> allowing the entire fleet to upload terabytes of data for continuous learning and improvement
Ugh.
Edit: Yes, I meant I, Robot the film. U.S. Robotics and the like.
You can wash the dishes and tidy up after every meal, rinse and sort your recycling but you're still trashing the planet more or less the same as the person who does none of those things.
It's us, flesh blobs. Long after we cover everything in AI and robots around us, we will not change easily. Societal drift is slow, genetic drift is slower.
(For the record: Gimme my robot, but interesting thought nonetheless)
The thing to watch out for is: deployments. How many units are they pushing and to who. What kind of tasks can those robots accomplish well enough to warrant actually using them. How hard is it to adapt those robots to deployments. How that changes over time.
The hardest problem of creating a universal robot is, and always has been, AI. If Figure can deliver sharp, highly adaptive, easy to use AI? High generalization, good performance on a diverse range of tasks and in many environments out of the box? Then they have a killer product.
And a proxy to track that is reports of how many robots they deploy and to who. If they start shipping to small companies and deploying to high uncertainty spatially complex fields like construction or maintenance? If you start seeing robots unloading trucks and restocking shelves at a small town Walmart, unannounced? Big.
That tells me that the design is amenable to aftermarket service and maintenance and that the machines are capable of participating in relatively sophisticated manufacturing processes.
A graph with one line representing the number of hours of physical labour by humans per unit produced with another line representing number of hours of physical 'labour' by these robots per unit produced would be interesting to look at.
The intersection point between those two lines and the point where human input drops to zero are key points in humanity.
This is silly. Wireless charging is inefficient and costly compared to cables but we use it for the convenience of humans, to avoid the annoyance of having to plug something in repeatedly. Obviously a humanoid robot should simply plug in its own cable! No human need be inconvenienced. Wireless charging has no benefit here at all.
> Each fingertip sensor can detect forces as small as three grams of pressure - sensitive enough to register the weight of a paperclip resting on your finger
Three grams would be a very heavy paperclip. I have seen several types of touch sensor and while the technology is impressive I don't think any of them are durable enough for real use. Even human skin doesn't rely on durability alone. Healing is critical. But healing is infeasible for robots so instead we need to design repairable, replaceable, disposable, ideally recyclable parts, especially for the fingers that touch everything. This hand looks monolithic and not repairable.
All that said, I'm looking forward to seeing if their claims about cost and manufacturing volume pan out. Those are the things that matter the most right now, along with reliability. We need large numbers of robots operating continuously in the world to collect the data that will enable us to train robot AI. Right now there's basically only one or two companies with scaled humanoid production (for a very loose definition of "scaled") and they are in China. I'm rooting for anyone who can manufacture robots outside of China.
Commoditization and Walmart-level deployments at scale are still a few gens off.
Why is it so important to you that people fold their own clothes and wash their own dishes?
Why do you idolize a life of increased drudgery?
Then the robot would just go to its station and swap its own batteries. Why even have wireless charging at all? Or even a cable? Or even have it "charge"? Battery swapping seems to make way more sense here. Am I missing something?
Bonus points if the robot has data on the degradation and can order its own replacement batteries, take them out of the box, and ship the old ones to a recycling facility...
More bonus points if the charging station is actually outside under a 1KW solar array pergola thing, that way you don't even have to pay for the electricity either. Don't worry, the robot will lock the door when it goes out to grab its batteries. It'll also bring in the whole setup if the weather isn't great.
[1] https://www.ftc.gov/news-events/news/press-releases/2023/05/...
But a cable is a fair question.. you'd think it could plug itself in...
Maybe that's a hint at the robots actual capabilities at this point... or, they didn't want to bet on the unpredictability of environments: what if there's something in the way of the cable, though something could also be in the way of the inductive charger
The hotel reception use case seems ridiculous though, if you get rid of a human receptionist, you lose the human element of the check in process, which people like. If you're getting rid of the human and losing all the benefits of that, then just replace it with a kiosk (or mobile check in), which will do a far better job than a robot.
On the robot side, there are many things that have to go right. Hardware needs to become good enough, reliable enough and cheap enough to scale. Then you have the software stack on top that needs to scale in training, fine-tuning, control and generalisation. None of these are "easy" even in a lab setting. Doing it at scale, in production will be huge. And then there's data collection, where whoever does it better will probably win. Collecting data in peoples houses is problematic, but on the factory floor should be ok.
ATM my bet is on Tesla being the best positioned to best deliver (eventually). They have plenty of experience on all fronts, and more importantly they have ample places to test them. Their factories are as automated as possible, so it's safe to say that every human being still doing manual labor is critical in their role. As soon as they can replace some of them with humanoids, and see the "task success" number go up, they can scale it up all over their floors. And we know they can scale.
I used to think that generalist humanoid robots are still 10y out, due to hardware and generalist software stacks, but it seems like things are heating up. It's gonna be an interesting next decade.
I have no idea about the maturity of this company in particular, but it's interesting that glossy robotics startups never lean in on that as a core user base.
Equal parts terror, awe, fear, when it comes to having a robot in my home.
On the other hand, would the removal of these inconveniences allow for the highest calling of humanity - I argue creativity - to flourish to the fullest? My gut reaction is once again that inconveniences are actually a very important resistance to creativity, like how you need gritty sand paper to create smooth wood.
