Loading…
The new bottleneck
Stack OverflowEira May
Summary
AI coding tools have significantly lowered the cost of generating software, yet many engineering organizations fail to realize overall delivery speed improvements. Applying the Theory of Constraints reveals that eliminating code production as a bottleneck shifts inventory directly into surrounding legacy processes that remain unadjusted. New friction points consistently emerge across underspecified requirements, prolonged design handoff gates, senior engineer review capacity, and external sign-offs from legal or security. Organizations can address these blockers by interrogating legacy agile ceremonies checkpoint by checkpoint to determine if their original constraints still exist. Practical remedies include adopting real-time co-development between product and engineering, treating initial designs as fluid starting points, and restructuring processes around running rapid, high-volume experiments.
Context
AI coding tools have reduced the cost and time of generating code, yet engineering teams often fail to see overall delivery velocity improve. Traditional development workflows—including two-week sprints, rigid handoff models, PRD templates, and checkpoint gates—were designed around the assumption that code generation was the primary system constraint.
Approach / What changed
Teams must interrogate workflows against current constraints instead of relying on legacy agile ceremonies. Recommended adjustments include co-developing features in real time across product and engineering, replacing strict design completion requirements with starting points for fast iteration, expanding review capacity, and designing lightweight processes focused on rapid experimentation.
Takeaways
- When code generation is cheap, underspecified product requirements and design handoff gates introduce greater cycle time penalties and rework.
- Doubling developer code output without modifying review, QA, or technical sign-off processes turns senior engineer review bandwidth into the primary constraint.
- Intuit shifted toward real-time co-development between PMs and engineers, turning designs into starting points to scale experimentation volume from few trials to hundreds.
Related reading
From PHP to team lead of agents: rethinking judgment, review, and data with Google's Andi Gutmans (Part 1)
Software development is experiencing an architectural shift as artificial intelligence agents transition individual contributors into orchestrators and reviewers of automated workflows. Rather than spending effort on manual coding, developers increasingly focus on defining system architecture, guiding agent execution, and managing operational risk through tiered human and agent review loops. The primary bottleneck for agentic systems lies in preparing organizational data so that agents can effectively reason over enterprise knowledge. To address this challenge, Google published an open knowledge format and is adopting agent-driven semantic ontology generation to infer relationships across messy data estates. This evolution transforms ontology building from manual human toil into automated generation, allowing engineers to focus on high-value data curation and strategic system design.
Eira May