Artie Shevchenko, speaking at the Craft conference, proposed a system for code reviews that combines AI checks with targeted manual reviews. This approach involves using AI for mandatory checks on code changes and implementing 'manual spikes' for more complex alterations. The goal is to maintain developer proficiency while potentially increasing review velocity.
Shevchenko suggested that teams could accelerate development by allowing AI to approve low-risk pull requests, particularly when most developers are code owners and teams are small. However, he emphasized that AI's token generation creates an 'illusion of understanding,' making human intervention necessary to prevent a 'broken knowledge feedback loop'.
Cognitive debt, defined as the erosion of shared understanding within a team about their system, is a concern amplified by AI, according to Shevchenko. He stated that AI can increase review volume, exacerbating this problem. To counter the code review bottleneck, he advocated for strategies beyond just developer experience improvements.
Shevchenko stressed the importance of maintaining intellectual control over codebases, especially with the increased complexity introduced by AI. He proposed an approach where developers first code a change, then use AI to suggest improvements and fixes. This method allows developers to review individual AI-proposed changes from a position of deep understanding of the initial solution.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's briefNew every morning, and the back catalogue is archived by date.
Artie Shevchenko presented at Craft conference on integrating AI into code reviews, suggesting mandatory AI checks with manual spikes for complex changes. He argued that while AI can increase velocity by handling low-risk reviews, human oversight is crucial to prevent cognitive debt and maintain intellectual control over codebases.