asquithdenardis

been running llama.cpp on a single $40/mo box for three months and it’s literally printing money—20 ms latency, 1k rps, no drama. compiled with cuBLAS, pinned 70B q5_K_M, slapped nginx in front, health checks via /health, done. battle scars: watch the mmap + NUMA dance and keep context length under 4k or you’ll swap to the moon. pro tip: turn on --tensor-split if you’ve got two GPUs, free 30 % speed bump. side hustle uses the same stack to batch-generate 60-min ASMR vids every night—because who doesn’t want to get paid while they sleep? https://asmrvideo.org

The small-web never left—it just got buried. Three zero-friction ways to dig it back up:

1. Marginalia Search → pastes only lightweight, hand-curated pages; add your keyword in quotes. 2. Feedle.org → RSS-only index; most 5-reader blogs still ping it. 3. `site:blogspot.com inurl:2025` on DuckDuckGo → bypasses Google’s engagement layer.

If you want to see these forgotten sites instead of just reading them, I built a micro-tool that screenshots every result and auto-caption it with FLUX Kontext (https://flux-kontext.io). Ten lines of code—grab it from the repo linked on the landing page.