sources

# Check Preferred Sources Monitor preferred sources for new Physical AI content. ## Usage ```text /sources ``` ## Instructions When this command is invoked: 1. **Read preferred sources** from `research/tools/preferred-sources.md` 2. **Check each source category** (in priority order): **Preprint Searches — Architectures & Methods** (Priority 1): - Search arXiv and TechRxiv for architecture/method queries (JEPA, EBM, world models, robot foundation models) - Filter to last 30 days only - Identify papers not yet in publications.md **Preprint Searches — Physical AI Platform & Infrastructure** (Priority 2): - Search for platform/infrastructure queries (Physical AI deployment, ROS2, digital twins, sim-to-real) - Filter to last 30 days **GitHub Organizations** (Priority 3): - Check listed GitHub orgs for new repos and releases - Include: facebookresearch, google-deepmind, nvidia-cosmos, Physical-Intelligence, ros2, Genesis-Embodied-AI, huggingface/lerobot - Filter to last 30 days **Company Blogs — Big Tech & Research Labs** (Priority 4): - Meta AI, Google DeepMind, NVIDIA, Siemens, Schneider Electric - Look for Physical AI announcements, product launches, research releases **Company Blogs — Startups & Emerging Players** (Priority 5): - Figure AI, Agility, 1X, Unitree, Physical Intelligence (pi.website/blog, pi.website/research), Covariant, Skild, Intrinsic - Foxglove, Formant, MathWorks - Look for funding, product launches, technical posts **Industry News & Startup Tracking** (Priority 6): - The Robot Report, TechCrunch Robotics/AI, IEEE Spectrum Robotics - Crunchbase (via WebSearch) for funding signals - ROS Discourse for community developments **Simulation & Digital Twin Ecosystem** (Priority 7): - NVIDIA Omniverse / Isaac Sim release notes - Open Robotics blog, Gazebo Community - Eclipse Foundation (Ditto releases) **Standards & Governance** (Priority 8): - ROS Enhancement Proposals (REPs) - CHAOSS, OpenSSF Scorecard updates - OSRA announcements **Researcher Pages** (Priority 9): - Check Google Scholar pages for new publications - Check personal websites/blogs for new content 3. **Rank findings**: - Relevance to Physical AI platform intelligence - Recency - Source credibility - Potential impact (citations, engagement, funding amount) 4. **Present top findings**: - Group by category (Papers, Projects, Ecosystem, Solutions, Industry News) - Show top 3-5 from each category - Include: title, source, date, 1-sentence description 5. **Ask which to add**: "Which would you like me to add? (e.g., 'P1, P3, E1' for Papers 1,3 and Ecosystem 1)" 6. **Add selected items** using `/add` workflow ## Output Format ```text Checked preferred sources (last 30 days): PAPERS (arXiv) P1. [2026-06-10] "V-JEPA 2.1: Temporal Coherence in Latent Space" Authors: Bardes et al. (AMI Labs) Extends V-JEPA with improved temporal prediction, 12% better on downstream robotics tasks P2. [2026-06-05] "Physics-Aware World Models for Industrial Digital Twins" Authors: Chen et al. (Siemens Research) Combines PhysicsNeMo surrogates with world models for factory optimization PROJECTS (GitHub) G1. [2026-06-08] Genesis-Embodied-AI/genesis-world v0.8 New release with ROCm support and improved physics fidelity (29K stars) ECOSYSTEM E1. [2026-06-12] Skild AI announces $300M Series A Building general-purpose robot foundation model, CMU spinout Source: TechCrunch SOLUTIONS S1. [2026-06-01] NVIDIA Isaac ROS 4.0 released Adds support for GR00T N1.7 inference, new perception pipeline Source: NVIDIA Developer Blog INDUSTRY NEWS N1. [2026-06-07] OSRA announces ROS2 LTS support policy 5-year LTS for Iron Irwini, enterprise certification program Source: Open Robotics Blog Which would you like me to add? (e.g., 'P1, G1, E1') ``` ## Avoiding Duplicates Before presenting results: - Check if paper title/URL already exists in publications.md - Check if project already exists in projects.md - Check if player already exists in ecosystem.md - Only show genuinely new content ## Error Handling - If a source is unavailable, skip and note it - If no new content found, inform user: "No new content in preferred sources (last 30 days)" - If many results (>15), present top-ranked only