Small misadventures in big data

The project was a complete failure. There was no other way to put it. No euphemisms, no learning from our mistakes, no vague gestures towards it all being an iterative process where it all happens incrementally, one less bad version at a time. After this, there would be no next version. It was done, over, no more

The initial aim was to use big data to copy people. By scraping everything that could be scraped about someone from databases, social media and everywhere else an algorithm could get a foothold, a sufficiently informed replica of an individual could be created that in all ways that mattered were that person, down to whether they had dandruff or not. Everything that could be known went into this doppelganger, and out came a copy so indistinguishable as to be identical

However

The process could not copy the small flaws that made up a person, the minute hesitations, the second-guesses, the imperceptible reroutings of thought processes through childhood memories that characterize a mind that has been around for a while. They got everything right, and thus, they got it all wrong

Nowhere could this be seen more clearly than in the cases when someone chose to have themselves doppelgängered. For whatever reason, they were too busy to be everywhere they needed to be, so they built a copy of themselves to keep up. However, these copies turned out to be better at being themselves than they were, and soon it all became a game of catching up with one’s better half just to stay in the game. Moreover, it happened at times that when someone figured out the person in front of them were the original rather than the copy, their barely disguised looks of disappointment could positively destroy those of a sensitive disposition

And thus it came to pass that no one copied this project. It was, once and for all, over

%d bloggers like this: