Bias in machine learning: We can do better | Dark Hacker News