min readfredrikpaulin.hashnode.dev
AgentBus: Borrowing from 40 Years of Vehicle Networking to Build Non-Fragile Agentic SystemsThe multi-agent landscape has an infrastructure problem. We've gotten remarkably good at making individual agents capable — tool use, reasoning, planning — but the moment you need five, ten, or a hundMar 16 · 10 min readfredrikpaulin.hashnode.dev
