In 2015, the bottleneck was writing code.
Teams needed coordination because many people worked in parallel. Estimation mattered because coding time was the constraint. Documentation came after because shipping was the priority.
The processes that emerged optimized for that bottleneck: - Sprint planning to allocate coding capacity - Story points to estimate coding effort - Standups to synchronize parallel work - Post-ship documentation
This made sense. For 2015.
The Bottleneck Moved
AI shifted the constraint. Writing code is cheap now. Understanding code is expensive.
New bottleneck: - Can we comprehend what we're building? - Can we preserve context across sessions? - Can we maintain what we've built? - Can new people ramp up?
Old process optimized for old constraint. Mismatch.
What Needs to Change
From Pre-Coding to Pre-Understanding
Old: Invest in planning before coding. New: Invest in intent clarity before generating.
From Post-Ship Docs to Real-Time Capture
Old: Document after shipping. New: Capture decisions as they happen. Context evaporates too fast for "after."
From Coordination Overhead to Minimal Sync
Old: Daily standups, sprint planning, grooming. New: Async by default. Sync when needed. Small teams need less ceremony.
From Velocity Metrics to Understanding Metrics
Old: Track story points completed. New: Track whether the team understands what they're building.
From Large Teams to Lean Teams
Old: Scale headcount to increase velocity. New: Small teams with AI. Coordination overhead exceeds value at scale.
The Transition Challenge
You can't flip a switch. Organizations have inertia: - Tools built for old process - People trained in old methods - Stakeholders expecting old metrics - Culture shaped by old assumptions
But you can evolve deliberately: - Question each process element - Experiment with lighter alternatives - Measure what matters now, not what mattered before - Let go of what no longer serves
The Update
Your process was probably designed for a different era. Maybe 2015. Maybe 2010.
The constraints changed. The practices should too.
Look at your process. Ask for each element: - What constraint was this designed for? - Does that constraint still exist? - What constraint exists now that this doesn't address?
Update accordingly. Your process should serve current reality, not historical assumptions.