Transcript 0:02 Robots make me feel weird. [gentle music] [object clattering] Shit. [laughs] Those clips aside, I think it's got something to do with the numerous sci-fi movies I've watched where things definitely don't go as planned. 0:20 Westworld, I, Robot, Terminator, but there are too many to list here. I call it the Frankenstein complex, the fear that something we've created will end up destroying us. 0:32 But with the unstoppable rise of AI and advances in robotics, it's not unrealistic to think that domestic robots will arrive sooner rather than later. After all, I keep seeing them on social media. 0:44 Boston Dynamics can do a lot more than backflips now. Tesla's Optimus keeps popping up, and then there's the rise of robot sports in China. 0:54 That last one is really quite strange and makes me no less uncomfortable than watching Blade Runner. But still, it's really quite awesome that we've made it this far. 1:04 Startups with billion-dollar valuations are racing to mass produce domestic robots, and some of the experts who've been skeptical about home robots for years are suddenly changing their tune, 1:15 but not all of them, and it's almost like something of a civil war has erupted in the world of robotics. 1:21 Recently, iRobot, the company behind the Roomba, arguably the most successful domestic robot ever made, filed for bankruptcy. 1:30 And Rodney Brooks, the MIT roboticist who co-founded iRobot, has called the vision of humanoid robot assistants pure fantasy thinking. 1:40 This is all very confusing given I, Robot is also a Will Smith movie that I mentioned earlier, where domestic robots rebel against humans. Oh, hell no. [gun firing] [dramatic music] So which is it? 1:54 Are we really gonna wave in the robot takeover of our homes, or is this another cycle of hype destined to disappoint? 2:02 I've been digging into the evidence, and the answer is more complicated and more interesting than taking either side suggests. 2:14 As of twenty twenty-five, the global market for home robotics was worth roughly fifteen billion dollars, but about seventy percent of that isn't coming from futuristic humanoids. 2:25 It's coming from something already in seventy-eight million North American homes, robot vacuums and mops. Now, a Roomba isn't exactly C-3PO, but these floor cleaners have done something crucial. 2:39 They've normalized the idea of autonomous machines in our homes. They've proven that robots can be reliable, affordable, and genuinely useful. 2:50 Robotic vacuums are now the most widely sold robots globally, with sales growing at around seventeen percent annually. And yet the company that pioneered this market just went bankrupt. 3:02 iRobot's value collapsed from three point five six billion in twenty twenty-one to about a hundred and forty million by the time it went under. 3:10 If the most successful domestic robot company couldn't make it work, what does that tell us about the prospects for far more ambitious humanoid machines? 3:19 Rodney Brooks, who spent three decades building robots at MIT before co-founding iRobot, thinks the current humanoid hype is wildly overblown. 3:29 "Today's humanoid robots will not learn how to be dexterous," he wrote recently, "despite the hundreds of millions or perhaps many billions of dollars being donated by VCs and major tech companies to pay for their training." 3:42 His argument centers on touch. The human hand contains seventeen thousand mechanoreceptors for detecting light touches, with density increasing towards the fingertips. 3:52 These receptors respond to pressure and vibration in sync with fifteen different families of neurons. 3:59 We've trained AI on vast amounts of speech and image data, Brooks point out, but we don't have such a tradition for touch data. Without that, humanoid robots will remain fundamentally clumsy. 4:12 But not everyone shares Brooks' skepticism. 4:14 Sergey Levine, a robotics researcher at Berkeley, who's seen plenty of hype cycles come and go, now gives a median estimate of twenty thirty for fully autonomous household robots. That's just four years away. 4:28 His reasoning centers on a crucial difference between home robots and self-driving cars. Autonomous vehicles need to be essentially perfect from day one. Mistakes can be fatal. 4:39 But home robots can afford to fumble, drop a dish, misfold a T-shirt. That's not a disaster. It's a learning opportunity. 4:48 This means domestic robots can be deployed gradually, starting with simple tasks under supervision and taking on more responsibility over time. 4:57 One particularly dexterous task robots have been able to achieve recently is perfectly folded laundry. I would love a robot that can do just that. The technological case for this rests on recent advances in AI. 5:11 Five years ago, robots were basically sophisticated puppets. They could follow preprogrammed routines, but throw something unexpected at them, and they'd freeze or fail. 5:22 Large language models and vision language action models have changed this. Levine's team at Berkeley has created robots with what they call a visual cortex connected to a motor cortex through an LLM. 5:35 When you tell this robot to clean up the kitchen, it doesn't need detailed programming for every step. The language model breaks down the task internally. 5:44 The vision system identifies objects, and the motor system executes movements. These robots now have something resembling common sense. 5:53 They can look at a wet floor sign and understand it means danger without ever having encountered one before. 