Why we’re at the beginning of the AI hardware boom | Caitlin Kalinowski (ex–OpenAI, Meta, Apple)

Caitlin discusses the trajectory of virtual reality (VR), acknowledging that despite advanced hardware developed by companies like Meta and Apple, mass adoption has yet to happen. She attributes this partly to social and device design factors, like the need for non-threatening, reactive, and soft-looking devices, which Pixar and Disney excel at. She predicts that the next major step is AR glasses, which will serve as a pragmatic way to access information comfortably without the drawbacks of current phone usage. She describes ongoing challenges with display tech, input methods, and production costs, but remains optimistic about widespread AR adoption.

A shift in interest towards hardware and robotics is evident, driven by the realization that AI's growth will saturate digital realms and the physical world will become the new frontier. Caitlin explains how foundational tech from VR—like perception and SLAM—is directly applicable to robotics, autonomous vehicles, and drones. This technological overlap indicates that physical AI and robotics are primed to grow, especially as AI techniques improve safety, efficiency, and cost-effectiveness.

Caitlin emphasizes that humanoid robots are still prototypes, mainly due to safety, cost, and technological maturity issues. She notes the importance of soft, lightweight designs to ensure safety and describes the complexity of hardware manufacturing, which involves thousands of parts with tight tolerances. Supply chain disruptions, especially for critical components like magnets and actuators, are significant hurdles, with geopolitical implications including the risk of adversarial threats like drones in war scenarios. She advocates for domestic re-industrialization and vertical supply chains to mitigate these risks.

Caitlin shares insights from her experiences at Apple, Meta, and working with iconic leaders like Steve Jobs and Mark Zuckerberg. She highlights Apple's meticulous focus on quality and design, and Zuckerberg’s decision-making efficiency. She stresses the importance of mission alignment, hiring generalists who can adapt, and recruiting AI-native talent to drive innovation. She views the current moment as a golden opportunity for individuals and companies to leverage AI to push technological boundaries.

AI is already transforming the hardware design process, enabling rapid prototyping, CAD modeling, and PCB routing. Caitlin discusses the potential for AI models specialized in physical engineering to accelerate innovation further, but notes the current limitations of AI in understanding physical properties like friction, pressure, and surface texture. She emphasizes that proprietary CAD data remains a barrier and suggests that hobbyists could be early adopters of AI-enhanced design tools, which could eventually disrupt traditional hardware manufacturing.

Caitlin raises concerns about AI safety in physical systems, especially humanoid robots operating near humans. She advocates for soft, compliant, and safety-focused designs. She also highlights how AI-driven drones and robotic systems could dramatically change warfare, underscoring the importance of developing independent supply chains for critical components to prevent strategic vulnerabilities. The conversation touches on the ethical implications of autonomous systems in defense and everyday life.

Caitlin explains her reasons for leaving OpenAI—primarily due to disagreements over decision-making and governance, especially surrounding a recent related Department of Defense deal. She remains optimistic about AI and hardware's potential, urging continuous innovation, responsible development, and multidisciplinary collaboration. She envisions a future where AI enhances individual creativity and capability, with a focus on building a desirable and safe future.