エピソード

  • Why Robot Arms Still Can't Recognise a Spilled Drink

    Why Robot Arms Still Can't Recognise a Spilled Drink
    2026/06/08
    In this episode of The Robotics Podcast, Lucas and Luna tackle a surprisingly hard problem: why even advanced robot arms struggle to identify a liquid spill on a counter. They explore the physics of specular reflection, the limits of today's vision systems, and how startups are using multispectral cameras and thermal imaging to solve it. Along the way, they reference a 2023 MIT study showing that standard lidar fails on transparent puddles, and discuss why this matters for home robotics and factory floors alike. A concrete look at one of the unsung bottlenecks in autonomous manipulation. #RobotVision #LiquidSpillDetection #SpecularReflection #RoboticsChallenge #MultispectralCameras #ThermalImaging #AutonomousManipulation #UnstructuredEnvironments #MITCSAIL #LidarLimitations #TransparentSurfaces #HomeRobotics #IndustrialAutomation #GraspingAndManipulation #ComputerVision #Technology #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    9 分
  • Why Robot Arms Still Can't Open a Jar

    Why Robot Arms Still Can't Open a Jar
    2026/06/08
    Episode 38 of The Robotics Podcast tackles a deceptively hard problem: why robot arms struggle with twist-off jar lids. Lucas and Luna break down the physics of rotational grip, the role of compliant torque sensors, and a 2025 MIT study that found even state-of-the-art grippers fail on 30 percent of standard mason jar tests. They contrast industrial automation with the surprising complexity of a household task, and explore what this means for assistive robotics in homes. Specific examples include the torque feedback loop, the problem of varying lid materials, and why a 2026 startup pivot from manufacturing to kitchen robotics ran into the same wall. A concrete look at how the gap between factory and home remains stubbornly wide. #RobotGrippers #JarOpening #RoboticsChallenge #MIT #TorqueControl #Compliance #AssistiveRobotics #DeformableObjects #Grasping #IndustrialRobots #HouseholdRobotics #Startup #Technology #FexingoBusiness #BusinessPodcast #TheRoboticsPodcast #AutonomousSystems #Hardware Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    6 分
  • Why Robot Grippers Can't Handle a Leather Belt

    Why Robot Grippers Can't Handle a Leather Belt
    2026/06/07
    Episode 37 of The Robotics Podcast explores the overlooked challenge of grasping limp, high-friction materials like leather belts. Lucas and Luna dive into the physics of friction, the limits of vacuum grippers, and why even a $50,000 robot arm fails at something a human does without thinking. They discuss real-world failures in automotive assembly and a surprising fix involving sandpaper and a software tweak. If you think robot dexterity is almost solved, this episode will change your mind. #Robotics #RobotGrippers #LeatherBelt #FrictionPhysics #AutomotiveAssembly #MaterialHandling #RobotLimitations #Grasping #IndustrialRobots #Manufacturing #TechChallenges #Automation #RobotDexterity #FexingoBusiness #BusinessPodcast #Technology #Hardware #Podcast Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    8 分
  • Why Robot Hands Still Can't Handle a Screwdriver

    Why Robot Hands Still Can't Handle a Screwdriver
    2026/06/07
    Episode 36 of The Robotics Podcast dives into one of the most deceptively hard problems in manipulation: using a screwdriver. Lucas and Luna explore why robots, despite having precise force sensors and vision, struggle with rotational tools. They break down the physics of torque reaction, the 'hand-eye coordination' gap, and a key paper from MIT's CSAIL showing that even state-of-the-art robots fail on 30% of screwdriver insertions. The hosts also touch on how this limits automation in assembly and repair, and why solving it could unlock huge value in manufacturing and construction. Plus, a brief moment on how listener support keeps the show ad-free. #Robotics #RobotHands #Screwdriver #Manipulation #Torque #Assembly #Automation #MITCSAIL #ForceControl #Dexterity #Hardware #IndustrialRobots #AutonomousSystems #Technology #FexingoBusiness #BusinessPodcast #Podcast #RobotFailures Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    9 分
  • Why Robot Hands Still Can't Screw in a Lightbulb

