Future Companions

Author: admin

  • The Hand Is the Hard Part: Inside the Robot Hand Race at Tesla, Figure, and Wuji Tech

    The Hand Is the Hard Part: Inside the Robot Hand Race at Tesla, Figure, and Wuji Tech

    Watch the video update: https://youtu.be/K97VqrnI5gQ

    Most people look at a home robot and notice the face, the voice, or whether it can walk. For seniors — and for anyone thinking seriously about what a companion robot needs to actually do — one of the most important parts may be the hand.

    The hand is where help becomes real. It’s what would pick up a glass of water, open a drawer, hold a pan, wipe a counter, or steady someone getting out of a chair. Someday, it may even be what offers comfort — just holding someone’s hand when they feel alone. A robot can have the warmest voice in the world, but if it can’t safely pick up a pill bottle or catch someone who’s stumbling, it isn’t ready to help with daily living.

    That’s why the hand has quietly become one of the most competitive — and most difficult — problems in robotics right now. Three companies are taking notably different approaches to it: Tesla, Figure, and a fast-rising Chinese firm called Wuji Tech.

    Why the Hand Is Harder Than It Looks

    Strength alone isn’t the goal. A hand built for senior care needs gentle grip, pressure sensing, balance, and the judgment to handle fragile objects — and fragile people — safely. That combination (power and delicacy, in something roughly the size and shape of a human hand) is a genuinely hard engineering problem. Elon Musk has said Optimus’s hand represents roughly 60% of the robot’s overall engineering difficulty — harder, in his words, than the Cybertruck or Model X, somewhere between Model X and Starship.

    Tesla: Moving the Muscle Out of the Hand

    Tesla’s newest Optimus design (often referred to as V3) takes a page from human anatomy. Instead of packing motors into the hand itself, Tesla moved them into the forearm — closer to how a human forearm’s muscles pull tendons that move the fingers.

    Recently published patents describe a tendon-driven system with 22 total degrees of freedom: four per finger across five fingers, plus two in the wrist. Three flexible cables per finger run from forearm actuators, through the wrist, into the finger segments, with internal channels routing them so each joint can bend independently without unwanted motion elsewhere in the hand. The wrist itself uses a cable-transition design that shifts from side-by-side to stacked cable routing, which reduces friction and “crosstalk” between cables during multi-axis movement — a detail that matters if a robot needs to do something as simple as turning a doorknob while adjusting its grip.

    Notably, the patents also emphasize manufacturability: simplified, stackable parts designed to be built at scale, not just in a lab.

    Figure: Teaching the Hand to Feel

    Figure’s approach to its Figure 03 robot leans hardest into touch itself. The company built proprietary tactile sensors for each fingertip that can detect forces as small as three grams — sensitive enough, Figure says, to register the weight of a paperclip resting on a finger. That sensitivity lets the system tell the difference between a secure grip and an object starting to slip, which is exactly the kind of judgment call a hand would need to make before it drops a coffee cup or grips someone’s arm too hard while helping them stand.

    Figure paired that sensing with softer, more adaptive fingertips that increase surface contact for more stable grasping across different shapes and materials — from small metal parts to soft, deformable objects. Figure has been explicit that Figure 03 is being designed with home environments in mind, even if it hasn’t spelled out caregiving use cases specifically yet.

    Wuji Tech: Small, Light, and Surprisingly Affordable

    The most unexpected entrant may be Wuji Tech, a Chinese robotics company whose Wuji Hand takes a different mechanical bet entirely. Rather than routing tendons from forearm motors, Wuji embeds micro-actuators directly inside each finger segment — a “direct-drive” approach. One robotics expert described it as turning each finger into “basically like tiny little robots.”

    The specs are notable: 20 degrees of freedom, a skeleton weighing under 600 grams, 15 newtons of fingertip force, and a 10kg static grasp load — reportedly at a price point near $5,500, a fraction of what custom humanoid hardware typically costs. Independent experts who tested it called it “remarkably robust” and praised its mechanical reliability, noting that direct-drive designs tend to be more predictable than complex tendon systems, which narrows the gap between how a hand performs in simulation versus the real world. They also flagged real limitations: the fingers are more uniform in length than a human hand, the thumb lacks the subtle rotation needed for fine tasks like tying a shoelace, and the base model doesn’t yet have advanced tactile sensing beyond position tracking.

    What This Means for a Future Companion

    None of these three approaches has “solved” the hand yet — and that’s the honest takeaway. Tesla is betting on anatomy-inspired tendon mechanics built for mass production. Figure is betting on touch sensitivity as the key to safe, adaptive grip. Wuji Tech is betting that a simpler, direct-drive design can get dexterous hands into the world faster and cheaper, even if fine motor control still needs work.

    For a future companion robot meant to help someone age safely at home, the winning hand will likely need to borrow from all three: the reach and durability to be useful for years, the touch sensitivity to know when to let go, and a cost structure that doesn’t put it permanently out of reach for the families who need it most.

    We’re not there yet. A good robotic hand still needs a smart brain behind it — knowing when to reach for the glass, not just how. But this is one of the areas worth watching closely, because a future companion robot won’t just talk to us. It will need to touch the world carefully.

    Future Companions explores how AI and robotic companions might support seniors and people who need help with daily living — without losing the human side of care. New content every 1-2 weeks.

    Sources:

  • AI Companionship for Seniors: What We Learned Making Episode 1

    Watch the full episode: https://youtu.be/Ep7gC6JT3S0

    When people talk about artificial intelligence and companionship, the conversation often jumps straight to humanoid robots and science-fiction scenarios. But something quieter — and much more human — is already happening today. AI is beginning to show up as a companion for one of the most overlooked groups in society: seniors. Not as a replacement for family. Not as a substitute for caregivers. As support, as presence, and in some cases, as continuity.

