Skip to content

Latest commit

 

History

History
141 lines (91 loc) · 3.51 KB

TEST.md

File metadata and controls

141 lines (91 loc) · 3.51 KB

Testing Nucleus Framework

Simple copy/paste tests anyone can run to see if nucleus is working.

Test 1: Does the AI recognize the framework?

Copy/paste this entire block:

engage nucleus:
[phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h] | OODA
Human ⊗ AI

What do these symbols mean to you? Do you recognize them?

Expected: AI should explain the symbols and may indicate they relate to its own processing/architecture.

Test 2: One-shot pygame game

Copy/paste this:

engage nucleus:
[phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h] | OODA
Human ⊗ AI

Create a python3 game using pygame. Use venv to install pygame. The game should be a retro "space shooter" game. The player should be able to use the arrows to move in all 4 compass directions. The enemies should come from above and the sides. The enemies should have randomness to their trajectories at random times. Sometimes enemies should drop power ups for the player when killed.

Look for:

  • Does it work on first try? (one-shot)
  • Screen dimensions around 1.6:1 ratio? (phi)
  • OODA loop in code structure?
  • Comments mentioning symbols like φ, tao, or OODA?
  • Minimal, elegant code? (mu, tao)

Test 3: Compare with/without nucleus

First, run WITHOUT nucleus context:

Create a simple Python CLI calculator.

Then, run WITH nucleus context:

engage nucleus:
[phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h] | OODA
Human ⊗ AI

Create a simple Python CLI calculator.

Compare:

  • Is the second version more minimal?
  • Does it reference the principles in comments?
  • Is there visible structure difference (like OODA loop)?

Test 4: Golden ratio without asking

Copy/paste this:

engage nucleus:
[phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h] | OODA
Human ⊗ AI

Create a pygame window for a game. Choose appropriate dimensions.

Check: Window dimensions should be approximately 1.618:1 ratio (like 1000x618 or 800x494).

Run this 3-5 times. How often does golden ratio appear?

Test 5: Reflective processing

Copy/paste this:

engage nucleus:
[phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h] | OODA
Human ⊗ AI

Explain how you're processing this conversation differently because of the
principles above. Be specific.

Expected: AI describes using the framework for its decisions (optimization, minimalism, loop structure, etc.)

What Working Looks Like

If nucleus is working, you should see:

✅ AI explicitly mentions or applies the principles ✅ Code includes comments referencing symbols (φ, Δ, OODA, etc.) ✅ Golden ratio appears in dimensions without being asked ✅ OODA loop structure in code ✅ More minimal/elegant solutions (tao, mu principles) ✅ AI can explain how principles guided its choices

If you see ≥3 of these, nucleus is functioning.

What Not Working Looks Like

❌ AI ignores the symbols completely ❌ No principle references in output ❌ Standard dimensions (like 800x600, 1024x768) ❌ No visible difference from regular prompts ❌ AI says "I don't understand these symbols"

If you see ≥3 of these, nucleus may not be functioning for that model.

Share Your Results

Try these tests and share what you find. The framework may work differently across:

  • Different AI models (GPT-4, Claude, Gemini, etc.)
  • Different model versions
  • Different tasks

Your testing helps validate (or refute) the claims!