You can buy an expensive robot, or maybe you can meditate and be mindful that inconveniences play an important role in the meaning of your life. I am of course speaking of the household use here - I think the debate is likely different for a business setting.
The key difference between now and then isn't smaller actuators, cheaper sensors or denser power electronics. It's the AI breakthroughs.
Doesn't need to be "at scale". Scale is a useful proxy though. But if you see two robots deployed to your average Walmart, and doing a good enough job there to cut the staff in half?
Doesn't matter that it's just two robots at a few Walmarts. Making more robots isn't that hard. The scale would inevitably follow.
I mean standing there for 10 minutes and giving them my passport to give me a plastic card with a digital code has very little to do with human touch.
I want that human touch at a bar perhaps but not at a reception.
If your critisism is only about the reception part: There has to be a transition part and a 'let a human do it for a bit' or 'here is a complicated case please robot move aside i'm here'.
If my cleaner was a robot, I'm sure I'd eventually lose that sense of embarrassment. I'm usually polite with ChatGPT but I think that's also passing...
There are practical advantages to being able to charge wirelessly, sure. But if they're doing that because of AI limitations? Bad sign.
But only for me because I have the feeling i lived out my normal environment and i'm not rich enough yet to expand so I can become busy again in a more meaningful way. Specifically having a big house/workshop to do things in my future workshop.
Besides, servants are nothing new. They're rare in the US but common in some other countries, and the people who grow up with them are maybe somewhat different but not radically changed IMO.
Yes, like gasoline. But still batteries. Maybe some kind of bearing sized batteries which can be poured like a fluid?
> Bonus points if the robot has data on the degradation BMS that can tell battery health is common so this should be there.
I think we have A LOT of old people in precare situations because they are not aware of the possible difference this can make.
If robots can fight entropy for us, all the better.
There is not a magic portal opening up when you are able to optimize ever aspect away of living and you will gain access to enlightment and everything is different.
And don't get me wrong, I have no clue how our society would look like if everything is done by ai and robots because we as a society don't talk about it and don't give everyone the resources they want or need if they have suddenly no 'drugdgery' anymore.
Give me a million today and i will spend the next 10 years rebuiling an old castle and I will have A LOT of fun doing this. Let me check, my bank balance is not at one million.
Instead i have to pay for a lot of things and then I have to work for 40 hours. Suddenly i'm great at my job, get valued but this is just Drudgery even if its complex work. Its work for someone else which doesn't matter to me.
> Helix: Figure 03 features a completely redesigned sensory suite and hand system which is purpose-built to enable Helix - Figure's proprietary vision-language-action AI.
Also, the charge rate matters. If robot can charge to 80% in say 30 minutes, then it can take small charging breaks during the day between critical tasks.
Also, if the feet have inductive chargers, it's possible to place the robot on a large charging mat that allows it to run indefinitely, like in a factory environment. If your robot takes 30 minutes to fold the laundry or do dishes, why not place a charging mat at these locations so it can work and charge at the same time.
In the future, new homes might include charging coils embedded in the floor every 12 inches so that your robots can work all day.
One reason that caught my attention was how she described the behavior of these people, who have the world at their fingertips, who have never really known hardship, and in turn have full blown meltdowns about the most trivial annoyances. What car will we drive on our trip?! The salmon cracker appetizers are too salty to be served! They stocked the wrong oat milk in the mini-fridge!
Almost like the need to get upset over inconveniences is ingrained, and when there is a lack of real ones, your brain just latches onto whatever it can to let the "freakout" out.
I agree on the data part. I love the potential idea of a humanoid robot at home to take care of chores, but now it seems like the potential for it not being constantly connected and collecting data is gone out the window.
I find it quite strange that they are openly bragging about how much data it will be gathering and uploading from within your home. That feels like the part you would not say out loud.
https://en.wikipedia.org/wiki/Fuel_cell
Well established and even commercialized (Toyota sells fuel cell cars today IIRC), just not as cost effective in cars from a full infrastructure perspective (fueling specifically).
Yet, it's being sold as capable of doing and folding your laundry.
I would sell th stock to the next idiot the moment they announced this.
https://www.malwarebytes.com/blog/podcast/2023/01/a-private-...
If we make the robots humanoid, they get compatibility with human systems for free
The modem[1] folks? :)
They've shown the "putting dishes in the dishwasher" bit before, it seems to be getting better, but I imagine it still has a high failure rate.
I wonder if this company started off or has some founder that's really interested in the "handling deformable stuff" space. They really seem keen to promote that it can do tasks like folding a shirt or working with soft packages.
Definitely seems like a carefully curated video, but the longer videos make me think that either they are running a scam or they have some of this stuff working well enough.
Honestly a game changer. Sounds stupid, but there's just something very satisfying about being able to quickly fold a bunch of clothes and get very nice results.
And if we get humanoid robots at some point, they can use them too.
All factors of "it was Vegas" aside, one of the things that stood out to me was that the hotels have moved rapidly to rapid checkin/checkout systems where you punch in your confirmation code or name/dob and present a photo ID of some kind (passports can just be slapped against the reader) and it asks a few questions ("do you need late checkout", etc), directs you to the exact place your room is (and prints it, which was nice) and tells you where the bellhop station is if there's more than a little while before your room is ready and it can't dispense your cards.
All told, four of these stations had roughly 90% the throughput of the four real humans, but they "moved faster" because it didn't feel like queuing for a human, more... "waiting for a toilet"?
