Syntax - Tasty Web Development Treats
Syntax - Tasty Web Development Treats

1019: LGTM, Ship It: The AI Code Review Problem

July 8, 2026

AI Summary

5 min read

When AI Writes Code Faster Than Humans Can Read It

A developer on the Syntax podcast described a nightmare scenario: individual contributors on his team are merging 60 pull requests per week, with changes spanning 20 to 70 files and thousands of lines of code. The workflow is simple—an engineer asks an AI agent to implement a feature or fix a bug, opens a PR, addresses automated review comments if necessary, waits for CI to pass, and merges. In many cases, there is little to no meaningful human review of the actual code. The result, he reports, is "a steadily increasing amount of technical debt, recurring bugs, architecture inconsistencies, fragile implementations, and code bases that are becoming progressively harder to extend and maintain."

The Speed Run Problem

Wes Bos and Scott Tolinski, the hosts of Syntax, recognized this pattern immediately. "Turns out all of the stuff that we talked about for 10 years about good software engineering still applies," Wes said. "You're just speed running it on fast mode now where it may have taken you six years previously to get to a point where your code base is a mess and things are duplicated."

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

What you'll learn

  • 1 (00:00) **Episode Introduction & First Question: Understanding AI-Generated Code** - The hosts introduce the Syntax potluck format and dive into a question about the difficulty of understanding code written by AI agents.
  • 2 (06:25) **The 60 PRs Per Week Problem: AI-Generated Tech Debt** - A question describes a team where engineers merge 60 PRs per week with little human review, leading to massive technical debt.
  • 3 (11:24) **Explainer: What Are Local AI Models?** - A question asks for an explainer on local models, clarifying the common misconception.
  • 4 (16:10) **Jujutsu (jj): A New Git Alternative?** - A question asks about Jujutsu (jj), a version control system, and whether it's a superior alternative to Git.
  • 5 (22:18) **Pricing Freelance Work in the Age of AI** - A question asks how contractors should price work when AI makes them much faster.
  • 6 (24:52) **CSS Oddities: Why `-webkit-box-reflect` Was Never Standardized** - A question asks why the `-webkit-box-reflect` property was never standardized or implemented in Firefox.
  • 7 (27:44) **Angular vs. React: The Library Debate in the AI Era** - A question from an Angular developer asks how to convince their team to use more external libraries, contrasting with the React ecosystem.

+ Full timestamped outline available in the app

Guests on this episode

Show Notes

This episode tackles the growing pains of AI-assisted development, from the struggle of reviewing thousands of lines of agent-generated code to the mounting technical debt when teams merge PRs without meaningful human review. Scott and Wes also dig into local models, whether jujutsu really beats git, how freelancers should price work in the AI era, and getting your team on board with external libraries.

Show Notes

Sick Picks

Hit us up on Socials!

Syntax: X Instagram Tiktok LinkedIn Threads

Wes: X Instagram Tiktok LinkedIn Threads

Scott: X Instagram Tiktok

Syntax - Tasty Web Development Treats