What if you were to feed a long short-term memory (LSTM) recurrent neural network (essentially, a deep-learning artificial intelligence engine) a bunch of science fiction screenplays and set it to work writing an original screenplay.
Then what if you gathered a cast and crew for one day and attempted to film the result.
I’ve been thinking a lot recently about the size of user stories in agile projects. The idea that I’ve been reflecting on is what if teams only worked with small, similarly-sized pieces of work, rather than exponentially larger blocks of work?
In theory, small user stories should be more predictable, should include less risk, less uncertainty and less complexity. They should, therefore, take less time to complete than larger user stories… you would think! Or as Mike Cohn put it in Agile Estimating and Planning (Prentice Hall, 2006), “small stories keep work flowing”.