6:00 They can predict that a glass will break if dropped without learning through destructive trial and error.As Levine puts it, the big benefit that recent innovations in AI give to robotics is the ability to leverage prior knowledge. 6:14 And crucially, when one robot learns something, they all learn it. 6:18 Companies like Figure AI are developing vision language action models trained on vast data sets of human movement, allowing robots to learn natural motion patterns rather than relying on manually programmed movements. 6:32 The money flowing into humanoid robotics is staggering. Figure AI, a startup few people had heard of three years ago, just raised over one billion dollars at a thirty-nine billion dollar valuation. 6:44 The investors were valuing a robotics company at the price of a major car manufacturer. They're already building a factory that can produce twelve thousand robots per year, with plans to scale to a hundred thousand. 6:56 Tesla's Optimus has gone from a promotional stunt in twenty twenty-one, literally a person in a robot suit, to functional prototypes doing yoga poses, sorting objects, and playing board games. 7:08 Elon Musk claims they'll cost under thirty thousand dollars, and there'll be a home-ready version arriving by twenty twenty-nine. 7:15 Whether Tesla can deliver on that timeline and price point is another matter, but they're leveraging everything they've learned from self-driving cars and putting it into a humanoid body. 7:26 Then there's 1X Technologies, a Norwegian company that's done something nobody else has dared. They've opened pre-orders for consumer purchase. Starting in twenty twenty-six, people can buy a Neo robot for their homes. 7:38 The first units come with restrictions, no hot objects, no sharp items, no homes with small children, and use a hybrid approach where human operators can take control remotely for unfamiliar tasks. 7:52 But it's a genuine consumer product, not a pilot program. When companies are building factories before they have customers, it's not speculation, it's a bet that a market is coming. But there's a big question. 8:05 Why are these robots humanoid at all? The standard argument is that our homes are designed for human bodies. Stairs, door handles, kitchen counters, light switches, everything assumes a roughly human-shaped operator. 8:21 A humanoid robot can navigate spaces built for us without requiring any modifications. But Brooks is skeptical of this reasoning. 8:29 He predicts that the successful domestic robots of the future will look nothing like humans. 8:34 They'll have wheels instead of legs, multiple arms, and possibly five-fingered hands, though they might still be called humanoid for marketing purposes. 8:44 The human form, after all, isn't optimized for folding laundry or loading dishwashers. It's optimized for bipedal locomotion across varied terrain, for throwing objects, for social signaling. 8:57 A purpose-built domestic robot might be better served by a completely different design. Dyson seems to agree. 9:05 The vacuum company, which knows something about domestic robots, is investing heavily in robotic arms rather than humanoid bodies. 9:12 Their approach focuses on specific tasks like washing dishes and handling kitchen work, rather than trying to build a general purpose human replica. Amazon's Astro takes yet another approach. 9:24 It's essentially a tablet on wheels. It looks nothing like a human, but it's teaching Amazon crucial lessons about how robots navigate real homes with all their mess and unpredictability. 9:36 There may be psychological reasons for the humanoid form. We might find it easier to interact with something that looks like us, or more willing to trust it in our homes. 9:46 But it's worth questioning whether the current humanoid fixation is driven by engineering logic or by the sci-fi imagery that shaped our expectations of what robots should look like. 9:57 Let's be honest about what's actually happening versus what's being demonstrated. Most autonomous robots in twenty twenty-five aren't fully autonomous. 10:07 Those impressive demonstrations often involve teleoperation, human operators controlling the robots remotely. Even 1X's Neo, which is furthest along in consumer deployment, relies on human operators for unfamiliar tasks. 10:22 The autonomy is partial, not complete. The testing happening in homes is heavily controlled. 10:29 These aren't robots dealing with the chaos of real family life, kids leaving Lego everywhere, pets knocking things over, that drawer everyone has full of random junk. 10:39 Current robots work best in structured, predictable environments. Your messy, lived-in home is still a challenge. Ken Goldberg, a robotics professor at Berkeley, offers important perspective. 10:52 Humanoid robots are still far behind the hype of AI chatbots, given the complexity of the real world. 10:59 Moving and manipulating objects in unstructured environments is exponentially harder than processing digital information. The physical world has a way of humbling even the most sophisticated AI. And then there's cost. 11:13 While Musk promises twenty to thirty thousand dollar robots, current prototypes likely cost six figures to build. 11:20 Mass production will bring prices down, but the early adopters paying for Neo robots in twenty twenty-six are essentially funding R&D for everyone else. 11:30 According to research from the World Economic Forum, robots could handle about thirty-nine percent of household chores within the next decade. 