Company
Date Published
Author
Git
Word count
1426
Language
English
Hacker News points
None

Summary

As software engineering evolves with AI coding assistants like ChatGPT and GitHub Copilot, traditional code review practices are being challenged by the increased volume of code from both human and AI developers. An analysis of over 5 million pull requests from 35,000 engineers reveals that small companies ship code three times faster than larger organizations, largely due to their higher likelihood of skipping code reviews. TypeScript is gaining popularity among startups, while JavaScript's usage declines, and larger companies tend to use more established languages like Java and Python. Companies using monorepos face significantly longer merge times compared to those with multiple repositories, often due to the strategic organization of code rather than speed concerns. The ideal pull request size for expedited review is around 50 lines, as smaller PRs are merged much faster than larger ones. Additionally, CI run times tend to be twice as long at larger companies, and these organizations are more likely to use merge queues to streamline the process, although advancements in technology are allowing earlier adoption of these systems.