Thought experiment - if we had airtraffic controller,, and pilot

Dev/OPS - Flight controller recording;

Act 1
====
Resilience in Incident Response. Hueristics / what changed / first thing is that engineers look at , second is to go wide and look at the universe of events and do a breadth wide search, thrid is convergent searching -  confirm and disqualify diagnoses by matching signals/symptoms, specific and past vs a general and recent - so like painful memories or things that are still in our cache;

So how do we do it better. Elements of expertise; recommend if it looks right, and able to make fine discriminations, mental simulation - they can imagine how to solve it or where it's going; higher order rules are knowing when to break the rules; 
Paper: Seeing the invisble: Perceptal-COgnitive Aspects of Expertise - "Klein & HOffman"

Experts see what is not there; 

Experience into expertise - personal, directed, manufactured, vicarious - on-call, teaching or code review, chaose engineering, and then I remember one time, 

Rasmussen Triangle - Money \ Acceptable Risk / and the lazy one i think __

The discrescionary space is where the money and acceptable risk meeets up before things blow up

Post moretems are also called retrospectives 

#Antipatterns

- Only reviewing failures
- Forgetting Biases, and taking them into account 
  - Counterfactuals - why didn't you do xyz, and that's hindsight biases
  - IKEA effect, we value things we build more than things we don't 
- De-prioritizing Retrospectives or Learning Processes
  - Time since incident is very important

Preventino / Act3
Paper title - Ironies of Automation Lisanne Bainbridge 
the more important an automation is the more important the human becomes
manual skills deteriorate 
speed vesus correctness
automation can camouflage the current state

joint cognitive system - foundations of coordinations;
connectivity versus coordination; the Animacy Paradox - as they increase in autonomy and authority we perceivethem in two ways, it is a deterministic machine | or animate agent, capable of activities independent of the operator; We think its deterministic before and after an incident, but during the incident we think it's behavig independently; 

Ok so how do we help to deal with it : Use simulations or chaos engineering, and then widen understandiing.

Expertise takes time and space to grow it otu

Take away: Success and failure are the two sides of the same coin, andd you have 72 hours to learn after an incident. 




##note# audio example, i didn't get it though