Code Reviews Are More Important Than Ever

In this episode Mike and Matt discuss how code review is becoming one of the most important developer skills as AI takes on more of the actual code writing. With AI generating larger and denser pull requests, reviewing code effectively has become harder - and more critical - than ever.

They break down the real cognitive limits humans face when reviewing code, including how many lines can realistically be reviewed at once and why reviews should be timeboxed to avoid missed issues. The conversation focuses on how to anchor reviews around what truly matters in a codebase, such as security, performance, testing, reliability, and user experience.

Mike and Matt also share practical tips for becoming a better code reviewer, including creating checklists around critical paths, doing multiple review passes, encouraging smaller cascading PRs, and relying on tools like linters, formatters, and AI to handle nits. They wrap up by exploring how AI can assist with code reviews - summarizing diffs, identifying risky areas, and generating edge cases - while leaving final decisions and tradeoffs firmly in human hands.

Show Notes

Intro

As AI progresses I’ve been doing more code reviewing then coding
Code reviews have also changed since AI can do so much in one shot it’s tough to review everything properly
Humans have some loose limitations when it comes to being efficient at reviews
- typically 200-400 lines of code is the optimal amount a human can review, after which effectiveness can drop
- The typical timeframe is you shouldn’t spend more then 1 hour for each review, and ideally less then 30min otherwise focus can drift and things can be missed
I think reviewing code is one of the skills that will be required heavily over the next few years so doubling down on becoming a better reviewer can help you stay competitive in the market

Tip to become a better code reviewer

Figure out what is critical for the codebase you are reviewing
- Access control
- Security
- Speed
- Testing/Reliability
- UX/UI
- Once you figure those out you can anchor your reviews by focusing on these elements deeply and only shallow reviewing the rest
- create a checklist based on the critical paths
Do multiple passes
Leave nits and formatting to the proper tools like prettier/eslint and AI
- if I see any nits that I need to make I’ll just ask the author to run the tools first and let me know when they run so I can finish review
When unclear don’t be afraid to ask for clarification
Timebox, you don’t want review sot last for more then 45minutes to an hour tops
- If you’re consistently reviewing for more then that then discuss cascading PRs so that smaller change sets can be reviewed
Run the code

How can AI Help

Use AI as a reviewer assistant, not the reviewer
- Ask it to:
  - Summarize the diff and list risky areas
  - Generate edge cases and test ideas
  - Spot common security issues in the touched code paths
You still decide: correctness, tradeoffs, and whether it fits your system

How to support the show

Patreon

Prices subject to change and are listed in USD

Support the show from as little as ~$1/month
Get a shoutout at the end of the episode (while supplies last) for just ~$3/month
Help support the HTML All The Things Podcast: Click Here

Scrimba Discount - Coding Courses!

Learn to code using Scrimba with their interactive follow-along code editor.

Join their exclusive discord communities and network to find your first job!

Use our affiliate link for a 20% discount!!

Click the link to take you to the Scrimba site
A pop-up should appear on your screen with the discount amount and instructions on how to claim it
Discount is for new accounts only

We receive a monetary kickback if you use our affiliate link and make a purchase.

Code Reviews Are More Important Than Ever

Listen

Who’s in This Episode?