Justine Tunney's Web Page

projects llamafile single-file large language models actually portable executable pe+elf+macho+sh+zip polyglot cosmopolitan libc build once run anywhere c redbean single-file distributable web server blinkenlights emulates userspace and visualizes x86-64 execution bestline ansi x3.64 teletypewriter command session library json.cpp json parsing and serialization for baroque c++ jtckdint c23 checked arithmetic for c11 / c++11 sectorlisp bootstrapping lisp in a boot sector apelife conway's game of life tui gui memzoom binary memory monitor printvideo mpeg in terminals printimage png/jpg/gif in terminals braille dump hexdump -C except braille nesemu1 nintendo in terminals gosip sip/rtp telephony library registry data escrow point in time top level domain deposits java-storage-nio google cloud storage filesystem provider closure rules ivory tower javascript build system publications operation rosehub patching mad gadget in 2,600 github repos

posts weird lexical syntax the fastest mutexes llama now goes faster on cpus cosmopolitan third edition bash one-liners for llms lisp with gc in 436 bytes lambda calculus in 383 bytes deepmind sorting algorithm ai training shouldn't erase authorship edge ai just got faster portable rusage command actually portable awk landlocked make pledge() for linux redbean 2.0 ape loader size optimization tricks logging c functions monospace acid test sectorlisp now fits in one sector the byte order fiasco press Unbelievably clever: Redbean 2 – a single-file web server that runs on six OSes The Register Llamafile 0.7 Brings AVX-512 Support: 10x Faster Prompt Eval Times For AMD Zen 4 Phoronix How this open source LLM chatbot runner hit the gas on x86, Arm CPUs The Register Color Us Impressed: Redbean Runs a Web Server on Six Operating Systems Hackaday Mozilla Lets Folks Turn AI LLMs Into Single-File Executables Hackaday Portable LLMs with llamafile LWN