Industrial AI Podcast Physics and Causality in Robotics
13 snips
Nov 19, 2025 Join Professor Stratis Gavves from the University of Amsterdam, an expert in physics-informed robotics, as he delves into the critical role of physics and causality in robot learning. Discover why traditional robot manufacturers struggle with adopting advanced world models, and how integrating multimodal data can lead to more explainable and efficient robots. Stratis also predicts a future of specialized models versus a universal one, exploring the balance between edge and cloud architectures for optimal robotics performance.
AI Snips
Chapters
Transcript
Episode notes
Industrial AI Cloud And Sovereignty Debate
- Peter Seeberg described the Industrial AI Cloud partnership between Deutsche Telekom and NVIDIA, citing a planned Munich data center.
- He framed the project within the broader European debate on digital sovereignty.
SAP Enters Tabular Model Space
- Robert Weber noted SAP released an open model for tabular data and promoted it heavily.
- He emphasized SAP's natural interest given its custody of vast enterprise tabular datasets.
AI Models Can Propose Valid Discoveries
- The Cosmos model (FutureHouse) produced validated scientific discoveries, illustrating AI's potential in research.
- Such models can generate testable hypotheses that humans later confirm.