Database Reference
In-Depth Information
Tunkelang:
One problem? There are so many! Data should help us make
better decisions. As we've learned from the work of Herbert Simon, Daniel
Kahneman, and many others, human beings are terrible at making rational
decisions. Data scientists should work to improve decision making wherever
data has to compete with human irrationality.
Gutierrez:
What is something a smallish number of people know about that
you think will be huge in the future?
Tunkelang:
By now, more than a smallish number of people should know
about what Dan Ariely calls our “predictable irrationality.” Books by him and
Daniel Kahneman have been best sellers. And yet I see little evidence that
people are applying the insights they should be deriving from those topics.
We make most of our decisions without recognizing the cognitive biases that
taint our decision-making process. Indeed, by doing so we are exercising a
form of overconfidence bias.
As software plays an increasing role in our day-to-day lives, I'm hopeful that
it will intercede in some of that decision making. Our computers, mobile
devices, and web-based services are witnesses to many of our daily decisions.
I look forward to the day that those devices play a more active role in helping
us make better decisions.
Gutierrez:
What has been the biggest thing you have changed your mind
about and how did that change come about?
Tunkelang:
When I was a student, I idealized theoretical work. My aspiration
was to be a professor contributing to theoretical mathematics and computer
science. Perhaps part of my reason was that the problems were so difficult,
and I equated difficulty with value.
As I've grown up, my values have been informed by experience. I still have a deep
respect for the intellectual acuity of theoreticians, but I'm much more impressed
by people who deliver practical impact. In fact, people who find simple solutions
to important problems especially impress me. Sometimes it's necessary to work
hard, but what matters are the results, not the effort expended.
I've tried to live according to those values myself. My goal is to produce the
most valuable results for the least amount of effort. If I have to solve hard
problems or make theoretical contributions along the way, then the end
justifies the means.
Gutierrez:
What personal philosophies and/or theories have you developed
from working with data?
Tunkelang:
Not sure it's a personal philosophy, but I assume that anything
that looks interesting is probably wrong. Even though I assume this, I look at
it anyway, because sometimes it really is interesting, and in those cases, it can
be extremely interesting.
