Skip to content

Roadmap

Project milestones and future plans. Updated each session.

Completed

✅ Project Setup

VS Code, venv, GitHub. Professional development environment. github.com/GravyLord/nci

✅ Concept Disclosure

NCI_Concept_Disclosure.pdf created under Lesley Ancion / Goatface Tech. Timestamped authorship established.

✅ Session 1 — Resonance Nodes & Compression Index

Built core data structure. Build: 0.87s, Search: 3.5ms, Efficiency: 99.1%.

✅ Session 2 — Concept Clusters & Context Boost

ConceptCluster via k-means. Sparsity fixed: 0.48% → 2.43%. Context boost added.

✅ Session 3 — Knowledge Landscape & .nci Format

KnowledgeLandscape, consolidation, inter-cluster connections, binary save/load. Round-trip verified identical.

✅ Session 4 — Visual Brain

brain_visual.html — D3.js force-directed graph visualizer. Cluster nodes, glow effects, hover tooltips, zoom and pan.

✅ Session 5 — Semantic Signatures

embeddings.py using all-MiniLM-L6-v2. learn(), learn_many(), query_concept(), concept_label(). Brain understands meaning, not just structure.

✅ Session 6 — MkDocs & Visualizer Update

MkDocs documentation site. Concept labels in visualizer tooltip.

In Progress

🔵 Native NCI Model Training

Train a small model entirely within NCI. No PyTorch. No GPU. Pure NCI from scratch.

Planned

⚪ goatface.ca Website

Public showcase for NCI and Goatface Tech. Single polished page — brand, NCI concept, core law, live visual brain demo, benchmarks, GitHub link. Hosted on Cloudflare Pages or GitHub Pages (free).

⚪ Easy Training UI

Upload a PDF or URL, type what to learn, NCI ingests autonomously. No code required to teach the brain.

⚪ Autonomous Paper Trading Bot

Test environment starting with a set amount of theoretical money. Signal engine → paper portfolio manager → dashboard view. Runs on Orange Pi alongside existing screener.

Future Vision

Living Brain Visualizer

Vision

Like the time threads in Loki — flowing, luminous, alive. Every thought leaves a trail of light.

  • Bezier curve threads instead of straight lines
  • Animated light pulses traveling toward firing nodes
  • Warm amber burst when a node activates
  • New node bloom when the brain learns something new
  • New connection formation — a thread fades in for the first time
  • WebSocket real-time mode — live events from a running NCI instance

Autonomous Learning

Plasticity, Hebbian strengthening, consolidation, curiosity drive. The brain gets smarter on its own, continuously, on modest hardware.

RPi Deployment

Full NCI running on a Raspberry Pi. 2–5 tokens per second. No GPU. No data center. Accessible to anyone.

NCI Language

A programming language where sparsity, compression, and hardware-awareness are native concepts — not optimizations.