Hi reddit, i know you hate AI slop so i indeed write the intro myself! iam dev and curios about local inference and long hoirzon coding on my own box. last day-ish i let my local model (qwen 27b on llama.cpp, 2x 16GB cards) go on a long coding tour inside deepseek harness while i slept / steered. no
