## 14:27 building first, building outright

it's often hard to realize or rationalize, but it's impossile to build
the perfect tool outright. the perfect tool is only built over many
iterations and trials; it's important to continue to iterate, rather

trends are cast over shorter time periods than we account for. when
working with data, it's often risky to extrapolate with just two or four
data points; with human behavior, these trends are often both pointed
out and lost.