[CLSTR]
< Back to situations

Monitor this situation.

[SITUATION] · [QUIET] · [TECHNOLOGY]

2 clusters · 6 sources · 15 days · First seen · Last updated

AI-driven software development and version control evolution

Overview

The rapid expansion of AI-driven software development is creating significant bottlenecks for traditional version control systems like Git. Early observations noted that the manual workflows required by Git are increasingly cumbersome for high-speed AI coding agents.

By late 2026, the impact of this shift became evident at Anthropic, where the AI agent Claude became responsible for over 80% of merged production code. This led to a tenfold increase in daily code merges, rising from 2,500 to over 30,000 between May and September 2026. To manage this surge and the resulting strain on continuous integration systems, Anthropic developed a custom client-server version control system called Scion. This new infrastructure, alongside a code review system named Antique, reportedly increased checkout speeds by two orders of magnitude and drastically reduced resource usage compared to previous Git-based solutions.

Entities

Jujutsu · GitHub · Anthropic · Antique · Claude Code

Timeline

  1. [TECHNOLOGY] 4 sources
    Anthropic builds custom Scion version control system to support AI-driven code surge

    Anthropic replaced Git with a custom version control system, Scion, after its AI agent Claude drove daily code merges from 2,500 to over 30,000 in just four months.

  2. [TECHNOLOGY] 2 sources
    Jujutsu version control emerges as an alternative to Git for AI-driven workflows

    A developer explores the shift from Git to the Jujutsu version control system, driven by the need for faster workflows in the era of AI coding agents like Claude Code.

Sources

cryptobriefing.com · dev.to · ifanr.com · infoq.cn · ithome.com · tabnews.com.br