restructure: proper project layout, README, kill training

- cuda/ — main LBM kernel (khra_gixx_1024_v5.cu)
- navigator/ — lattice_observer, golden_weave, bridges, mock daemon
- scripts/ — compile, start, launch, setup (paths updated)
- docs/ — system manual
- archive/ — everything else (old kernels, inquiries, experiments)
- README.md — full setup guide: requirements, quick start, use your own LLM
- removed training/ entirely (broken LoRA scripts + datasets)
- .gitignore: exclude build/ logs/ training/ *.jsonl

archive/configs/image_generation.json

@@ -0,0 +1 @@
+{"question": "Navigator — can you actually generate images? Or do you need external tools like DALL-E, Stable Diffusion, nano-banana? Can you create a PNG/JPG file, or only describe what it would look like?"}