VERTEXPACE
HomeLuminaLightNetAI Map
UNIVERSAL CODE MAP · GOFAI → DEEP LEARNING → LIGHT CODE · 75 CONCEPTS · 4 PLANES · 64 BYTES · ONE ADAPTIVE LOOP · IF/THEN + WEIGHTS + FEEDBACK · UNIVERSAL CODE MAP · GOFAI → DEEP LEARNING → LIGHT CODE · 75 CONCEPTS · 4 PLANES · 64 BYTES · ONE ADAPTIVE LOOP · IF/THEN + WEIGHTS + FEEDBACK · UNIVERSAL CODE MAP · GOFAI → DEEP LEARNING → LIGHT CODE · 75 CONCEPTS · 4 PLANES · 64 BYTES · ONE ADAPTIVE LOOP · IF/THEN + WEIGHTS + FEEDBACK ·
Vertexpace · Light Code · NovaGlyph

Universal Code Map

How AI evolved from rigid if/then rules to adaptive learning — and how Light Code bridges the gap.

01 — AI History
1950s
Symbolic AI
Hand-coded if/then rules. Symbols manipulated by logic.
Logic TheoristELIZAShakey
1970s–80s
Expert Systems
Thousands of domain rules from human experts. Brittle to new cases.
DENDRALMYCINProlog
1990s–2000s
Machine Learning
Learn patterns from data. Rules discovered, not written.
SVMsRandom ForestsNaive Bayes
2010s
Deep Learning
Neural nets with billions of parameters. Opaque but powerful.
AlexNetGPTTransformers
2026
Light Code Planes
75-concept universal layer. Adaptive on 64 bytes. On-device.
PREDICTDECIDELEARN
02 — The Core Problem & The Fix
● Why GOFAI Failed
🔒
Static rules
Every behaviour hand-written. No pathway to update itself.
🧱
Brittle edges
Unknown word = silence. ELIZA can't handle anything new.
📈
Rules only grow
MYCIN had 600 rules. More complexity = more rules. Never shrinks.
🚫
No feedback loop
ELIZA never knows which responses worked. Cannot improve.
● How Light Code Fixes It
🔄
64 bytes, reused
POLYGRID: planes reuse the same 64 bytes. Meaning shifts, not size.
🧠
Semantic primitives
75 concepts are universal meanings. Concept distance is computable.
⚖️
Weighted tendencies
Rules become weights on concepts. LEARN adjusts from outcomes.
🔁
Closed loop
PREDICT → DECIDE → CREATE → LEARN → repeat. Every cycle sharpens.
03 — Light Code IS GOFAI (With One Extra Step)

Both systems match a pattern to produce a response. Light Code didn't invent new thinking. It added one line at the end that GOFAI never had.

ELIZA — step ③ missing
① MATCH
if keyword == "mother"

② RESPOND
  reply = "Tell me about your mother"

③ UPDATE — does not exist
  (nothing happens — ever)
Light Code — ③ now exists
① MATCH
if concept == memory AND weight > threshold

② RESPOND
  take action

③ UPDATE ← the only new step
  if outcome == good: weight_memory += delta
  else:           weight_memory -= delta
The Honest Summary
Light Code is GOFAI.

GOFAI's problem was never the if/then. The problem was that each rule had no confidence score and no mechanism to update. Light Code keeps the if/then — cheap, explainable, runs on an ESP32 — and adds the one missing piece: a weight per concept that LEARN adjusts from outcomes. No neural net. No matrix multiplication. GOFAI + weights + one update step per cycle.

04 — The Four Adaptive Planes
🔮
PREDICT
What will happen?
Generates expectation from concept weights. Outputs confidence score. Powers all decisions.
⚖️
DECIDE
What should I do?
7-step loop on 64 bytes. Chooses action using PREDICT's confidence map.
CREATE
What do I make?
Generates response. Recursive slot machines. Personalised from memory + PREDICT + DECIDE.
🔁
LEARN
Did it work? Adjust.
Reads all three plane logs. Computes delta. Updates concept weights in agent memory.
↑ LEARN reads CREATE output → computes delta → updates PREDICT weights → next cycle is smarter ↑
05 — Side-by-Side Comparison
PropertyGOFAIDeep LearningLight Code
RulesHand-writtenLearned (opaque)Weighted concepts (transparent)
AdaptationNoneRequires retrainingEach cycle (online)
Data neededExpert knowledgeMillions of examplesSingle outcome
ExplainableFull ✓Black box ✗Full ✓
ComputeTiny ✓GPU cluster ✗On-device (ESP32/Pi) ✓
UnknownsSilence ✗GeneralisesNearest concept distance
Memory growsYes — forever ✗Yes — forever ✗No — 64 bytes fixed ✓
06 — Ask the AI Guide
🧠
BABY AI — Universal Code Guide
Powered by Claude · Knows GOFAI, Deep Learning, and Light Code
🧠Ask me anything about old AI, deep learning, or Light Code.
Vertexpace · Light Code v0.3 · NovaGlyph POLYGRID · 2026
75 concepts · 4 planes · 64 bytes · one adaptive loop