11:38 The tasks most ready for automation are repetitive, time-consuming chores that don't require complex decision-making, like loading dishes, folding laundry, and tidying up. 11:50 But the real driver of adoption might not be general cleaning. It might be elderly care. Globally, we're facing a massive caregiver shortage. 11:59 Countries like Japan simply don't have enough young people to care for their aging population. 12:04 By twenty thirty, about twenty-five percent of elderly individuals living alone could benefit from robot-assisted care.Stanford's Professor Alison Okamura is developing soft robots specifically for caregiving, machines with compliant arms that can help someone out of bed without bruising them. 12:23 As she notes, many older adults would actually prefer a robot's help with intimate tasks like bathing rather than depending on a human caregiver. 12:31 There's something about the impersonality of a machine that can paradoxically feel less invasive. The robots don't need to be perfect. 12:39 They just need to be good enough to help someone live independently for a few more years, help carrying heavy items, reminders to take medication, assistance getting up from chairs. 12:50 These simple aids could be transformative. Okamura thinks specialized helpers like these will arrive in five to 10 years. The general purpose robot butler, that might not come until the 2040s. 13:03 Brooks thinks today's humanoid efforts will be long gone and most conveniently forgotten by then. But before we welcome robot roommates, we need to confront some uncomfortable questions. Privacy is the obvious one. 13:18 These robots need cameras to navigate, microphones to understand commands, and AI to process it all. One X's CEO, Bernd Bornick, says it's absolutely unforgivable to have any breach of privacy in these systems. 13:32 But can we trust that promise when the robot's AI is learning from everything it sees in your home? We've been here before with smart speakers and smartphones, and the track record isn't exactly reassuring. 13:44 There's the employment question too. If robots can do 39% of household chores, what happens to the millions of people employed in domestic work, household cleaners, caregivers, maintenance workers? 13:56 These jobs might not disappear overnight, but they'll certainly transform. And there's the psychological dimension. What happens to human relationships when we can outsource care to machines? 14:08 Does having a robot companion reduce loneliness or increase isolation? When children grow up with robot helpers, how does that shape their understanding of work and responsibility? 14:20 But let me just return to the beginning of this video for a moment. 14:23 As I said, a lot of people may just think the whole thing is weird and uncanny, and this psychological bridge may take a lot longer to cross than is expected. And that's if we can afford it. 14:35 If a robot costs as much as a car but offers less utility, then there may not be much of a market after all. And let's just consider when something goes wrong and a domestic robot ends up killing someone. 14:48 Somewhere, somehow, this will eventually happen, and a lot of people will see it as a risky, if not fearful, investment. William Gibson once wrote, "The future is already here. It's just not evenly distributed." 15:02 That's exactly where we are with home robots. In research labs, robots are folding laundry and making coffee. In a handful of homes, Neo robots are learning from their human families. 15:14 In millions of homes, robot vacuums are quietly proving that we can live alongside autonomous machines, even as the company that pioneered them files for bankruptcy. 15:25 The truth is probably somewhere between Levine's optimism and Brooks' skepticism. We're unlikely to have humanoid butlers running our households by 2030, but we're also unlikely to be waiting until 2050. 15:38 The more probable future involves specialized robots that do specific tasks well, folding laundry, assisting elderly relatives, maintaining cleanliness, rather than general purpose humanoids that do everything. 15:51 And those robots might not look anything like us. They might have wheels instead of legs, multiple arms instead of two, forms optimized for function rather than familiarity. 16:02 C-3PO might arrive, but he might arrive on wheels. I'm cautiously optimistic. There are real challenges and legitimate reasons for skepticism. 16:13 But imagine a world where nobody has to choose between career and caregiving, where the drudgery of daily chores doesn't eat up our precious free time. 16:21 And the biggest one for me, where elderly people can maintain independence for longer. That's not a dystopia. 16:28 That's a world where humans can focus on being human, creating, connecting, and living, while machines handle the repetitive tasks we never enjoyed anyway. 16:38 The future of domestic robots isn't about replacing human connection. It's about freeing us to have more of it. So what do you think? Are you ready to welcome a robot into your home? 16:49 Or like me, do you find the whole thing just a bit weird? Mm-hmm. You've been watching Absolutely Agential. 17:04 We're a media consultancy and startup focused on AI, and we create weekly YouTube videos and a newsletter we send a few times a week. 17:12 In a world of AI hype, we're trying to strike a balance, looking at a range of opinions to cut through an increasingly complex world. 17:19 The best thing you can do to support us is to sign up to the newsletter or subscribe to the channel. But also, don't forget to like and leave us a friendly comment. See you next time. It'll be Absolutely Agential.