Why I joined Blacksmith to work on storage again
Summary
This post explains how CI workloads create unique storage challenges at fleet scale, arguing that ephemeral microVMs and high churn require storage designed for speed, correctness, and selective durability. It outlines five core problems—getting the base image and dependencies onto hosts quickly, warming environments efficiently, deciding which writes to durability, preserving artifacts, and safe, shared caching—before inviting engineers to join Blacksmith to tackle these storage foundations.