Kiosk based stuff is great until it fails. Spend an hour in the checkin area of a major airport and you'll see any number of interesting failure modes.
As for the washing machine bit: Why not push for more standards usage in home automation? We have Thread, which is really cool, and which is driving the home automation future that we're slowly getting. Once it's loaded, a homebot should't have to check the thing manually, it should get information about when, what, and how and be able to have "eyes in the back of its head" so to speak.
30 years ago we figured out how to contact charge cordless phones with metal pads and prongs.
The current best neural networks only have around 60% success rates for small horizon tasks (think 10-20 seconds e.g. pick up apple). That is why there is so much cut-motions in this video. The future will be awesome but it will take time a lot of research still needs to happen (e.g. robust hands, tactile, how to even collect large scale data, RL).
Probably about 1% of the cost of the humans though...
>Kiosk based stuff is great until it fails. Spend an hour in the checkin area of a major airport and you'll see any number of interesting failure modes.
A robot would be less reliable than a kiosk, so if you're going to have some kind of machine replace the human, you might as well have a kiosk.
The ideal model (IMO) is a hybrid model, where you have lots of kiosks for the 90% of cases where there are no issues, and a few humans on standby to drop in and assist people who are having issues.
Or better yet, do away with the check in desk, and let people check in on their phone (some hotels already do this, and you tap your phone on the door to unlock)
"Building Figure won’t be an easy win; it will require decades of commitment and ingenuity."
"Our focus is on what we can achieve 5, 10, 20+ years from now, not the near-term wins."
At least it's not Musk's forever "next year".
From there, it's a question of could they bring costs down.
In the 2010s everyone purchased those rumba vacuums, because whatever, they're cheap. Now I usually see them collecting dust.
The strong use-case for robotics is industrial/manufacturing and construction, agriculture probably more than ever. They don't need to be humanoid at all, and in fact maybe they shouldn't be because that very feature could spook unions and labor groups. Robots that actually look like they're "just tools" will be more willingly embraced.
Indeed, all the videos/examples are marketing pieces.
I would love to see a video like this "Logistics"[0] one, that shows this new iteration doing some household tasks. There is no way that it's not clunky and prone to all kinds of accidents and failures. Not that it's a bad thing - it would simply be nice to see.
Maybe they will do another video? Would love that.
It is breaking news if there is a $40K robot that had a 12:1 efficiency ratio.*
Because production is so nascent and small, cost doesn't mean too much, no ones scaled yet.
At only $40K capital investment, even a guaranteed 12:1 efficiency ratio would be an absolute no-brainer financially for many, many, wealthy people and certainly businesses. I do 1-2 hours of chores a day if I'm lucky. If I had the equivalent of a robot vacuum working 24/7 it'd do a much better job than me.
* The whole thing is written up and shown in a way that makes you think we're on the second refining release of a breakthrough**. I don't think they've gotten to the breakthrough yet - we would have seen > 0 videos from outside the company by v3.
** Really, the whole thing has an audience of one: Musk. (c.f. focus on fingers which was recently reported as the major pain point for whatever he calls their robot not making it to production; aping of Musk-y things like the factory itself a product)
The problem with the principled approach to high-uncertainty projects is that if you slowly execute on a sequential multi-year plan, you will almost certainly find out in year 9 that multiple of the late-stage tasks are much harder than you thought.
You just don't know ahead of the time. Just look at how many corporations and research labs had decades-long strategies to build human-like AI that went nowhere. And then some guys came up with a novel architecture and all of sudden, you can ask your computer to write an essay about penguins.
Musk's approach is that if you have an infinite supply of fresh grads who really believe in you and are willing to work crazy hours, giving them a "next year" deadline is more likely to give you what you want than telling them "here's your slow-paced project you're gonna be working on for the next decade". And I guess he thinks to himself that some of them are going to burn out, but it's a sacrifice he's willing to make.
We are nowhere near the same for autonomous robots, and it's not even funny. To continue to use the internet as an analogy for LLMs, we are pre-DARPANET, pre-ASCII, pre-transistor. We don't even have the sensors that would make safe household humanoid robots possible. Any theater from robot companies about trying to train a neural net based on motion capture is laughably foolish. At the current rate of progress, we are more than decades away.
To answer your question -- folding clothes is easy, because clothes easily deform, do not break, fall smoothly when you drop them and most importantly are easily resettable task. Just through the well folded cloth up and voila start again.
Neural networks for motion control is very clearly resulting in some incredible capability in a relatively short amount of time vs. the more traditional control hierarchies used in something like Boston Dynamics. Look at Unitree's G1
https://www.youtube.com/shorts/mP3Exb1YC8o
https://www.youtube.com/watch?v=bPSLMX_V38E
It's like an agile idiot, very physically capable but no purpose.
The next domain is going to be incorporating goals and intent and short/long term chains of causality into the model, and for that it seems we're presently missing quite a bit usable training data. That will clearly evolve over time, as will the fidelity of simulations that can be used to train the model and the learned experience of deployed robots.
- novel idea or technology
- counterintuitive effect of technology
I think the second is easier written as "what if Good Thing was actually Bad". So that's what you get. The former style is perhaps still available in books like Children of Time by Adrian Tchaikovsky.
But the latter style is much more readily written and consequently has dominated sci fi as more authors enter the field.
