AI for Science heats up as CurieOS structures long-running research tasks with DAGs
dr_alphalyrae · x · 2026-09-07
A thread on labs pivoting to AI for Science. James Gin introduces CurieOS, which automatically structures long-running scientific tasks with DAGs behind the scenes for coherence in hard scientific and engineering problems. Commenters note Orbital has been compounding research and closing the silico-to-real gap for years.
More from coding & agent
- Codex agent burns 13.3M tokens over 26.5 hours on a single build — nijfranck · 2026-09-07
- LlamaIndex founder uses GPT-6 Astra to fully auto-generate product demo videos — llama_index · 2026-09-07
- Bjarne Stroustrup: AI-generated code is bloated and buggy, natural language as programming language is 'idiotic' — bendee983 · 2026-09-07
- AI rewrites TrackMania physics engine in pure C, then trains PPO agents to beat author's records — burny_tech · 2026-09-07
- Zero critical vulnerabilities found across 120 Replit-built apps, CEO weighs in — amasad · 2026-09-07
- Unverified: claims of using 'GPT-6-ASTRA' inside Claude Code — BLUECOW009 · 2026-09-07