FPR EDU — Learn the Build
“The Build Generates the Curriculum, and the Curriculum Becomes the Next Ranch.”
1. Why We Build
A family-led mountain property project in Pound, Virginia, documenting practical land stewardship, useful skills, resilient infrastructure, and community-building work.
2. How We Fund It
Merchandise and original Art by Abbe apparel help support ranch infrastructure.
3. Build Your Own (EDU)
Learn the exact blueprints, local AI setups, and trade skills so you can replicate this architecture for your own farm or project.
[START HERE] Find Your Starting Point — Adaptive Intake & Pretest
Tell Gary your background and goals to get an instant domain placement and personalized learning plan. No generic one-size-fits-all tracks.
How We Built Gary: Foundations of a Sovereign AI System
A practical, evidence-backed curriculum covering how to build an autonomous, local-first GPU node with zero recurring cloud subscription lock-in.
Module 1: The Local-First Sovereign Philosophy
Why dependence on closed cloud APIs is fragile. The core doctrine of local execution, data sovereignty, and fenced-yard governance.
Module 2: Sovereign Hardware Architecture
Evaluating the Ryzen 9 3950X, 64GB RAM, and RTX 5060 Ti 16GB VRAM budget. Storage allocation between OS (C:), Workspace (S:), and Blobs (D:).
Module 3: Local GPU Model Serving with Ollama
Quantized GGUF inference. Running Qwen 2.5 14B, Gemma 4 E4B, and Qwen 3.5 27B locally on Port 11434.
Module 4: Unified Model Gateway with LiteLLM
Configuring LiteLLM (:4000) as the central OpenAI-compatible endpoint with transparent model aliases and automatic local-first fallbacks.
Module 5: 5-Layer Institutional Memory & Provenance
Layer 0 (Raw SHA-256) → Layer 1 (Normalized) → Layer 2 (Knowledge) → Layer 3 (Canonical State) → Layer 4 (Public-Safe RAG).
Module 1 Checkpoint Quiz
Question: What is the purpose of LiteLLM in the Sovereign Core?
Gary Tutor Preview
The tutor interface is a preview. A live knowledge service is not connected; this control does not send your question to an AI service.
Learning preferences on this device
This pretest saves preferences in your browser on this device. It is not an account, an admission decision, or a cross-device learner record. The tutor preview has no connected AI backend.