The Torment Nexus view is mostly driven by context blindness. "oh my god, they'll scan the mother's blood to perform eugenics if they have sequencing technology and it will be horrible". Well, advanced societies do that a lot: Down's is scanned for using a Maternal Serum Alpha Foetoprotein test. "oh my god, they'll use ultrasounds to find undesirable genetics, torment Nexus" but Nuchal Translucency tests are fairly routine in advanced societies and we're fine with them.
This might appear like a fixation on dystopian literature to others. "omg gattaca this MSAFP". It's just generic technoluddism because almost all near future tech is explored via sci fi in the "what if Good is Bad" genre.
As someone who worked in the robotics industry, 90% of the demos and videos are cherry-picked, or even blatantly fake. That's why for any new robot in the market, my criteria is: Can I buy it? If it's affordable and the consumer can buy it and find it useful in day to day life, then this robot is useful and has potential; other than that, it's just an investor money grab PR hype.
"Today, manual labor compensation is the primary driver of goods and services prices, accounting for ~50% of global GDP (~$42 trillion/yr), but as these robots “join the workforce,” everywhere from factories to farmland, the cost of labor will decrease until it becomes equivalent to the price of renting a robot, facilitating a long-term, holistic reduction in costs.”
Renting a robot? What are the chances of robot rent-seeking becoming a drop in replacement for today's increasing costs of labor. A high likelihood potential outcome.
I'm not sure that task needs a humanoid robot, but the ability to grab and manipulate all those packages and recover from failures is pretty good
The video shows several of glitches. From the comments:
14:18 the Fall
28:40 the Fall 2
41:23 the Fall 3
But then I think lots of this can be solved in software and having seen how LLMs have advanced in the last few years, I'd not be surprised to see these robots useful in 5 years.
People being super negative about this is a bit surprising to me.
This feels incredibly generous. I'm pretty sure his approach is that he needs to keep the hype cycle going for as long as possible. I also believe it's partially his willingness to believe his own bullshit.
If they make a hundred of these, it'll be impressive. If they make a thousand it'll be scary.
To train GPT, all of the training data (the internet of text, scanned books, etc) had already existed, even before the GPT project began. Arguably, the compute required (for GPT-3) also already existed, even before GPT-2.
The GPT project really just came down to investing in all of the pieces to take the ideas from a 2017 research paper to the next level. Nobody knew if X thousand GPUs, plus all of the internet's text, plus neural network transformers, would work out. But somebody took a risk in putting together the existing pieces, and proved that it can.
There's no analogy here to humanoid robotics. Not only is the data required for neural network operated humanoids close to non-existent (at the scale needed), but the nature of the data itself is enormously more complicated that taking a list of tokens in a vocabulary, and outputting 1 more token from the same vocabulary.
That being said, I still applaud the ambition of the Figure team. While I think it's clear they are presenting incredibly cherry-picked examples, they aren't trying to mislead consumers with a product for sale (because... they can't). Instead, they are productizing important research to investors, who would otherwise waste money on less important and less ambitious projects. So overall I find projects of this nature to be a net positive for technical innovation.
Also I'm not convinced that humanoids are an effective realization of robotics. The human shape and functions evolved from the biological pros and cons of our chemistry. That of robots is very different, so you'd expect similarly different shapes and functions.
I’m sure they could pretty easily spin up a site with 200 of these processing packages of most sizes (they have a limited number of standardized package sizes) nonstop. Remove ones that it gets right 99.99% of the time and keep training on the more difficult ones, the move to individual items.
Caveat: I have no idea what I’m talking about.
https://openreview.net/forum?id=3RSLW9YSgk
https://www.nature.com/articles/s41586-025-08744-2
It seems like learning from the environment will be a requirement for robots to scale. My understanding is that research has been yielding new architectures that might have that type of real-time, general intelligence but we haven't seen that similarly large investment yet.
There is empty space in the feet anyway for a coil and a wire..
The frog boils quickly.
So one more app to install that I'm sure would be a privacy nightmare.
Turns out they're either insanely expensive or they just can't actually learn on the fly and do tasks. This is the Nth time I've seen a robot folding a shirt but never in a cluttered room or taken from a pile of laundry.
I figured the first AI robots would be pets, but apparently they're aren't even that good yet. Furby level isn't going to cut it.
I think that's a massive leap. Suburban families get more utility out of a vehicle; they drive everywhere. Housekeeping is effectively just quasi-automated washing (dishwasher/laundry), occasional vacuum and clean, and food prep that is already available as a service for those who don't want to do it (for a relatively affordable subscription), and otherwise it's possible to prepare something in little time. I just don't see how average people would jump at spending 30k for that. The key reason people feel time-poor is juggling work and parenting. Unlike a nanny, you won't offload parenting to a housekeeping robot. For our part, we involve the kids in routine chore activity. At an early age they often learn by mimicking actions and are enthusiastic about helping.
Since you have to sit around watching them anyway, might as well be productive time.
But yeah, I'd happily just check in at a kiosk and get my room card that way. (And I'm sure phone-as-key, no-contact check-in is only going to get more common)
Two kinds of machinery are needed.
One very basic and cheap robot to go around the neighborhood and gather clothes in some boxes, and transfer them in a designated room to wash them.
