Tom Dörr shares a project involving simulating one million AI agents on social media. The post links to an external resource and includes a media attachment, suggesting a large-scale experiment or research effort exploring how AI-generated agents behave and interact within social media environments. This type of simulation could provide insights into emergent behavior, information propagation, and the dynamics of AI-driven online communities at massive scale.
For National Robotics Week 2026, NVIDIA showcases a comprehensive roundup of breakthroughs in physical AI — the convergence of robotics, simulation, and foundation models that is accelerating how intelligent machines are built, trained, and deployed. Key announcements from GTC 2026 include the NVIDIA Isaac GR00T open models for natural language-driven robot control, Cosmos world foundation models for synthetic data generation, the open-source Newton 1.0 physics engine for dexterous manipulation, and expanded simulation tools like Isaac Sim 6.0 and Isaac Lab 3.0. Together, these form a full-stack cloud-to-robot workflow enabling developers to iterate in virtual environments before real-world deployment.
The post highlights diverse applications across industries: PeritasAI is integrating physical AI into surgical robotics for real-time coordination in operating rooms; Maximo (incubated by AES Corporation) completed a 100-megawatt solar installation using autonomous robot fleets; and Aigen deploys solar-powered agricultural rovers running inference on NVIDIA Jetson Orin to perform precision weed control without herbicides. Research efforts from the University of Maryland focus on humanoid robots for household tasks, while projects like OceanSim from the University of Michigan bring GPU-accelerated simulation to underwater robotics.
Additional highlights include Doosan Robotics using Cosmos Reason for adaptive warehouse palletizing, Mimic Robotics achieving dramatic sample efficiency gains with video-action models, and the OpenClaw platform running locally on Jetson Thor for private, low-latency edge AI. The second cohort of the AWS MassRobotics Fellowship features nine NVIDIA Inception startups spanning humanoid robotics, haptics, and agricultural automation. The overall message is clear: the combination of simulation, world models, and edge computing is fundamentally reshaping how robots learn and operate in real-world environments.
