"A complex system that works is invariably found to have evolved from a simple system that worked.
Gall is not praising simplicity. He is describing a pattern: complex systems that work usually earned their complexity. They grew from something smaller that already behaved, and the early version taught people what mattered.
When teams start with the final architecture in mind, they skip the hard part: discovering which constraints are real, which failure modes matter, and which interfaces deserve permanence. That is how rewrites become brittle. You end up locked into decisions you have not tested. Complexity added upfront is often guesswork about constraints you haven’t met yet.
The non-goal is the perfect system. Start with a simple system that works and let reality add constraints. Name what you are not building yet, so the first version can actually ship and the boundaries stay defensible.