Definitely an amazing tech given the tight budget but I struggle to see the appeal beyond pleasing my curious engineer mind. Any idea how such robot could be of any use for a “normal” private citizen (beyond carrying my water bottle while running)?
Excellent replacement for a security guard. If power consumption is low when not in motion, they can patrol, pause for a short time, watch their cameras, then move to a new position. Most aerial drones have to hover to patrol, which would make power consumption dramatically increase.
Knightscope has tried to deliver this (security guard mobile robot) for years, with decidedly mixed results:
>In December 2017, the San Francisco SPCA rented a Knightscope K5 robot to patrol the area next to their animal shelter, including the public sidewalk, to deter vandalism. The SPCA received complaints about using a robot on a public sidewalk where homeless people were encamped, and the robot was covered with a tarpaulin and smeared with barbecue sauce; the SPCA discontinued the contract.
> This is going to be a problem everywhere as robots are rolled out - people don’t like them.
Yeah this is a big problem, especially for unaccompanied autonomous robots.
I’m very bullish on this stuff in technological progress terms, but I’d still dislike seeing these things on the street unaccompanied. At best you’re being remotely recorded/streamed/observed/stalked by some anonymous remote operator (and the internet has taught us a lot about what human behaviour is like when there is total anonymity).
At worst it’s packed with explosives.
I’m not sure how we deal with this issue. I guess there will be a big difference between how accompanied vs unaccompanied robots will be viewed.
A bit like a large dog wondering about on its own would be treated with a lot more caution than one accompanied by its owner!
People don't like security guard robots that look like security guard robots. There's an easy answer to that: make it cute. I don't have any hard numbers, but I'd be willing to bet people are much less likely to vandalize a robot that looks like Wall-E than one that looks like ED-209, even if they do exactly the same thing.
Social ladder climbing for rich kids?
Buy groceries and bring them home one pack at a time?
If it can self recharge It would probably make some sense for patrolling property and report suspicious movement. To do that however it should be 100% trustworthy, that is, no mysterious Internet connections to phone home, no closed phone apps, etc. and the user should be able to integrate it with any video surveillance system in place, program it extensively from the paths to follow to associating sensors signals to events, alerts etc. That way it could also become useable aside the fact that it is indeed an engineering marvel.
Another possible use is helping blind people to walk around, however to do that it should become good at leading by finding the right path, not just following an human.
I'd be seriously interested in it as a mobility aid, as a "replacement" for a guide dog. Could be useful to some types of users. But yes, thats not really the average guy... I guess it is like with Alexa. First, there is the product, users will find usecases for it, and they will try fucking hard to legitimize their own investment.
Companionship with lower maintenance cost. The diversity of behavior is the problem but perhaps with the advances in neural nets we will get the behavior people desire in pets:
* unpredictable behavior within boundaries
* possible to alter behavior through repeated action (training)
* diversity of response to novel stimuli
Ultimately, what is it about a dog that makes it lovable? If we can replicate that and exploit the human desire to personify and love anything, we could create pets with intermediate intelligence and I think many people would enjoy that.
The actual product risk, I think, is making the robotic pet too useful. I suspect that would transform the pet into a tool. For instance, a robotic pet that reacts strangely and hides before the human owner detects a thunderstorm is far preferable to one which (using its link to the Internet) says “Thunderstorm approaching; batten down storm shutters”
The imperfection of understanding and the fact that the pet is a net cost center are two things that I think are crucial to mass market robotic pet adoption among those who desire companionship.
I personally would want companionship with something squishy and biologic. The idea that I'd be relying on something built and designed is extremely off putting.
It's masturbation vs sex, video games vs life, humanity vs borg-like collectivism.
If your friends are going to be machines, why not plug into a matrix like virtual reality where you can just live out your wildest dreams? Well, because it's fake.
To me, nothing has meaning without the constraints of reality and relationships to other biologic beings. That said, I suppose there may be a day when machines are even more sentient than I. Still, until I'm convinced there is a 'ghost in the shell', I just don't want it.
