Making assumptions & cutting down on scope are immensely powerful tools. There is a tendency among engineers to create systems that are universal and work in al

Making assumptions & cutting down on scope are immensely powerful tools. There is a tendency among engineers to create systems that are universal and work in all sorts of environments, when in fact, this is not what is absolutely immediately needed, or is closer to a "nice to have". A widespread example of this is linear regression analysis. Mathematically, it only works because we make assumptions about the data (and even these can be relaxed) & limiting the scope of the problems it can solve. This also applies to delivering projects. A fully engineered & robust project could possibly require a great deal of human, physical resources and time. But if we start iterating over a list of problems your system solves (sorted by decreasing value), the project suddenly costs only a fraction of what it would otherwise. This doesn't come for free though: your MUST understand the limitations & assumptions of your systems, otherwise you or someone will pay dearly. Look at Zillow & failure to understand its forecasting system. Look at so many other startups applying ML where it has no place. Zillow's failure is the best case (though catastrophic): the principal paid the cost. But take something like Google, applying a fully automated system to make decisions about banning accounts. These are real world decisions that have a huge effect on people & businesses. In this case, Google saves money on providing human support & offsets the costs to third parties who often can't afford it.