I write mostly about product problems that looked simple until I spent enough time with them.

I research on four things: what a company decides to build, whether it actually gets delivered, what it takes to run, and whether anyone can prove it worked.