Muse ships reliability fixes after SEVs left cron jobs and scheduled tasks unrecovered
alexandr_wang · x · 2026-10-05
- The Muse team posted an update on scheduled tasks/cron reliability after several SEVs: the inference tier went unresponsive and VM scheduling left VMs unavailable during deployments, and many crons failed to resume after recovery.
- Fixes deployed: improving KV cache hit rate after enlarging the model's context window, a better VM scheduling algorithm to cut deployment downtime, and a new fleet-wide auth token format with finer-grained access.
- Cron systems inside VMs now have hardened monitoring and alerting to catch unresumed tasks.
More from coding & agent
- Designing a programming language for AI agents, not humans — mark_k · 2026-10-05
- Thoughtworks engineer tests local models for agentic coding on Apple M3 Max and M5 Pro — bibryam · 2026-10-05
- AI agents are about to flood the workforce, and no one's ready: WIRED — Dr_Singularity · 2026-10-05
- New Paper IGP-Bench Teaches AI Agents When to Stop Chasing a Wrong Idea — _sathvikr · 2026-10-05
- AI decompilation era: PS5 is 80% ported to PC and 'closed source' is ending — almmaasoglu · 2026-10-05
- Claude Code ignores steering messages, Codex always acknowledges them — Angaisb_ · 2026-10-05