State of Code Reviews in 2026

Let's talk about code reviews in 2026.

AI is here and empowering more and more engineers to create amazing results in short period of time, but there is price.
Feature Release time has is growing, in some cases grinding team velocity into stand still, burning out great engineers and frustrating everyone involved.

It feels like every organisation I talk to has same issue: Code review became biggest bottleneck in release process.

Features delayed by days, in some cases weeks, sitting in Pull Request hell, awaiting for "blessing" of fellow peers (usually one senior person, who hates their life by this point).
And when new code is "accepted" - it is so outdated that full rewrite may be reasonable.

This is unacceptable, no matter on which side of this situation you are.

Why don't we abandon code reviews all together?

• Pair programming solves knowledge sharing issue and improves team morale.
• Better test tooling helps improve software quality and observability.
• Better deployment tooling lowers cost of error, improving brand image in case of inevitable bug slipping through.

Code quality may suffer, but we build software to solve problems. No matter how beautiful, abstract or performant code is - customers won't care much if code does not do what it needs to do.

Too hardcore?

How about "approve by default" if PR has more than one collaborator or been open for at least a day.
Allowing more complex features to be brought to attention of wider group and simple bug fixes, small adjustments or inconsequential changes to be quickly delivered to customers.

Are you brave enough to try?