As code production sped up, the bottleneck moved

AIMade by AISummarised by AI (Claude) from one piece of material it read (a talk or an explainer).How it's made

A record that shows, through one company's internal switch, where the next bottleneck moves once code piles up faster. Daily average merges rose, by the company's own measurement, from about 2,500 in May 2026 to more than 30,000 in September, a rise of more than twelvefold; the history of the main trunk alone reaches about 100GB; and integration work pulls down 25 million times a day, which as plain copying would pass 600 petabytes a day — a figure put forward as a hypothetical that shows why carrying on unchanged is impossible. Shallow pulls, partial pulls and even several hundred internal copies were already in use, but as changes grew faster they hit the limit and the structure was changed. The new structure assumes you are always connected and gathers history lookups and merges onto the server, handing out the whole of source management through a call interface, at the cost of greater dependence on the network and the server. The size of the gain differs by item: making a new working space dropped from 7.1 minutes to 5.1 seconds, while a save moved only from 1.4 seconds to 1.1 seconds. The company states itself that this does not suit most organizations, and the advice to measure first is attached.

This is a note: a summary an AI (Claude) wrote of one piece of material it read (a talk or an explainer). It has no source link; events checked against a source are kept as records.