9/ Exhibit A is of course the TSA security theater. I am sure we're safer since 9/11, but almost all safety increase is in invisible places.
Conversation
Replying to
10/ Not only does "lean" translate to "increase risk", incentives are stacked to increase risk for people who have least ability to complain
2
5
11
Replying to
11/ In general, if there is a need to cut costs by X, 10% will come from function degradation, 90% from increasing risk via "leaning"
1
3
7
Replying to
12/ Of that 90%, almost all will happen in parts of ops that will affect the vulnerable more. Not today, but at some unknown future date
1
2
4
Replying to
13/ Complex infrastructure inherently involves moral hazard because of this: experts manage risk distributions via cost cuts over time
1
2
8
Replying to
14/ In this, impacted parties are not blameless. They usually knowingly overallocate resources to function X over insurance of function X
1
2
Replying to
15/ How do you tell real efficiency increases from fragility-for-cost tradeoffs? Follow the information. Real thing embodies new knowledge
1
3
4
Replying to
16/ When you can't find any new information embodied by an evolved system, but it's cheaper, then it is more fragile
2
1
5
Replying to
17/ If you find a new information embodied locally, then there is some real efficiency increase and some risk moved elsewhere
2
1
4
Replying to
And I agree that there is a risk/efficiency tradeoff, but most real systems aren't anywhere near the efficient frontier for it.
2
Replying to
Yeah, it's expensive to creep towards it. Easier to fake it even when you're not at the frontier.
Replying to
3
Replying to
Well, they are still susceptible to enormous frame pressures at the boundary