    Why Robot Hands Still Can't Screw in a Lightbulb
    2026/06/06
    Episode 35 of The Robotics Podcast tackles a deceptively simple task that still stumps every robotic hand on the market: screwing in a lightbulb. Lucas and Luna break down the three specific failure modes — thread alignment, torque feedback, and fragile grip — using the example of a standard E26 bulb and a $120,000 industrial arm. They discuss how human instinct for 'wiggle and push' is brutally hard to code, why force-torque sensors aren't sensitive enough, and what the latest research from MIT's soft robotics lab might change. The episode also explores a surprising parallel with an earlier challenge (USB insertion) and why adding vision doesn't fix the core problem. A concise, example-rich deep dive for anyone curious about why your future robot butler still can't change a lightbulb. #RobotHands #Robotics #LightbulbChallenge #ForceTorqueSensors #SoftRobotics #MITRobotics #IndustrialRobots #Grasping #AssemblyAutomation #Technology #Podcast #TechPodcast #FexingoBusiness #BusinessPodcast #Automation #RoboticsResearch #E26Bulb #ThreadAlignment Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    8 分
  • Why Robot Hands Still Can't Fold a T-Shirt

    Why Robot Hands Still Can't Fold a T-Shirt
    2026/06/06
    Episode 34 of The Robotics Podcast tackles the deceptively difficult problem of robotic garment manipulation. Lucas and Luna explore why folding a t-shirt remains a grand challenge in robotics, from the high-dimensional state space of fabric to the limitations of current grippers and perception systems. They examine the Berkeley AutoLab's work on speed-scaling cloth manipulation, the 'Bidirectional Cloth Manipulation' paper from 2024, and why even the latest deep learning models struggle with deformable objects. The episode also touches on the broader implications for automation in logistics and home robotics, and why this problem is a litmus test for general-purpose manipulation. #Robotics #Technology #RobotHands #GarmentManipulation #DeformableObjects #BerkeleyAutoLab #ClothFolding #DeepLearning #Grasping #Perception #Automation #Logistics #HomeRobotics #FexingoBusiness #BusinessPodcast #RobotLimits #Manipulation #Research Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    9 分
  • Why Robot Arms Still Cant Pick Up a Lego Brick

    Why Robot Arms Still Cant Pick Up a Lego Brick
    2026/06/05
    In this episode of The Robotics Podcast, Lucas and Luna explore a deceptively simple question: why can't a state-of-the-art robot arm reliably pick up a single Lego brick? They dive into the physics of precision grasping, the limits of current force-torque sensors, and the gap between industrial automation and human dexterity. Along the way, they discuss a 2025 study from the University of Tokyo that found robot hands fail on small, rigid objects over 30% of the time, even with advanced vision systems. The conversation touches on why Lego bricks are a perfect stress test for robotic manipulation and what this means for future automation in electronics assembly and packaging. #Robotics #RobotArms #Lego #Manipulation #PrecisionGrasping #ForceSensors #Dexterity #IndustrialAutomation #UniversityOfTokyo #GraspingFailure #Technology #FexingoBusiness #BusinessPodcast #Podcast #Manufacturing #Hardware #AI #Automation Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    9 分
  • Why Robot Arms Still Struggle to Wipe a Counter

    Why Robot Arms Still Struggle to Wipe a Counter
    2026/06/05
    In this episode of The Robotics Podcast, Lucas and Luna explore the surprising difficulty of one of the most mundane household tasks: wiping a countertop. They break down the physics of contact forces, the challenge of maintaining consistent pressure across uneven surfaces, and why roboticists call this 'the wiping problem.' Using examples from recent research at Carnegie Mellon and Boston Dynamics' latest cleaning demos, they explain why a task that takes humans zero thought requires millions of data points for a robot. The conversation also touches on how this connects to broader challenges in mobile manipulation and why your Roomba can't clean a spill. If today's tech insight was worth a coffee to you, check out buy me a coffee dot com slash fexingo to support ad-free episodes like this one. #RobotWiping #Robotics #Manipulation #ForceControl #CarnegieMellon #BostonDynamics #MobileManipulation #CleaningRobots #AutonomousSystems #IndustrialRobots #Technology #FexingoBusiness #BusinessPodcast #Podcast #RobotChallenges #TactileSensing #StretchRobot #HouseholdRobots Keep every episode free: buymeacoffee.com/fexingo
    続きを読む 一部表示
    11 分