Flexion's AI Intern: Why Humanoids Need New Brains
flexion robotics, founded by ex-nvidia engineers, unveiled an ai system that lets humanoids execute complex multi-step tasks autonomously by combining reinforcement learning with video observation, signaling a shift toward general-purpose robot brains.
flexion robotics, founded by ex-nvidia researchers, released a demo of a modified unitree humanoid performing complex office chores like retrieving packages and using elevators without human teleoperation. the system operates via a master ai model that analyzes human videos to match necessary skills with those learned in simulation, relying heavily on reinforcement learning for both planning and low-level motor control. this technology addresses the critical bottleneck in robotics: shifting from single-function machines to general-purpose systems capable of handling unfamiliar environments. analysts estimate the market for these robot foundation models could reach $150 billion by 2036, though flexion will face stiff competition and must closely integrate with hardware manufacturers.