Shifting focus from delivering more code to creating more value requires aligning architecture with user needs and business outcomes.
The Ohio State University sidewalk example shows that real value emerges when systems adapt to how people actually use them, not based on planned assumptions.
Good products solve real problems within the user's context, while bad products simply add functionality without relevance, creating complexity without value.
The "build trap" measures success by output like features shipped instead of outcomes, a problem that AI can accelerate by helping build the wrong things faster.
Design systems around the impact they should enable by starting with observable outcomes, which changes the questions architecture needs to answer.
Architecture should enable learning by making impact observable, supporting experiments with safe rollbacks, and optimizing for the reversibility of uncertain decisions.
Connect technical decisions to business assumptions, treat data as a feedback infrastructure, and make creating impact a shared responsibility across the entire team.
The ultimate goal is to move from shipping features and optimizing output to enabling outcomes, optimizing learning, and measuring real product impact.
Jobs that call for the skills explored in this talk.