@leovido.eth

Bought this book for my wife when she first started learning Scrum. I ended up learning a lot from her, and have mentored and upskilled my team to use Scrum properly a few years ago. Lots of companies simply rely on "Scrumbut", ignoring the essential parts that allow companies to ship and improve their products. After reading the first few chapters, there were a couple of things that fired up some neurons in my brain, around how humans work and AI. The book starts off with the FBI, and how hundreds of millions of dollars have been spent to build a reliable software system. The problem here was the Waterfall methodology. Everything is planned beforehand, taking a lot of time, requirements, etc. This way, nothing could go wrong, right? Well, history shows us that this has failed consistently, multiple times. It's not that people weren't smart, it's about the processes involved. The key element about Scrum, is that it focuses on feedback from the customer, product owners, stakeholders, to make sure they are all aligned with building the right product. Often times, it happens that a project might be completed, but in the end, the user won't use it because of lack of initial feedback. This is why we have beta versions and groups that provide feedback to improve the quality of the product. Relating it to AI, we have an amazing tool at our hands, but without the right processes in place, we can end up with an app that will likely have an expiry date. Vibecoding can be enhanced if we use AI skills, rules, project charters, business information, etc., that will produce and build an app that aligns better with our vision. If Scrum works beyond software development, what if we bring its benefits to AI? What if someone created a system that follows Scrum values? Agents could be used as QA testers, or even act as product owners. If Scrum already proved to delivery of software in half the time, could we see an even quicker delivery if implemented correctly with AI?

4 replies

11 recasts

151 reactions