What happens when you build a machine learning algorithm—but do it fundamentally wrong? The graphs below demonstrate the difference between planning and attempting to hack something together on-the-fly. This was try one — My attempt to hack it together and make it work failed utterly. This, hack-and-hope approach, works for simple software or aged technologies that you have lots of experience in. If you try it in a machine learning problem — you get [very] wrong predictions or in a more real-world case dead people and crashed cars.
