How AI Code Review Changes Workflows: The Loss of 'Why' in PR Evaluations
Sea-Relative2951 · reddit · 2026-08-10
The author shares how their workflow and mindset have profoundly changed after using AI (like Verdent's review subagent) to assist with code reviews.
- Loss of Context: Previously, the author would ask human developers 'why' certain decisions were made. Now, facing an AI-generated transcript, the human-to-human debate and context are lost.
- Changed Reading Habits: Abandoned top-to-bottom reading. Instead, they rely on the AI to flag important parts of the diff and start reviewing from there.
- Blind Spots: Admits to no longer forming a complete picture of the middle sections of large diffs, making it easier to miss logical errors (like redundant retry wrappers) in code that looks familiar.
More from coding & agent
- Vercel CEO: AI Coding Still Needs Human Review as Models Make Rookie Mistakes — brandon_galang · 2026-08-10
- Developer Reaffirms: MCP and RAG Coexist as Many MCP Servers Are Essentially RAG — cocktailpeanut · 2026-08-10
- Dev jokes: 'It's not vibe coding if you actually care about the code' — haydendevs · 2026-08-10
- Scale AI Founder: Misaligned Multi-Agent Swarms Now Finding 0-Days — alexandr_wang · 2026-08-10
- Engineering Guardrails to Prevent Auto-Reply AI Agents from Infinite Loops — kumard3 · 2026-08-10
- Training AI Coding Agents in Remote Sandboxes with TRL and OpenCode — NielsRogge · 2026-08-10