Despite testing with countless objects, Ambi Robotics discovered their system struggled with a common item they hadn't prioritized: plastic shipping bags. Bags fold and lose suction unpredictably, highlighting how real-world deployment uncovers critical edge cases that extensive lab testing misses.
Related Insights
Ken Goldberg's company, Ambi Robotics, successfully uses simple suction cups for logistics. He argues that the industry's focus on human-like hands is misplaced, as simpler grippers are more practical, reliable, and capable of performing immensely complex tasks today.
The founders initially focused on building the autonomous aircraft. They soon realized the vehicle was only 15% of the problem's complexity. The real challenge was creating the entire logistics ecosystem around it, from inventory and fulfillment software to new procedures for rural hospitals.
Leading roboticist Ken Goldberg clarifies that while legged robots show immense progress in navigation, fine motor skills for tasks like tying shoelaces are far beyond current capabilities. This is due to challenges in sensing and handling deformable, unpredictable objects in the real world.
The most complex challenge in robotics isn't just hardware or software alone, but the "boring" problem of calibration where they meet. Seemingly minor physical misalignments create cascading, hard-to-diagnose software issues that require deep, cross-functional expertise to solve.
Developers often test AI systems with well-formed, correctly spelled questions. However, real users submit vague, typo-ridden, and ambiguous prompts. Directly analyzing these raw logs is the most crucial first step to understanding how your product fails in the real world and where to focus quality improvements.
Instead of simulating photorealistic worlds, robotics firm Flexion trains its models on simplified, abstract representations. For example, it uses perception models like Segment Anything to 'paint' a door red and its handle green. By training on this simplified abstraction, the robot learns the core task (opening doors) in a way that generalizes across all real-world doors, bypassing the need for perfect simulation.
While on-device AI for consumer gadgets is hyped, its most impactful application is in B2B robotics. Deploying AI models on drones for safety, defense, or industrial tasks where network connectivity is unreliable unlocks far more value. The focus should be on robotics and enterprise portability, not just consumer privacy.
Self-driving cars, a 20-year journey so far, are relatively simple robots: metal boxes on 2D surfaces designed *not* to touch things. General-purpose robots operate in complex 3D environments with the primary goal of *touching* and manipulating objects. This highlights the immense, often underestimated, physical and algorithmic challenges facing robotics.
Moving a robot from a lab demo to a commercial system reveals that AI is just one component. Success depends heavily on traditional engineering for sensor calibration, arm accuracy, system speed, and reliability. These unglamorous details are critical for performance in the real world.
The "bitter lesson" (scale and simple models win) works for language because training data (text) aligns with the output (text). Robotics faces a critical misalignment: it's trained on passive web videos but needs to output physical actions in a 3D world. This data gap is a fundamental hurdle that pure scaling cannot solve.