The wild early years

Free
0/8 steps visited

Step 6 of 8

Open-weight — a different branch

Read · ~1 min

Open-weight — a different branch

Here's the part the subscription adverts skip: not all AI comes from a Silicon Valley subscription box.

Some organisations publish their model weights — the trained pattern file, the "brain" itself — for anyone to download and run. Meta's Llama family is the famous example. Note the difference: OpenAI sells ChatGPT as a product; Meta publishes Llama for you to run yourself. Two completely different deals.

Why that matters for you:

  • Run it on your own PC with a tool like Ollama (free, covered hands-on in Off the Grid — it's just software that runs AI locally).
  • Privacy by default — your prompts can stay on your machine; you're not feeding Google or OpenAI your daily life unless you choose to.
  • No per-message meter ticking away in the background.
  • Your guardrailsyou decide what it'll say, instead of inheriting a corporate morality filter you never picked.

The honest flip side: when you own it, you own the upkeep too — updates, basic safety, and not leaving a raw model exposed to the open internet without a second thought.

And local isn't automatically better. Smaller local models can be less capable than frontier cloud on genuinely hard tasks. So the course teaches when to reach for which — cheap, local models for the bulk grunt work; frontier cloud for the polish when a job truly needs it.

The install steps wait for Off the Grid. For now, just plant the flag: there is an own-it path here — self-sufficient, open-source-friendly, your knowledge on your hardware. That's the direction everything on this site points.

Continue — the recap.

Warning

Real power. Educational use only.

What we teach you to build is genuinely powerful — uncensored assistants, agents, and automations on your own hardware. In the wrong hands, that is as dangerous as malicious code in the wrong hands. We do not teach illegal, malicious, or harmful use. You are responsible for what you deploy.

See what we mean →