NVIDIA post-trained Nemotron with Factory's agent harness inside its air-gapped network
maddiehfaulkner · x · 2026-11-19
Factory reveals NVIDIA post-trained its Nemotron models using Factory's platform, running entirely within NVIDIA's air-gapped network — the only commercially supported harness deployed there, giving researchers long-running agents and full execution traces.
More from coding & agent
- Open-source prefill engine hits 3.2× vLLM throughput serving GLM-5.3-Flash on 4 B200s — bariskasikci · 2026-11-19
- User reports Muse agents running slow and stuck: 3 of 10 idle for up to 9 days — RichardsonDx · 2026-11-19
- LangChain's Harrison Chase calls Deepagents the best base harness for building agents — BraceSproul · 2026-11-19
- Biohub launches MCP server to run protein models ESMC and ESMFold2 inside Claude, ChatGPT, Cursor — AllThingsApx · 2026-11-19
- Hands-on: GPT 6.2 Astra tackles 6 Blender MCP 3D builds in one-shot prompts — reddit0_r · 2026-11-19
- Designers need a prototype playground: a copyable prompt-based AI design workflow — round · 2026-11-19