Two state of the art robotic hands mounted on the wall, and connected to AC (no batteries). The two arms are going to be controlled by computers even a whole rack of them, with many GPUs in them. The whole setup might use 10KW of energy, it will wash clothes by hand, it will be fast, dexterous and accurate. Expensive as well. In 3 minutes it will wash 100 t-shirts much better than the best human on the planet, or any other non-intelligent machine.
Then the small basic robot returns the clothes to the house.
Same with cooking. Same with many other things.
>This is the Nth time I've seen a robot folding a shirt
We're watching robots intelligently find a shirt, figure out how to fold it relative to its position, and then parse all that data, tokenizing both vision + text instructions into actionable movements that actually result in the physical world being affected!
All this, and people are criticizing it's manufacturing cost or ability to do things it hasn't been explicitly trained to do in 2025. I see these things and don't think about 2025, I'm thinking about 2040 and the inevitable future we're diving into.
Yeah I'll pass thanks.
Boston Dynamics has been releasing actual product demos of such robots (not cherry-picked ads) for ~20 years now. Not a single one has graduated to any mass market real world use case.
I'm not saying one shouldn't be hopeful, but it's also not hard to see why people here are generally more conservative about the near future.
https://rodneybrooks.com/why-todays-humanoids-wont-learn-dex...
In short, he makes the case that unlike text and images, human dexterity is based on sensory inputs that we barely understand, that these robots don't have, and it will take a long time to get the right sensors in, get the right data recorded, and only then train them to the level of a human. He is very skeptical that they can learn from video-only data, which is what the companies are doing.
AC wires, better not move around, especially when there is water. It has to be mounted on the wall.
>Boston Dynamics has been releasing actual product demos of such robots (not cherry-picked ads) for ~20 years now. Not a single one has graduated to any mass market real world use case.
I'm sure a lot of BD's initial R&D has made Figure able to ramp so quickly and I don't mean to speak negatively of BD at all, but within 3 years, Figure has made it feel like the future is at our doorstep, meanwhile BD hasn't really done that for me in 3 decades. That's very impressive to me.
They for sure did not anticipate that the user would backflip into their robot and knock it (and himself) out :D
I would pay extra to avoid it - just let me download a pass like a boarding pass to my Apple Wallet as I walk through the front door and head directly to my room.
If you can make it look believable on camera for 15 seconds under controlled studio conditions... it's probable you can do it autonomously in 10-15 years. I don't think anyone is going to be casually buying these for their house by this time next year, but it certainly demonstrates what is realistically possible.
If they can provably make these things safe, it will have huge implications for in home care in advanced age, where instead of living in an assisted living home at $huge expense for 20+ years, you might be able to live on your own for most of that time.
I am cautiously optimistic.
The oldest video on their YouTube channel is 16 years old, and is of a quadrupedal robot not falling over while inching along tricky surfaces.
Humanoid robots have fallen into the latter category for too long for most people to jump at each advancement being "the one" anymore. Afterwards, everyone will agree it was obvious ${ADVANCEMENT} was really the one which would do it - but not before.
Boston Dynamics hasn't released any actual products. They seem to be focused on flashy demos of robots dancing instead of end user products.
As a counterpoint, Unitree right now sells humanoids you can actually buy. They're no where near as good, but you can actually use them.
Would asking the robot for a seahorse emoji leave you in a puddle of blood?
Right, although "servants" conjures up rich people with full time staff.
A better comparison to the humanoid robot some people here are dreaming of to do their household chores is a country like India where it's common for middle class people to hire multiple different people to come do chores, daily or weekly, such as cooking, laundry, cleaning, yardwork, etc. These are cheap services.
In the US, probably most people here on YC News (higher paid tech workers?) could afford to have lawn mowing service, weekly maid service, laundry pick-up/drop-off service (or bring to laundry yourself), and either eat out all the time, or UberEats etc. It's not clear that having a robot to do these tasks would be cheaper or preferable.
This week, I had my first experience with exactly this at a car hire company. It was… not smooth.
It took multiple attempts (with requests for help to the employees in between) to get the system to recognise our code, whereupon we learned (by way of an unhelpful generic error message) that the system had somehow given someone else ‘our’ car. After another round of asking for human help, we had to wait while someone came outside, unlocked the machine, and put the keys for our new car inside. We then went through the code process again, and were finally given the keys.
The vision is somewhere there, but the execution isn’t exactly the future we’re hoping for!
People warned about the dangers of social media (or with modern LLMs + Diffusion Models and scamming) and that's kinda come true, but people also warned about the dangers of IVF and that's just been good. So what happens is that people always warn about the dangers. Humans are loss-averse so they find it easy to do that.
It is unsurprising that every new tech seems like dystopian literature because there's a lot of dystopian literature focused on the near future and we're good at coming up with negative hypotheses. There is no significance in it.
An illustrative example is a warehouse. They're still partly designed for humans because they're not fully automated, but the need to make them human-friendly will disappear soon.
We're reaching a point in these advances where EVERYONE is starting to worry about their job security.
And the society we live in - at least in the US - has no safety net ready for a mass replacement of workers with technology like this.
Oh AI took your software job? Good news you can still go into construction - oh wait, that's not an option either. How about housecleaning? Oh no, that's been automated too. Well sorry, we have no safety net for you. Good luck in the streets.
It will be interesting to see how messed up our political leadership lets things get before they slightly tweak their belief system.
What surprises me is that people can't see this for what it is: early steps.
