Nobody Knows How Large Software Projects Work

Large software systems are so complex that even their creators often cannot answer basic questions about how they function. This isn't a failure of competence but an inherent characteristic of scaled systems.

Companies capture significant value through "fiddly, awkward features" -- self-hosting options, organizational controls, localization, regulatory compliance. Each of these "affect every single other feature you build," creating exponential complexity impossible to fully comprehend.

Static documentation fails because "the system is rapidly changing as you try to document it." Many system behaviors "don't necessarily have a lot of conscious intent behind them" -- they emerge from accumulated design choices.

"Simple questions like 'can users of type Y access feature X?' often can only be answered by a handful of people in the organization."

"The ability to answer questions about software is one of the core functions of an engineering team."

"It's easier to write software than to explain it."

Answering questions about large systems requires investigative work through code reading, experimentation, and trial-and-error testing. This investigative capability is more valuable than commonly recognized and explains why engineers hold significant institutional power.