ROME team open-sourced full agent infra: ROLL, ROCK and iFlow CLI before training the model
thisguyknowsai · x · 2026-10-06
A breakdown of the open-source ROME + ALE ecosystem from a Chinese team, which built infrastructure first and trained the model last.
- ROLL: RL training framework with async train-rollout multiplexing
- ROCK: sandboxed execution engine handling 10,000+ concurrent environments
- iFlow CLI: agent orchestration ensuring training-deployment consistency
The author argues most labs try to train agents without this foundation — production agents are about infrastructure, verified data and proper RL, not prompt engineering or raw model scale.
More from coding & agent
- t3 code spawns agent types to review PRs with Alchemy's 20s deploys — samgoodwin89 · 2026-10-06
- Hugging Face adds an RL Environments filter to the Hub, supporting 4 frameworks — lmoroney · 2026-10-06
- Karpathy's arrival makes Anthropic's coders better, says Kuprel — Kuprel · 2026-10-06
- 'DOCX Will Be Replaced by Markdown Within 5 Years' Sparks Pushback from Formats Veteran — bytebot · 2026-10-06
- A browser tab refresh bug kept a server at 100% CPU for 4 days—load scaled with tabs squared — Ok_Negotiation_2587 · 2026-10-06
- Loop-engineering: 8 unattended agent loop patterns to run your repo while you sleep — JafarNajafov · 2026-10-06