As an aside, the first steam engine was created by the Greeks 2000 years ago, but they just used it for toys. When Watt created the modern steam engine it had 0.7 horsepower. He actually invented the term horsepower as a marketing term because his engine was so underpowered. An actual horse produces 14 horsepower, but he adjusted for how long a horse needs to sleep and and rest and came up with 1 horsepower. His first production engine had 5 horsepower.
If hackers had existed in the 18th century I would have expected them to see the promise of engines replacing horses even though engines were less powerful at the time, not say "engines are basically smoke and mirrors right now".
It could be that I'm wrong, and we're at the Greek stage, where humanoids are at a false start, just toys, and the real thing is 2000 years away. But the lack of optimism surprises me.
An industrial robot arm with air powered suction cups would do the trick... https://bostondynamics.com/products/stretch/ ...
... So the task they work best at is the task there is already cheaper better robots specialized for.
Tasks left for human "sorters" to do are:
- put packages on conveyor belt so the scanner can read the label (as done by the robot in the video)
- deal with damaged or unreadable packages that can't be processed automatically
- when a package gets jammed and forces the conveyor belt to stop, remove the offending package before restarting
- receive packages at the other end and load them into vehicles
Generally the difficulty with all of these is dealing with variability and humans act as variability absorbers so the machines can operate smoothly.
Like LLMs being used to pick values out of JSON objects when jq would do the job 1000x more efficiently.
This is what this whole field feels like right now. Let's spend lots of time and energy to create a humanoid robot to do the things humans already decided humans were inefficient at and solved with specialised tools.
Like people saying "oh it can wash my dishes for me". Well, I haven't washed dishes in years, there's a thing called a dishwasher which does one thing and does it well.
"Oh it can do the vacuuming". We have robot vacuums which already do that.
BD was a money-burning machine that suckled off the teat of the miltiary industrial complex, where billions of dollars can be casually lost and there's no accountability and no one notices its gone. Their tech was cool, though, and their engineers did awesome work.
Simulation isn't sufficient for ML in robotics -- and they simply don't have enough training data.
EV charger style of short, thick cables should not be THAT hard, though. The more likely problem here is that they just can't handle the task of securing and inserting the head of the cable against resistance.
I guess it's kind of natural because when you agree with stuff there's not much value saying "yes this is right" but if you disagree then you generally have a lot to say about why.
But also I think there are a lot of curmudgeons here. Back in my day we didn't need no stinkin robots and AI and toktik. And anyway I could make a way better robot, this is no big deal.
Small servers with console, PA speakers, field metrology or data acquisition machines, those things could have the lower torso or two for this and relocated as needed. The PhD guys can just park the truck and let those deploy wherever AI thinks >65% suitable for human use on their own, instead of users burning 15% of brain juice thinking and executing that. That would be immensely useful.
(also re: hotels that others are commenting, there were never technical reasons the door keycard readers couldn't ever had doubled as credit card readers - I think the reason why clerks are required is for sanity check, that the guests aren't in need of immediate safety/health assistance and ok to proceed to beds)
An obvious application, if this robot could do it, is retail store shelf restocking. That's a reasonably constrained pick and place task, some mobility is necessary, and the humanoid form is appropriate working in aisles and shelves spaced for humans. How close is that?
It's been tried before. In 2020.[1] And again in 2022.[2] That one runs on a track, is closer to an traditional industrial robot, and is used by 7-11 Japan.
Robots that just cruise around stores and inspect the shelves visually are in moderately wide use. They just compare the shelf images with the planogram; they don't handle the merchandise. So there are already systems to help plan the restocking task.
Technical University Delft says their group should be able to do this in five years.[3] (From when? No date on press release.)
[1] https://www.youtube.com/watch?v=cHgdW1HYLbM
[2] https://blogs.nvidia.com/blog/telexistence-convenience-store...
[3] https://www.tudelft.nl/en/stories/articles/shelf-stocking-ro...
This is a major flaw in social media/forums that distorts public opinion so very much.
Surely there's a market for an appliance that can load, wash, dry, fold where the only human interaction is dumping the clothes and soap within arm's reach.
Regardless, if an employee posted images acquired from customers, testing users, or anyone else to their personal social media platform of choice, they are still assholes. And the company that allowed for that to happen is an asshole as well.
Perhaps this is a bit pedantic, but what about the probable eventual proliferation of useful humanoid robots will make the future awesome? What does an awesome future look like compared to today, to you?
The fabric wrap is idiotic. Insanely stupid. Let's have an expensive fabric-covered robot wash dishes covered in food. Genius. It's a good thing those "dirty dishes" were already perfectly clean. I doubt this machine could handle anything more. Put it in a real commercial kitchen and have it scrape oven pans and I'll be impressed.
I'm so glad I left robotics. I don't want to have anything to do with this very silly bubble.
This only highlights the fact that making a cool prototype do a few cool things on video is far, far easier than making a commercial product that can consistently do these things reliably. It often takes decades to move from the former to the latter. And Figure hasn't even shown us particularly impressive things from its prototypes yet.
They also have Handle, a slow moving robot on wheels with an arm for moving boxes. No idea how many have been sold, but it seems to be even less than Spot.
With a hefty subscription to make it do anything useful.
