Using Machine Learning to Pick Tests

The way that CI typically works today is that you maybe have different phases of tests. And, but we've kind of created this stacking of tests when it seems like maybe a better way is for the system to learn what tests are likely to be needed to be run instead of just taking this like blunt hammer to it every time. This is one of the more exciting areas starting to actually use real machine learning for real problem solving as opposed to just having it in the loop.