Article is the latest app to make me revisit how I measure the value of software. Questions to ponder:

1. When does speed to market generate more value than a full feature set?
2. How might a suite of little apps, each doing one thing well, outperform a single app that does many things?
3. What if Article were an experiment, and actual user behavior—not stated preferences—drove functionality?
4. And furthermore, AI personalizes such functionality to each user? So no two copies of apps such as Article are exactly alike?

Questions 3 and 4, especially, are neither desired nor possible using today’s definition of software.

But if we change how we measure the value of software, all these questions may be crucial for shaping the software of tomorrow.