7/ For example, in this tweetstream, item 4 is missing (incompletion), and there are 2 item 5s (overdetermination) *pats self for nice save*
Conversation
Replying to
8/ Those who are strong at being data-driven are actually hit hardest, since they find many *more* clever ways to overfit data.
2
1
2
Replying to
9/ Overfitting overdetermined reality = compounded noise = IO = Sense of mounting chaos. Anxiety provoking to all, paralyzing to data-driven
1
1
1
Replying to
10/ The obvious alternative is narrative filtering (see recent Uber valution kerfuffle for example). This is vulnerable to confirmation bias
1
Replying to
11/ You mitigate confirmation bias through analytical rigor. When this is hard you get analysis paralysis.
1
1
2
Replying to
12/ Us narrative/analytic types have long history of dealing with confirmation bias and analysis paralysis to learn from.
1
Replying to
13/ Data-driven types otoh have only about 20 years of history dealing with overdetermination/overfitting problem and data paralysis.
1
Replying to
14/ So... IO can be defined as being caught between rock and hard place: analysis paralysis and data paralysis.
1
Replying to
15/ When you're not paralyzed, you're in an stressful arms race between outbound narrative filtering and inbound data filtering.
1
2
2
Replying to
16/ So IO subjectively feels like manic depression: caught between paralysis and snowballing crazymaking, orbitting around 'sensemaking'
2
3
Replying to
My favorite fictional novel, as in, the novel doesn’t exist, is Marking/Making books.google.com/books?id=fu5cA (click first page link)
Replying to
a psychoanalytic answer - developing negative capability en.wikipedia.org/wiki/Negative_ and the capacity to bear frustration
1
1


