<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Llama-Cpp on Soil &amp; Silicon</title><link>https://soilandsilicon.dev/tags/llama-cpp/</link><description>Recent content in Llama-Cpp on Soil &amp; Silicon</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Fri, 22 May 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://soilandsilicon.dev/tags/llama-cpp/index.xml" rel="self" type="application/rss+xml"/><item><title>Optimizing a Chatbot on macOS, Part 2: Baseline</title><link>https://soilandsilicon.dev/posts/optimizing-a-chatbot-on-macos-part-2/</link><pubDate>Fri, 22 May 2026 00:00:00 +0000</pubDate><guid>https://soilandsilicon.dev/posts/optimizing-a-chatbot-on-macos-part-2/</guid><description>Install llama-cpp-python with Metal support, download Qwen2.5 14B, write a minimal decode loop, and measure our baseline TTFT, ITL, and eval score.</description></item><item><title>Optimizing a Chatbot on MacOS, Part 1: Introduction</title><link>https://soilandsilicon.dev/posts/optimizing-a-chatbot-on-macos-part-1/</link><pubDate>Thu, 21 May 2026 00:00:00 +0000</pubDate><guid>https://soilandsilicon.dev/posts/optimizing-a-chatbot-on-macos-part-1/</guid><description>Introducing a practical series on making a local LLM chatbot fast on macOS and eventually on any laptop.</description></item></channel></rss>