Operating reality
What actually exists, who maintains it and what observations support the representation.
What depends on what?
Try a simple example.
If the power goes out, water pumps, food storage and communications may be affected too. A simulation is a model for exploring connections like these before making changes in the real world.
Look at the connections
A food business needs water, power, transport, refrigeration and skilled people. Changing one part can affect the others. A useful model makes these links visible, so people can explore a change and ask what they might have missed.
ObserveRecord conditions, origin and credits and what remains unknown.ModelExpose relationships and alternative assumptions.RehearseExplore normal operation, changes and disruptions.ActAuthorised people and systems carry out suitable work.ReviewCompare the outcome, correct and share agreed learning.
Sources for this section
D09 · The B.E.L.A.U. Protopian Gambit P10 · The Long Game: Grain by Grain
Know the limits of a model
A model can help people ask better questions without predicting everything. It needs clear assumptions, information about the real system and checks against actual outcomes. A detailed picture of a business or town is not enough to show that its predictions are reliable.
Alternative futures
Changes to staffing, ownership, infrastructure, rules and resource allocation.
Contingencies
Loss of suppliers, equipment, networks, finance, skills or access to a particular AI service.
Sources for this section
D03 · A Fair Go for the AI Age D09 · The B.E.L.A.U. Protopian Gambit
Try a small example
The example below starts with an electricity interruption because a documented South Australian outage shows why power is a useful place to begin. The service links in this small model are teaching assumptions, not a reconstruction of that event. No live sensors, backup capacity, timing or probability are modelled.
Learn from what really happens
Businesses taking part could test a specific improvement and compare the result with the model. They would share only information they were authorised to share. Recording failures and unexpected effects would be as useful as recording improvements.
Sources for this section
P02 · Aura of Intelligence P05 · Native Nations of the World D05 · Fiji-Australia Vuvale Union Submission
Start with power: South Australia, 2016
AEMO’s final report records about 850,000 customer connections losing electricity in South Australia on 28 September 2016. The report describes effects on households, businesses, transport, community services and industry.
Why the tool starts this way
That documented event is why electricity is selected first. The nine services and their links below are a simplified teaching model chosen for this site. They are not outage measurements, probabilities or a claim that every linked service failed in 2016. Select a different interruption to compare paths. The model does not include backup systems or time to recovery.
AEMO: final report on the 2016 South Australian black system event
A simple example
Service connections example
Select an interruption and trace a small service network. A path in this graph is not a prediction of real-world failure.
Which service is unavailable in this sketch?
Gold: selected interruptionViolet: dependency pathTeal: no path traced