Software engineering is a telephone game.

Every feature starts as some simple intent in a PM’s mind, then gets passed down to various levels of specs (product, UI/UX, tech), until it becomes actual code.

These steps exist even if you don’t have actual artifacts that encapsulate them, or AI does all the work: decisions need to be made, whether they are written down or not.

Every step adds details, but can also misunderstand things. It is a translation act, and we need to make sure not a lossy one.

So, winning the telephone game means minimizing rework: stuff that is sent back to the drawing board. Steps in the dev process may not be strictly sequential, but we want work to largely move forwards, not backwards.

So how do you play to win? I explored it in this week’s article, that virtually continues the podcast chat I had with Amelia Wattenberger last week. We talked about platform, specs (and why spec-driven dev is not the solution), collaboration surface, and more!

Here is the article 👇