ClueCon Weekly with Pavel Oborin [Sn. 15 Ep. 18]: Why Code Reviews Become the Bottleneck

ClueCon Weekly with Pavel Oborin [Sn. 15 Ep. 18]: Why Code Reviews Become the Bottleneck

Code reviews are supposed to protect quality, but in a lot of teams quietly become the slowest part of delivery. In this episode, host Jon Gray is joined by Pavel Oborin, Lead Engineer at Revity, to break down why pull requests stall, what open-source communities do differently, and how teams can cut idle time without cutting corners.
​
We dig into the real-world causes of review bottlenecks (busy reviewers, missed notifications, time zones, multi-round iterations, and competing priorities), then walk through practical “low-effort / high-impact” improvements: linting and CI automation, merge queues, code owners, and moving big architectural debates earlier into a design phase using RFCs/ADRs. The conversation also gets into the newer reality: AI can generate a lot of code quickly—so the job shifts toward reviewing intent, architecture, and risk, not just syntax. Pavel shares how to treat AI like a super-productive teammate who lacks your domain context, and how specs and tests can keep the output grounded.
​
Topics covered:
🔹Why PRs sit idle (and what causes multi-day feedback loops)
🔹Automation that removes “human-delay” steps
🔹Reducing nitpicks so humans focus on architecture and domain logic
🔹AI-assisted review + spec-driven development