Sentiero Ventures’ Post

This recent demo by Toyota Research Institute and Stanford University School of Engineering is easily one of my top 3 inspirational videos for Control Systems and Robotics. The first is Atlas doing parkour by Boston Dynamics and the second is Chamipion-level Autonomous Drone Racing by the University of Zurich, Robotics and Perception Group. The most noteworthy technical point was the usage of PINNs (Physics-Informed Neural Nets) in an NMPC that works in real-time (50Hz) for a safety-critical and complex system with a low data regime. We should also not overlook the role of excellent hardware engineering that enabled this incredible performance.

What’s better than a self-driving and auto-drifting Supra? Two of 'em. Toyota Research Institute asked us to produce a video to help explain how tandem drifting can help us all avoid crashes. Nothing more fun than a night shoot with fully autonomous race cars. Full video here...https://lnkd.in/djCPbPUx

2025 is shaping up to be the year AI truly transforms our world. From personalized learning to autonomous systems, augmented working, and even quantum AI—are we ready for the future? 🌐🤖 #AI2025 #FutureOfWork #TechTrends

2025 is shaping up to be the year AI truly transforms our world. From personalized learning to autonomous systems, augmented working, and even quantum AI—are we ready for the future? 🌐🤖 #AI2025 #FutureOfWork #TechTrends

The dawn of evolving AI is here! With intelligent computers and robots capable of spatial intelligence, they're not just tools, but reliable partners. A world where AI’s understanding of our spatial reality enhances human productivity and prosperity is within reach! #AI #FutureTech

As the final contribution of my PhD at Stanford University I developed the motion planning and control algorithm that enabled the world’s first autonomous tandem drift! This approach integrates a carefully trained neural network vehicle model, with real-time optimization-based planning and control. After months and months of testing and validation, it was a surreal experience to be the safety driver in Takumi, the chase vehicle, and finally flip the switch to go “autonomous”.  The result is what you see in this video — an autonomy framework that can perform highly dynamic maneuvers at the limit of friction (drifting) and react to avoid collisions with other moving agents in the environment. This project demonstrates that carefully designed autonomous systems, enhanced by data-driven models, can deliver capabilities that rival even the best human drivers in the world. A huge shout out to the teams at Toyota Research Institute and my lab at Stanford, the Dynamic Design Lab. Without their support, such an incredible project would be impossible. Full video: https://lnkd.in/g-rAAWaV Find more technical details here: https://lnkd.in/g-jhBxg7 #AI #artificialintelligence #neuralnetwork #autonomousvehicles #autonomy

Synthetic data is revolutionizing AI by providing artificially generated datasets that mimic real-world data while addressing challenges like data scarcity, privacy concerns, and cost. Created through techniques like generative adversarial networks (GANs), simulations, and data augmentation, synthetic data enables AI systems to train on diverse, scalable, and customizable datasets. It finds applications in autonomous driving, healthcare, robotics, and finance, offering advantages like ethical compliance and cost efficiency. However, challenges such as bias, validation, and limited realism remain. As adoption grows, ongoing research focuses on improving the fidelity, fairness, and cross-domain utility of synthetic data, making it an essential tool for advancing AI development. #SyntheticData, #ArtificialIntelligence, #MachineLearning, #DataPrivacy, #GenerativeAI, #AutonomousDriving, #AIResearch, #AIEthics, #DataAugmentation, #GANs, #AIApplications, #FutureOfAI, #AITrends, #TechInnovation, #AIPOD-CAST, #AIGenerated https://lnkd.in/g-HbGGwY

Robotics and AI. Listen to this explanation of how manufacturing might change. https://lnkd.in/eFVi_HXj

Today’s AI has no explicit world model, a prerequisite for true intelligence and reasoning. In some aspects a cockroach is more advanced than our most sophisticated robots. Superintelligence Computing Systems fills the gap with a betondcnext generation AI robot brain, which is not an LLM.