Texas sharpshooter

I really like this formulation of a common apophenia-caused fallacy: a Texan guy who shoots several times at a barn wall and then paints a target centered on the tightest cluster of hits, calling himself a sharpshooter.

Texas sharpshooter is a metaphor for the process everyone uses to some extent in their daily lives. It maybe useful in some cases (to generalize), but when applied to complex systems with a lot of contradicting data, it usually leads to errors.

The examples are numerous: climate change deniers choose specific regions/time-intervals to illustrate their point (although that is becoming increasingly difficult), parapsychologists find clusters and patterns in large amount of random data, betting companies cultivate the myth that lotteries can be predicted.

Self-driving trolley

It was all fun and games pondering with ridiculous unrealistic trolley problem until we started to think about the self-driving cars.

Now the trolley dilemma forcing us to kill either 1 or 3 people is not so unrealistic and far-fetched anymore. Imagine a car driving itself along the road and sensing three people crossing the road – it quickly calculates that accident is unavoidable: the car has to decide in split-second – either to drive forward and likely kill three pedestrians or swerve and likely kill the driver.

We can debate about the insufficiency of the information available, but ultimately the car has to decide. Will it go for utility? Does it have to consider number of people inside the car then?

relevant read