A personal blog covering fitness, nutrition, AI tools, and self-hosted tech on a Raspberry Pi

PPL vs 5/3/1: How Each Program Applies Progressive Overload

Your muscles do not care how hard you worked today. They care whether today was harder than last time — that's called progressive overload. Here's the principle behind every good program, and how PPL and 5/3/1 put it to work.

Creatine for Fat Loss: What It Actually Does to Your Reps, Muscle, and Brain

Everyone talks about creatine for lifting. Almost nobody talks about the extra reps, how it helps with fat loss, the muscle it protects during a cut, or what it might be doing for your brain along the way.

The SEO MCP Servers That Audit My Site Every Morning

I used to check Search Console once a month if I remembered. Now two MCP servers do it every morning before I've had coffee.

How to Lose Fat Without Losing Muscle: The 5 Rules That Actually Matter

Most diets burn muscle right along with fat. Here are the 5 rules I use to lose fat without losing muscle: protein targets, training, and pacing that actually holds up.

MCPs(Model Context Protocol): Peripherals for the AI

I used to think of AI as one big brain typing back at me. Then I plugged in an MCP(model context protocol) server and understood, these are peripherals, and they change everything.

RAG on a Raspberry Pi: Building a Private Assistant with pgvector + Ollama

I finally gave my local Pi assistant something to actually know. Here's how pgvector and Ollama turned a chatbot into a private assistant with real answers.

I Tracked My Health with AI for 30 Days: What Actually Changed

I spent 30 days feeding my health data to an AI every single morning. Some things genuinely changed. Some things were pure theater. Here's the honest breakdown.

Sleep Tracking in 2026: What the Data Actually Tells You (And What It Doesn't)

A tracker can tell you how you slept. It can't make you sleep. Here's what the 2026 sleep tracker data actually means, and the ritual that does the real work.

Running a Local LLM on a Raspberry Pi 5 (or Home Server)

Everyone assumes local AI needs a GPU. Here's how I got a local LLM running on a Raspberry Pi with Ollama, what speed to actually expect, and where the limits are.

RAG: AI Gets a Personal Touch

RAG (retrieval augmented generation) is how AI answers questions about your own data. Here's how it works, why chatbots on websites use it, and why it beats fine-tuning.

Consistency Beats Motivation: Winning Daily

Motivation fades. Consistency compounds. Here's the actual math behind 1% daily improvement, why your body responds to repetition over bursts, and how to keep showing up even on your worst days.

OpenCode for AI Coding: Evade Vendor Lock-In

OpenCode is the open source, model agnostic coding agent that lets you plug in Claude, GPT, Gemini, or your GitHub Copilot subscription without rewriting your workflow. Here's the full setup guide.