Ideas sobre tecnología, liderazgo, productividad y crecimiento personal en la industria tech.
Página 13 de 34
I put Ollama and Open WebUI on the same 8GB M1 Mac Mini that runs my image generation, gave it the Metal GPU, and tuned it to fit. It worked, with one honest catch: 8GB is the ceiling, and that ceiling decides everything.
I built Syncthing to run entirely on my LAN and even wrote that you should trust the listen addresses, not the checkboxes. Months later my Pi-hole logs caught it resolving the public discovery servers anyway. The checkbox wasn't enough.
One of my three Pi-holes was handling 99.9% of DNS queries while the other two sat idle. The fix I wanted (even load balancing) turned out to be something RouterOS just won't do, and the reason is which firewall chain the queries land in.
Building a two-player courtroom card game in Godot using Cursor IDE's custom modes and AI agents. A hackathon experiment in AI-assisted development.
Running Stable Diffusion locally on an 8GB M1 Mac Mini. Why I picked ComfyUI, why the GPU work has to run native instead of in Docker, and the FP16-plus-lowvram tuning that kept it from crashing.
Página 13 de 34
