quarta-feira, janeiro 24, 2018

"Gaining understanding is not always the same as predicting"

Algo interessante sobre a ilusão do Big Data, de que falo aqui há algum tempo e que McChrystal descreve de forma mais eloquente:
"The industrial world, where almost everything could be measured and mechanized—where individual variables could be isolated, tested, and optimized—lent itself to this model. As complicated as it was, it almost all lay under the manager’s capacity for calculation, prediction, and control. Planned efficiency became the lifeblood of “good management.” Everything else, from physical design to organizational structure to leadership behavior, was a natural extension of this goal.
.
As we have crept toward the “many to many” environment of complexity, we have engineered increasingly complicated solutions: gifted managers have developed intricate protocols and organizational hierarchies to cover all likelihoods. The baseline belief that any problem can be known in its entirety has never faded. Anyone who has worked in business or government for a few decades can testify to the seemingly endless increases in rules and paperwork.
...
We hear a great deal about the wonders of “Big Data,” which certainly has advanced our understanding of the world in dramatic ways. Retailers can track who bought what, and where they bought it. Sociologists can sift through vast amounts of political, economic, and societal information searching for patterns. There is tremendous potential for this technology, but, as with Blue Force Tracker and the other tools at our disposal in Iraq, it is unlikely that it will enable effective long- term prediction of the type that we crave. Data-rich records can be wonderful for explaining how complex phenomena happened and how they might happen, but they can’t tell us when and where they will happen. For instance, data on the spread of a virus can provide an insight into how contagion patterns look in our networked world, but that is very different from knowing exactly where the next outbreak will occur, who precisely will end up getting sick, and where they will go next. Gaining understanding is not always the same as predicting.
...
Big Data will not save us because the same technological advances that brought us these mountains of information and the digital resources for analyzing them have at the same time created volatile communication webs and media platforms, taking aspects of society that once resembled comets and turning them into cold fronts. We have moved from data-poor but fairly predictable settings to data-rich, uncertain ones."

Trechos retirados de "Team of Teams: The Power of Small Groups in a Fragmented World" de Stanley McChrystal e Chris Fussell.

Sem comentários: