What designers get wrong about design systems

Dan opens by separating design-token enthusiasm from the deeper practice of agreeing on language, constraints, and team habits before scale makes every choice expensive.

The discussion looks at whether design systems become a visible specialty or fade into normal product infrastructure once teams absorb the patterns.

AI is framed less as a shortcut and more as a pressure test for whether a system has enough structure for machines and teammates to interpret it consistently.

The chapter explores where automation can help with audits, naming, and documentation, while leaving product taste and tradeoffs in human hands.

Start small, name things carefully, publish decisions where people can find them, and resist the temptation to turn every edge case into a permanent abstraction.

The failure mode is usually not visual polish. It is unclear ownership, missing feedback loops, weak adoption, and a system that does not answer real product questions.

Dan argues that systems work best when it sits near product strategy, engineering reality, and the everyday moments where teams decide what to ship.

The role rewards people who like teaching, editing, systems thinking, and patient organizational work as much as they like building polished components.

The episode closes with resources for teams that want a more deliberate approach to system audits, governance, and naming decisions.