Resources / Go deeper

Local models (Ollama, LM Studio)

Picks checked October 2026.

A local model runs on your own machine, so nothing you paste leaves it. It's the answer for shops that won't allow cloud AI, with a catch: small local models are slower, and they fail in quiet ways.

Where I'd go

Ollama Quickstart

Docs · Free

The command-line path most local-model guides assume. Start here before any tutorial.

LM Studio docs

Docs · Free

The app path, and the easier one on a work laptop where you want to see what's loaded.

Simon Willison, local LLMs

Blog · Free

A working engineer's running notes on which open models are worth trying, dated per post. This one's for staying current, not learning from zero.

Mine, for the basics

Local AI for Network Engineers: what actually works, measured

A free model on my own hardware, fed real configs with zero redaction, and the four silent failures that make local AI look broken. Network-specific, measured, and it needs no lab.

I cover the network engineer basics there.