Technical leadership is not a promotion for being the best coder — it's a shift from solving problems yourself to building a team that solves them without you. The core concept is 'invisible impact': when you lead well, the team appears to do everything themselves. Key skills include knowing when to stay quiet and let engineers learn from survivable mistakes, distributing context rather than making decisions alone, and managing the human dynamics underneath technical problems. The IC-to-tech-lead transition involves an identity shift away from visible output like commits and PRs toward harder-to-measure contributions like reducing ambiguity, mentoring, and protecting team autonomy. In the AI era, judgment becomes the scarce skill — evaluating whether generated code fits system constraints, owns risk, and handles edge cases. Success is measured by team autonomy: if the team keeps making good decisions when you're absent, the leadership is working.