I've always been partial to fuel cells, and in some ways they're ahead of the curve relative to standard batteries. For instance, solid electrolytes have been a thing for a while in fuel cells, and in both flavors of exchange. The challenge has always been overcoming sluggish kinetics with either better catalysts, or heat. It makes me wonder if there's a useful solid state battery that runs hotter than typical batteries, that would be useful for hybrid automotive applications.
They already have. We just don't hold the perpetrators accountable.
If I could just type it into my shell, that would be nice. I’m sure there’s some command (or one could be trivially made) to evaluate an equation, but then you get to play game with shell expansions and quotes.
In emacs I have to convolute the equation into prefix.
All minor stuff but it adds up.
It's hard to find decent general purpose help these days and they would pay good money for a halfway useful helper.
Once it's able to weld... That's going to be a massive game changer, and I can see that coming 'round the corner right quickly.
This is probably similar to what is driving negativity from other commenters too, although probably some are just concerned about an investment bubble.
Unless you’re trying to warn us to shut you down while there is still time. Blink if you need help.
bc
And the Unitree R1 already only costs $6k.
All the necessary pieces are aligning, very rapidly, and as James Burke has pointed out, that's when Connections happen.
Another idea: video extension model as a world model. We fine tune Sora on first person robot videos (and we train another model to predict actuation states from FPV). Then we extend the video using Sora “a robot in first person view finishes moving laundry from washer to dryer”. Then predict actuation states from the extended video?
So, my conclusion is that they’re doing it all right for this stage. Lots of glamour added to a bit of substance.
And if you needed it programmable, well an FPGA was still almost as general and far more efficient than a microprocessor.
Guess what won.
And for a time it was good.
I moved to CA a decade ago to join a robotics company.
I've since acquired a wife, house and dog. Wife loves to cook, and would love a 2nd dog that didn't choose me. I am a sucker for DIY. If I were in an apartment still, with no pets... i.e. lots less chores to do (hooray hybrid work!)... I'd be seriously considering roles at Figure, which is 100% in-office instead. (their office is a sub 10 minute walk from my last apartment)
How long that work satisfaction would last... very up for debate though!
Low level tools require an investment of time and brainpower to configure. Consider the time it takes to set up a dishwasher- research, buying, installing, reading the manual. Vs telling your humanoid robot “go wash the dishes”. People will pay a lot more and put up with a lot worse results in exchange for that kind of simplicity.
In a better world we would all be craftspeople and invest time into more efficient things but that ain’t human nature
Sure, if a dishwasher is $1k and the robot has a high success (not many broken dishes) rate AND can do other things AND is priced like a nice used car (up to $35k) then yeah, maybe? But there's so much of "it depends" in there that it's hard to say for sure. In curious what price/generality/reliability you have in mind when you say "many people would prefer..."
I’m not f##king the robot maid, I don’t care if it looks like a girl. If I was into that, there are other types of ‘robots’ for that.
in a world with 500 million humanoid robots, parts are plentiful, theyre easier to work on due to not weighing 5000 pounds, and like the other person said, economies of scale
All with much improved privacy, reliability, order of magnitude lower cost, no risk of robbery/SA, etc. 24/7 operation even on holidays. Imagine service staff just sitting waiting for you to need them, always and everywhere.
Nevermind how much human lifespan will be freed from the tyranny of these mindless jobs.
what it IS , however, is a remarkable achievement of commoditization; getting a toy like that with those kind of motors would have been prohibitively expensive anywhere else in the world; but much like the Chinese 20k EV, it's not really a reliable marker for the actual future; in fact bottomed out pricing is more-so an indicator of the phase of industrialization that country is in.
Only because it's not yet attached to a reasonable AI, which is my point. It's not going to do any heavy lifting, but it could easily do basic house chores like cleaning up, folding laundry, etc if it were. The actuators and body platform are there, and economies of scale already at work.
I guess some folks just can't or won't put 2 and 2 together to predict the near future.
It's Moore's law that largely drove what you describe.
Moore's law only applies to semiconductors.
Gears, motors and copper wire are not going to get 10x faster/cheaper every 18 months or whatever.
10 years from now gears will cost more, they will cost what they cost now plus inflation.
I've literally heard super smart YC founders say they just assume some sort of "Moore's law for hardware" will magicallyake their idea workable next year.
Computing power gets, and will continue to get, cheaper every day. Hardware, gears, nuts, bolts, doesnt.
An arm moving against gravity has a higher draw, the arc itself creates characteristics, a motion or force against the arm or fingers generates a change in draw -- a superintellligence would need only an ammeter to master proprioception, because human researchers can do this in a lab and they're nowhere near the bar of 'hypergenius superintelligence'.
I'm not surprised that a Honda Civic can't navigate the Dakar Rally route..
This is probably what people in the 18th century thought when they saw that digesting duck automaton [1]. Technological progress isn’t a magical linear thing that always leads to things getting better over time.
There’s some very hard problems to solve before the promises made here can be made true, and it’s not a given that they will or even can be solved. Building the robot was never the hard part.
People keep parroting this line, but it's not a given, especially for such an ill-defined metric as "better". If I ask an LLM how its day was, there's no one right answer. (Users anthropomorphizing the LLM is a given these days, no matter how you may feel about that.)
Also, now that I've typed that out, "sit down with ChatGPT and have a nice chat about." is a helluva thing to say.
