Nine Mac Minis power 42-way concurrent macOS CI; a code agent setup burns just ~20B tokens/month
yangyi · x · 2026-08-30
A first-hand recap of a home rack built to solve the review bottleneck in AI automation: 3 Proxmox servers, 2 Synology NAS (21TB, running MinIO/Verdaccio/Portainer), and 9 Mac Minis — 8 forming a CI runner pool (five 16GB at 3 concurrent each, one 48GB M4 Pro at 9, two 64GB M4 Pros at 9 each), totaling 42 concurrent macOS CI lanes.
Even so, full checks on every MR still queue, prompting a split of the review structure: which tasks run full checks, which run only dependency-affected ones, and how to tier review standards. Only after these rules are set does Harness automation truly begin.
A notable data point: a code-agent workload burns just over 20B tokens/month; setups hitting hundreds of billions are mostly text/multimodal workloads — likely with Harnesses exploding context in loops. Burning that many tokens is itself an engineering feat.
More from coding & agent
- Simon Willison: code mode is the future of tool calling, if it can be sandboxed — springrod · 2026-10-16
- Agent named 'Ilya' reportedly scores 1.1pt higher on benchmarks — daniel_mac8 · 2026-10-16
- Gemini CLI fix: substring matching inlined unrequested binary files into model context — priyanshu1976 · 2026-10-16
- Building AI agents on real healthcare claims data: the domain work that makes them work — alex_verem · 2026-10-16
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Vibe-coded iPhone Duo sniping bot built with Codex in 17 minutes, 41 tests passing — xiaohu · 2026-10-16