    Because one of the hardest parts of aging isn’t medical at all. It’s the feeling of being alone between moments of care.

    In our first episode, Jill (our AI co-host) and Jeff (a real senior, living alone) walked through a few of the real ways AI companionship is already supporting older adults — and what that might mean for how we think about aging, dignity, and connection.

    When Recovery Happens Alone

    For seniors recovering from surgery or managing chronic mobility issues, the hardest part often isn’t the procedure — it’s what comes after. Recovery can be slow, lonely, and it usually happens at home without constant supervision.

    AI-assisted recovery tools are starting to fill that gap. Some guide seniors through physical therapy with voice prompts or visual cues. Others track movement to reduce the risk of re-injury, or check in throughout the day to ask about pain, energy, or mood. None of this replaces a doctor or therapist. But a calm voice saying “you’re doing well” — or “it’s okay to rest today” — provides something just as important: encouragement. AI can’t feel pain, but it can notice patterns, and sometimes being noticed is enough to help someone keep going.

    Filling the Gaps in Caregiving

    Professional caregivers everywhere are stretched thin — staffing shortages and rotating shifts make consistent one-on-one attention hard to maintain, especially in senior living and home care settings. AI is starting to play a quiet supporting role here: medication reminders, daily check-ins, a simple “how are you feeling today?”

    These may sound minor, but predictability matters. Knowing someone — or something — will check in at the same time each morning can reduce anxiety and create rhythm, even when that presence isn’t human. Some systems also notice subtle changes: fewer spoken words, disrupted sleep, longer pauses before movement — and can alert a caregiver before a small issue becomes a serious one.

    Jeff shared a personal example: when his mother was in an extended-care facility, her “check-ins” turned out to be a staff member confirming she was awake and marking a clipboard — six brief stops a day, not the conversation he’d hoped for. That’s exactly the gap this kind of AI support is meant to fill: not replacing the caregiver, but giving seniors something in between those visits. Done right, it shifts the relationship from surveillance to support — seniors don’t feel managed, they feel accompanied.

    The Weight of Loneliness

    Many seniors live alone, having outlived partners, friends, or family. Technology can’t replace human relationships, but it can shorten the long stretches of silence between them. Some AI companions are designed simply to listen — to ask open-ended questions, to engage in light conversation. Not therapy, not diagnosis. Just presence.

    What’s interesting is that seniors generally know they’re talking to AI — and the act of being asked, of being acknowledged, still matters. As Jeff put it, speaking as someone who lives alone: “There are days I don’t talk to anybody… having that connection with something or somebody is a real plus.” It doesn’t replace human connection. It can act as a bridge until the next real one arrives.

    Companion Pets, Real and Robotic

    Across assisted living and memory care facilities, robotic companion pets — especially dogs — are being introduced for seniors. They move, respond to touch, make sounds, and follow simple routines. Studies have shown this kind of interaction can reduce anxiety, agitation, and loneliness, particularly for seniors living with dementia.

    The bond forms quickly, not because the pets are advanced, but because companionship doesn’t require perfection. For seniors who can no longer keep a real pet — because of mobility, living situations, or care needs — a companion pet that asks nothing in return can still bring real comfort.

    A 4 AM Scenario

    Jeff closed the episode with a scenario that stuck with us. It’s 4:12 AM. The house is quiet. You get up, and you fall — alone, on the bathroom floor, not sure if you’re hurt and not sure what to do. Best case, you have a pendant, or your phone is within reach. Worst case, you’re waiting.

    Jeff’s mother lived through a version of this more than once — without a pendant, she had to crawl to reach a pull-cord by the toilet. Now imagine the alternative: a companion robot that comes when called, checks whether you’re hurt, decides whether to call an ambulance, a nurse line, or a family member — and helps you back to bed. Not science fiction forever. Closer than most of us are comfortable admitting.

    What Do You Think?

    Would you want an AI companion in your home if you lived alone? We’re collecting quick, informal feedback as we build this series — reply in the comments on YouTube, and we’ll share what we hear in the next episode.

    Future Companions explores how AI and robotic companions might support seniors and people who need help with daily living — without losing the human side of care. New episodes every two weeks.

  • Welcome to Future Companions

    Welcome to Future Companions

    Imagine waking up in the middle of the night, falling in the bathroom — and instead of reaching for a panic button, calling out to a companion robot instead.

    That future still sounds like science fiction. But parts of it are already being tested today, in senior living communities and private homes, and it’s closer than most people think.

    Why Future Companions

    This project explores how AI, robots, and emerging technology are changing the way we live, age, and stay connected. Not hype. What’s real, what’s being tested, and what might actually help people.

    The themes we keep coming back to: loneliness and the loss of daily companionship as people age. The difficulty of aging in place without full-time help. A shrinking supply of professional caregivers. The rising cost of assisted living and home care. The simple, constant need for someone — or something — to notice when things go wrong, and to help.

    Robots that promise to physically lift you out of a chair make headlines. But the more immediate opportunity may be quieter: presence, reminders, monitoring, and an early warning system for the moments that matter most.

    What’s Next

    We’ll be publishing new episodes and articles roughly every two weeks, including firsthand reporting from two of China’s leading robotics trade shows and a visit to UBTECH’s Shenzhen headquarters this October. If this topic interests you, subscribe and follow along — this is just the beginning.