Welcome to our reading recommendations of the week!
This week, we suggest you three articles:
- How automation can go wrong and break people lives
- Some tips to get into testing of machine learning and predictive analytics
- Riskiest Assumption Tests (RAT) to replace the MVP
When automation goes wrong
Sometimes, automation can go horribly wrong. Gojko Adjic tells stories of people losing their job or a lot of money because of a system called Michigan Data Automated System (MiDAS) in Michigan. Around the world, computers are replacing human interaction with robots, these stories remind us we should be very wary about this, only automate “difficult and time-consuming parts of work and assist people in making better decisions”. The following article is an excerpt from his upcoming book Computer Says No you can already pre-order: When automation goes horribly wrong by Gojko Adjic
Testing machine learning systems
How to test machine learning systems and predictive analytics? In these days where Data Science is eating the world of software, we, testers, must find new ways to test nondeterministic applications. In this article written by Peter Varhol, you will learn the differences between machine learning systems and predictive analytics, then a few guidelines and best practices in order to prepare yourself to this new era of software testing: Moving targets: Testing software in the age of machine learning
Replace awful MVP with RAT
Have you ever heard about the RAT? Riskiest Assumption Tests is what you should focus on instead of a MVP, be it the Minimum Viable Product or the Minimum Valuable Product (or the Minimum Lovable Product). Read this interesting article by Rik Higham: The MVP is dead. Long live the RAT.