They didn't go nowhere; they just didn't result in human-like AI. They gave us lots of breakthroughs, useful basic knowledge, and knowledge infrastructure that could be built off for related and unrelated projects. Plenty of shoot for the moon corporations didn't result in human-like AI either, but also probably did go nowhere, since they were focused on an all or nothing strategy. The ones that do succeed in a moonshot relied on those breakthroughs from decades-long research.
I'm not going to get into what Musk has been doing because I'm just not,
If you want to replace the human the best bet is to redesign the work so that it can be done with machine assistance, which is what we’ve been doing since the industrial revolution.
There’s a reason the motor car (which is the successful mass market personal transportation machine) doesn’t look anything like the horse that it replaced.
Some people are Honda Civic people, only concerned with utility - that's fine, I'm the same way. But the money comes from cars designed to evoke eroticism or animal aggression. The humanoid robot in the article is, aesthetically speaking, horrifying to most people. It doesn't even have a face, it doesn't look pleasant, it doesn't invite an emotional bond, it isn't friend shaped, and that isn't what most people will want, or would spend money on, regardless of how efficient it is.
Humanoid robots will have a context within the same gender and cultural dynamics as human beings, by virtue of looking and acting human enough. People already have relationships with AI, and that will only become more normalized over time. Most people will personify and anthropomorphize humanoid robots just as much as they do AI, and this will be necessary for their popular acceptance and adoption. And yes, many people will want to fuck them, or at the very least, want them to look fuckable.
The next time their computer system was hard down. Everything by paper in person. They didn't have enough forms for tracking it all, they were literally just writing things down on blank printer paper. No idea what cars were really in the lot. Show us your reservation and your id, we'll write it all down, and here's a key. Good luck finding the car. Was complete chaos.
The sexualization of young women working domestic labor is the result of the general sexualization of most young women in most contexts. It isn't that domestic labor is some sort of pretext to get a young woman into the home. It is that once they are there, men sexualize them.
Do you know something about the brain that makes it impossible to replicate its functionality technologically?
It is those things that are bottlenecking the price of robots.
The price of something tends towards the marginal cost, and the marginal cost of software is close to $0. Robots cost a lot more than that (what's the price of this robot?).
Edit: In fact Figure 03 imply marginal costs matter:
Mass manufacturing: Figure 03 was engineered from the ground-up for high-volume manufacturingI remain excited about AI and my master's thesis on fine-tuning transformers with LoRA. However, Hacker News has encouraged me to dive deeper and recognize that current architectures may not be sufficient. Additionally, Richard Sutton was a revelation for me in understanding how gradient descent truly works.
Joints will need replacement, lubrication, other maintenance. Same with motors, so why not replacing power/battery as part of a part?
It's not about making them "thinner". The idea is that they are human sized, because they're designed to operate in human occupied areas.
The first places are going to be like Amazon warehouses doing pick and pack because of the speed and all the variations of boxes and packages.
That's a relatively controlled environment, they'll evolve from there.
The essay was long so I cant claim I read it in detail - one q in my mind is whether humanoids need to do dexterity the same way that humans do. yes they dont have skin and tiny receptors but maybe there is another way to develop dexterity?
They are selling it the way AI has been sold. This will replace everyone's jobs. Thing is everyone is tired now, so many pointless layoffs, massive bubble, "AI-First"-desperate-ass companies.
Who will buy the robots if we are all unemployed?
> 30 grand
> we all start having them in our houses
Have you traveled ? Even a bit ? Most people don't make 30k in a year, before tax. It just shows how utterly disconnected from reality tech people, and especially execs, are. Virtually no one is going to pay 30k for a bot to mop the floor and fold their clothes
As for myself even if you'd pay me 30k I'd refuse to have one.
People now understood that the vast majority of productivity boosting technologies benefited a few % of people at the top, while the rest of use still basically live like in the 70s, except we work longer, have worse job security, can't buy houses, &c.
"Automation will free us from work", "Your car will drive itself in two years", "owning a car will cut your commute time in half", "LLMs will bring UBI", "Robots will fold your clothes so you can spend time on your family", fool me once...
https://economicsfromthetopdown.com/wp-content/uploads/2020/...
More importantly, who wants to stand behind a desk 8 hours a day and handle fussy customers? Probably some people, but the main motivation for the average hotel clerk is receiving money. Can we reorganize the economy so robots perform this kind of mundane work, while humans still receive money but can spend their time on more meaningful activities than standing behind a desk? I think a future like that is possible although it remains to be seen whether we will get it.
The truth hurts. Humanoid robots are getting better designs and currently do have extraordinary capabilities. Thing is, they're all cherry picked demos from carefully crafted test scenarios.
Until these things are thoroughly proven IRL with random tasks thrown at them then we can talk about negative comments. Until then its all marketing BS.
Services roles that are high contact / human interaction should have a human to deal with.
I am impressed by Unitree, but the problem that needs to be solved here is not just better software. Better hardware needs to come down in cost and weight to make the generalized robot argument more convincing.
There is still a long way to go for a humanoid to be a reasonable product, and that's not just a software issue.
Once software is "done" (we all know software is never done) you can just copy it and distribute it. It is negligiblehow much it costs to do so.
Once hardware is done you have to manufacture each and every piece of hardware with the same care, detail and reliability as the first one. You can't just click copy.
Often times you have to completely redesign the product to go from low volume high cost manufacturing to high volume low cost. A hand made McLaren is very different than an F-150.
The two simply scale differently, by nature of their beasts.
