Musings on AI and LLM
- A SYCL iGPU in a CUDA Box: Great for Building and Testing, Useless for Inference
- KV-Cache Quantization vs Positional Recall
- Tuning llama.cpp Batch Sizes on Intel Arc Pro B60
- How to Fix Sites Forcing Platform Passkeys (and Get Your QR Code Back)
- You're Already Using This Library (You Just Don't Know It)
- The Dead Upstream Trick: Forcing Nginx to Prefer Stale Cache
- The Global Tarpit: How to Stop 100k Cloud VPS Bots with Nginx and 128 Bytes of RAM
- fail2ban Not Matching Journal Entries from Unprivileged Users