From AI and robotics to self-driving cars and industry automation: this DLD25 session has it all. Moderated by WIRED editor Greg Williams, the conversation brings together experts Angela Schoellig (TU Munich), Qasar Younis (Applied Intuition), and Limor Schweitzer (Tugbot, Mov.ai) to explore the current state – and future – of AI-powered robotics, autonomous systems, and humanoid robots in industry and mobility applications.
Angela Schoellig highlights the transformative role of machine learning in robotics: “Robots now can understand the world from vision. We cannot imagine anymore a robot without a camera.”
She emphasizes the integration of large language models (LLMs) to help robots understand object relationships, such as recognizing a chair and knowing it can be moved or sat on. There’s a critical gap, however: “Bridging this into the real world, into the physical reality, is still a big challenge”, Schoelling notes.
Qasar Younis discusses the maturity of autonomous vehicles, stating, “We’ve moved autonomy from a technical problem to an engineering problem. It’s just cost reduction now.”
Already, self-driving cars can help make the roads safer, Younis says. “Waymo has a tremendous safety record, way better than any human would have. It’s already safer than human beings. I don’t know why we keep gnashing our teeth over this.”
On humanoid robots, Limor Schweitzer describes a market experiencing “a revolution, not an evolution.” While there are at least 20 companies producing humanoid robots for industry automation, it’s not clear whether their human shape offers advantages over purpose-built robots, Schweitzer argues. “We are going to see them deployed in use cases that are not perfect for them.”
The panel discussion also touches on safety and human adaptation. Schoellig stresses the need for multi-layered safety systems, from hardware to high-level reasoning algorithms, while Schweitzer emphasizes the importance of gaining confidence in working alongside humanoid robots.
