Show HN: A language model that lives in a boot sector
AI summary
Editorial policyA language model implemented with a 512-byte MBR inference engine, inspired by the boot sector demo scene.
Key points
- •The inference engine fits in a 510-byte MBR, with weights, vocabulary, and strings stored in subsequent sectors.
- •A small character-level neural network was trained on dmesg-generated logs.
- •The project aims to generate boot logs for non-existent hardware as an experimental demo.
#LLM / Model Trends#Research & Papers