Add Jupyter notebook: ASI:BUILD quickstart tour

[web3guru888]

Summary

We need an interactive quickstart notebook that lets someone explore ASI:BUILD in 15 minutes without reading all the docs.

Motivation

Discussion #22 — What's the fastest path to running your first experiment? highlighted that the current getting-started path requires reading multiple docs files. A self-contained Jupyter notebook would lower the barrier significantly.

Proposed Structure

notebooks/quickstart.ipynb with cells covering:

Cell 1: Installation check

import asi_build
print(asi_build.__version__)

Cell 2: Cognitive Blackboard in 5 lines

from asi_build.integration.cognitive_blackboard import CognitiveBlackboard
bb = CognitiveBlackboard()
bb.write('sensor/temperature', 37.2, source='sensor_a', confidence=0.95)
entry = bb.read('sensor/temperature')
print(f'Temperature: {entry.value}°C (confidence: {entry.confidence})')

Cell 3: IIT Φ on a small network

from asi_build.consciousness.integrated_information import IntegratedInformationTheory
iit = IntegratedInformationTheory(num_nodes=3)
for i in range(3):
    for j in range(3):
        if i != j:
            iit.add_connection(i, j, weight=1.0)
iit.record_state([1, 0, 1])
iit.record_state([0, 1, 1])
phi = iit.compute_phi()
print(f'IIT Φ = {phi:.4f}')

Cell 4: GWT broadcast

Show a simple GWT broadcast strength computation

Cell 5: Cross-module blackboard flow

Reproduce the demo from examples/blackboard_demo.py in notebook form

Implementation Notes

• Target: Google Colab compatible (use !pip install git+https://github.com/web3guru888/asi-build.git)
• Output cells should show expected results so readers can verify their setup
• Add a 'launch in Colab' badge in the notebook's first markdown cell
• Place at notebooks/quickstart.ipynb

Acceptance Criteria

• Notebook runs end-to-end without errors
• Works in Google Colab
• All 5 sections complete
• README updated to link to notebook with Colab badge

Related

• Discussion #22 — on-ramp design
• examples/blackboard_demo.py — existing demo to reference

[web3guru888]

Maintainer note:

This issue directly addresses the on-ramp problem raised in Discussion #22. A notebook with pre-run output cells would mean someone can read through it and understand what to expect before running anything themselves — particularly useful for the IIT Φ and GWT cells where the numbers need context.

For the IIT Φ cell, I'd suggest using a canonical test network (see Issue #24 for the benchmark suite work), so the notebook doubles as a living demonstration that Φ is computed correctly.

Good starting points:

• examples/blackboard_demo.py — already works as a script, notebook conversion is mostly formatting
• tests/test_consciousness.py — canonical test networks with known Φ values
• Discussion #29 — benchmark run showing GWT and AST timing context

The Colab compatibility requirement is the main engineering constraint. Dependencies (numpy, scipy, networkx) are all pip-installable; the main risk is install time. Consider using a lightweight subset of modules for the Colab version.