Monday, October 4, 2010

Soft is Good

"This was a time to robotically identify sources of diversionary patterns. If successful, the countermeasure could be to simply  lower the weighting of repetitive themes to account for the manipulation. Realistically, one would have to accept that statistical methods required and probability would limit precision. Still, there was good reason to be confident in the methodology.

"Linking ostensibly descriptive labels to discover their value, a good place to start.  For example, "Economic Freedom" and "Tax Evasion" or "Market Based Management" and "Unfair Competition". As we have seen more and more lately, relying on human input is necessary to move at optimal speed even though automation is considered very early. It is all inexpensive because the resources can be re-purposed with little effort or time."

How often they would hear from otherwise intelligent people: I don't get involved in that stuff, I am not bright enough. And it wasn't false modesty. These people were behaving just like one should expect. They are human. We have the luxury of doing this as our 'day job'. It is part of our job description to beware of a high-order abstraction engine's vulnerability.

No comments:

Post a Comment