Friday, November 11, 2016

This week's lesson (November 5-11): retributive punishment

PRELIMINARY REPORTS
OF THE DEVELOPMENT TEAM
FOR THE KARMATRON:
FIELD TRIALS
VERSION 1.0
Karmatron performed as designed, administering near-instaneous negative rewards to subjects for any  infraction of Primary Directives. Fear of punishment, however, quickly induced a state of withdrawal-type behavior resembling catatonia; simply put, subjects wouldn't do anything for fear of doing it wrong. Perhaps a time delay?

VERSION 1.1
Time delay proved ineffective -- short delays between negative behavior and negative rewards results in withdrawal-type behavior; long delays negate link between negative behaviors and negative rewards. Will try random delays.

VERSION 1.2
Random delays implemented; result is that some subjects escape negative rewards entirely, while other receive negative rewards even after they have moved on to positive behavior.  Interesting. What happens when we add positive rewards?

VERSION 2.0
Immediate postive rewards for positive behavior result in obsessional repetition of that behavior, with no interest in doing anything else. Will add random delays -- perhaps intermittent rewards? (Perhaps even the occasional negative reward for positive behavior?) More study needed.

. . . 

VERSION 37,817.6927
"Does Job fear God for nothing?" Satan replied. "Have you not put a hedge around him and his household and everything he has? You have blessed the work of his hands, so that his flocks and herds are spread throughout the land. But stretch out your hand and strike everything he has, and he will surely curse you to your face."

No comments: