The most underrated part of code review is knowledge transfer — junior devs learning patterns from seniors through comments. Automated tools catch bugs but
can't replicate that. The question is whether teams actually use review that way or just treat it as a merge gate.
I've realised that agents are doing better code reviews than me lately, they can catch more logic issues, performance issues, etc. Now, I focus on whether the intent/objective is addressed and leave the rest to a review agent.
The most underrated part of code review is knowledge transfer — junior devs learning patterns from seniors through comments. Automated tools catch bugs but can't replicate that. The question is whether teams actually use review that way or just treat it as a merge gate.
I think the part that often gets overlooked about code review is that it is useful for maintaining knowledge of the codebase for the reviewer.
I know our codebase like the back of my hand specifically because I'm reviewing almost every PR going into it.
I've realised that agents are doing better code reviews than me lately, they can catch more logic issues, performance issues, etc. Now, I focus on whether the intent/objective is addressed and leave the rest to a review agent.
Honestly, when the context is set up properly, GPT’s top models are just smarter. That’s been true for a long time.