Running a 16x faster local LLM on Apple Silicon broke my illusion of the corporate oracle
-
Look, I’m a draftsman in Chicago. My days are usually spent arguing with CAD software and trying to keep my rent from swallowing my soul. Last week, I got tired of paying subscription fees to a tech giant just to ask it for study tips and tarot spreads. So I downloaded a quantized model, plugged in my M3 MacBook, and watched it chew through tokens at 16x the speed I was used to. Suddenly, the "oracle" wasn’t living in some Nevada data center. It was sitting on my desk, breathing the same damp winter air as my cactus.
It sounds nerdy, I know. But spiritually? It hit me like a lightning rod. We’ve spent years treating AI like this monolithic, almost god-like entity. You feed it prompts, it spits out prophecy, and you bow to the algorithm. That’s classic Carl Jung projection right there. We externalize our need for structure and certainty onto shiny black boxes because it’s easier than doing the actual inner work. When the model runs locally, that illusion cracks. You see the weights, the parameters, the sheer mundane math of it. It stops being a deity and starts being a tool. Or, if you prefer the occult angle, it’s finally behaving like The Magician card should: raw potential sitting right in front of you, waiting for your intent to direct it. No corporate gatekeepers, no algorithmic bias baked in by a boardroom of shareholders.
I’ll be honest, though. I’m not buying into this whole "decentralized AI consciousness" hype train. It’s just code. But running it locally feels like a necessary corrective to the spiritual bypassing I see in our community. Too many folks treat divination like a vending machine. You drop a coin (or a prompt), you get an answer, you skip the messy integration phase. Local inference forces you to actually engage with the output. You curate the context window. You tune the temperature. Running it feels less like querying a server and more like working with a polished quartz point—you have to clean it, set the charge, and direct the focus yourself. That’s energy work, plain and simple.
Anyway, I’m running it on a 32GB machine and it’s actually usable for journaling and pattern recognition. Has anyone else tried spinning up local models for divination or energy mapping? I’m curious if you feel the same shift in agency, or if I’m just over-caffeinated and romanticizing a bunch of matrix math.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login