CareMotion: Video-Taught Assistive Robot for Family Care
A consumer-facing assistive robot (or phone-based robotic arm) that a family member teaches to perform one specific personal task — pouring a glass of water, fetching a mug — by filming themselves doing it.
Family caregivers of people with mobility limitations, and adults aging in place, who want one trusted 'helper' to perform a specific daily task.
- Teach-by-video onboarding so a caregiver can customize the robot to one home's layout and routine in under 15 minutes
- Single-task focus per robot to keep UX and reliability tight (one skill done well beats 20 skills done badly)
- Caregiver-share mode so a spouse, adult child, or home health aide can each add demonstrations to the same memory
- Privacy-by-design: all video learning stays on the device unless explicitly uploaded by the user
WAM-TTT shows it's now technically feasible to steer a robot foundation model with raw human video at test time — that's the missing piece that has kept consumer assistive robotics stuck on rigid pre-programmed behaviors.
Elder care assistive robot market is a clear macro tailwind: $3B in 2024 growing ~14% CAGR toward $11B+ by 2034, with 22% of adults 85+ needing help with personal care and well-documented caregiver burden that assistive tech measurably reduces.Elder Care Assistive Robots Market (2024-2034) ↗Effect of assistive technology on the burden of caregivers to older adults ↗
Direct consumer competitors in 'video-taught single-task' are essentially absent — Hello Robot Stretch is research-grade and pre-programmed, Jibo/Kuri are dead, ElliQ is social/voice only; the WAM-TTT-style video-steering wedge is genuinely novel even as Tesla Optimus and Figure loom as future general-purpose threats.Stretch 4: Hello Robot's Assistive Mobile Manipulator ↗Jibo is dead (again) ↗
Consumer home robotics has a graveyard of failures (Jibo, Kuri, Anki) and current elder-care robots typically sell to institutions or via subscription; willingness-to-pay among family caregivers exists but is unproven at consumer hardware price points for physical robots.Healthcare Companion Robots Market Size & Share 2024 - 2032 ↗What Happened to Jibo? The Friendly Robot That Got Left Behind ↗
Demographic aging is a 20+ year structural certainty and caregiver shortages are worsening; the video-imitation paradigm aligns with the broader foundation-model trend, but single-purpose robots risk being undercut by arriving general-purpose humanoids from Tesla/Figure.Robots are learning to do housework from videos of humans doing chores ↗Informal caregivers' perception of assistive robots in eldercare ↗
WAM-TTT is still an arXiv preprint, not a productized stack; consumer-grade hardware plus safety/liability for frail users, plus the gap between demo and reliable in-home manipulation, makes this a multi-year, capital-intensive build rather than a quick app.US household humanoid robot learns kitchen chores by watching humans ↗