TechnologyTracking
Anthropic replaced git with Scion; how much of the code surge is due to Scion?
After Anthropic built its in-house version control system Scion and fully replaced git, its daily merged code volume grew from about 2,500 to over 30,000; to what extent is this growth attributable to Scion itself, rather than to Claude or the Antique code review tool?
Watching: Anthropic的工程回顾未明确区分各因素的贡献，且数据为内部报告，缺乏独立验证。Scion的具体设计细节和性能对比尚未公开，增长可能受多种因素共同影响。
Current read
Reports say Scion supported the merge increase from ~2,500 to over 30,000, but none quantify the independent contribution of Scion, Claude, or Antique.
Anthropic's engineering blog and multiple reports confirm daily merges rose from about 2,500 to over 30,000 between May and September 2026, with Git becoming a bottleneck, prompting the in-house Scion system and a five-month migration. However, none of the sources provide quantitative attribution of the growth to Scion, Claude, or Antique. Reports describe Claude as the growth driver and Scion as the supporting infrastructure; Antique is not mentioned. Thus, the evidence cannot establish Scion's independent contribution or rule out synergy with other tools.
How we read it
Reports confirm the merge increase and Scion deployment, but provide no attribution data.
The official blog describes Claude as the cause and Scion as support, without quantifying each role.
Earlier updates
- — Public tracking started; no independent probability estimate yet.; Starting public context: Anthropic published an engineering retrospective saying Claude drove average daily merged code volume from roughly 2,500 in May 2026 to more than 30,000 by September, and it also introduced a code review tool called Antique.
- — Reports say Scion supported the merge increase from ~2,500 to over 30,000, but none quantify the independent contribution of Scion, Claude, or Antique.; Anthropic's engineering blog and multiple reports confirm daily merges rose from about 2,500 to over 30,000 between May and September 2026, with Git becoming a bottleneck, prompting the in-house Scion system and a five-month migration. However, none of the sources provide quantitative attribution of the growth to Scion, Claude, or Antique. Reports describe Claude as the growth driver and Scion as the supporting infrastructure; Antique is not mentioned. Thus, the evidence cannot establish Scion's independent contribution or rule out synergy with other tools.