Pedro’s 5090 Upgrade: Full Migration Recap + LLaMA.cpp Setup on WSL

In this video, I walk you through migrating Pedro, my AI bot, from a refurbished Windows machine to a brand-new RTX 5090-powered Homelab setup.

We cover:

Unboxing the MSI PC with RTX 5090 🔥

Setting up the OS, WSL, and Tailscale

Installing and running LLaMA.cpp (community-chosen via Discord poll!)

Serving the model across the private tailnet with Tailscale Serve

Benchmarking the upgrade: from 11 tokens/sec ➔ 94 tokens/sec 🚀

Pedro is now fully GPU-accelerated, lightning fast, and live inside the SoyPete Tech Discord — where you can interact with him directly using /askpedro!

🖥️ Want to build your own AI HomeLab? Check out my Raspberry Pi AI setup too:
👉 AI on a Raspberry Pi - Full Tutorial

🌐 Connect with the SoyPete Tech community:

🔗 Linktree (courses, streams, projects): https://linktr.ee/soypete_tech

💬 Join the Discord (ask Pedro yourself!):   / discord  

👋 If you enjoy HomeLab builds, AI experiments, and self-hosted tech—subscribe and follow along!

#RTX5090 #Homelab #LocalAI #SoyPeteTech #LLM #PedroBot