But then fooling humans into believing that about a robot dog is kinda your point I suppose. I just don't see the need other than some tech cos profit, and I find replacing a real dog with a contrived copy a bit cold.
Not sure why you’re being downvoted — many people would and indeed did find that useful vis a vis Sony’s Aibo. In fact, when those robotic dogs became unrepairable, some people would have a funeral for them: https://www.theguardian.com/world/2018/may/03/japan-robot-do...
Many of the dogs are accompanied by notes written by their former owners. “I feel relieved to know there will be a prayer for my Aibo,” one said. Another wrote: “Please help other Aibos. My eyes filled with tears when I decided to say goodbye.”
I think that if it can carry a water bottle, then it could carry a meal. The dog robots could automate restaurant food delivery for short range delivery in cities. They could possibly even climb stairs to reach walk up apartments.
In cities like NYC, I don’t think you’d see any buildings accept robots carrying unidentified packages in, at least within the next couple decades.
Considering people also still often steal iPhones and other electronics to pawn off, I can imagine this device would be the equivalent of a completely unaccompanied several thousand dollar iPhone on legs to a thief.
Hmm, maybe you could have some anti-theft system if it can detect it's being "kidnapped", by recording video/audio of the thief, and transmitting GPS and phoning a security service. Although I guess you could still capture it by throwing a blanket over it and putting it in a faraday cage.
But you can still detect exactly where it was robbed, and if the robots of each restaurant chain have a distinctive shape people would know it was stolen, reducing the likelihood of reselling it.
I am now chillingly reminded of Battle Angel Alita. Also in the live action remake the character called McTeague was a Hunter with robot dogs... this is not a good future.
It doesn't sound like they're equipped to autonomously navigate long distances. Not sure they understand things like street lights, crosswalks, construction manholes... my impression is they basically just stay "tethered" to your waist via an imaginary leash.
Regarding comments about theft... I think plain old vandalism would also be an issue.
It reminds me of the first home computers in the late 1970s. The product is not yet very useful, but attractive because home products like it soon will be.
This looks really cool and I’m sure the tech can be used to do all kinds of interesting things, but I’m struggling to think of an actual use case for the product. From the marketing and price it seems to be directed towards consumers—is there anything this can do for me that a backpack can’t do? Still want one, anyway.
I thought a dead simple military application would be sentries. Have them patrolling around FOBs, return and charge when they need to, fresh ones go out. Unarmed of course.
I like Black Mirror. Like an extension of the argument that "guns don't shoot people, people shoot people", Black Mirror displays the equation, XYZcoolTech+people = horrifying.
Apply that here and you can bet that "florida robot dog does stupid thing" will be a headline in the future.
The grandparent said autonomous armed robots that are currently in use (which excludes under development).
As far as we know, the answer to that question is zero, unless you might count a missile as an autonomous flying robot (but even that still has to be launched by human).
> is there anything this can do for me that a backpack can’t do?
I agree it’s early enough that the tangible uses for this type of tech are still a bit blurry to us. But I also think this is a bit like saying “is there anything a car can do for me that my legs can’t do?”
It depends on what criteria you evaluate it on. Yes you can move from A to B on your own steam, of course. But automation is incremental. These things aren’t a significant obvious benefit, until they are.
Scalability is one thing. Comparing writing notes in my iPad to writing them on paper, at first glance doesn’t look much different, apart from that the iPad can effectively store a lifetime worth of notes, easily searchable, shareable, linkable etc. Its scalability potential (for this use case) is almost infinite.
When evaluating these sorts of things it’s worth bearing in mind the “leverage” that the tech might give us, that we didn’t have before. This scalability/leverage benefit is often not immediately obvious until you look a little bit below the surface.
This robot dog might start off by carrying your water bottle but later it might be able to also carry a foldable chair, a waking stick, food and water, and a communications device, suddenly meaning that someone with limited mobility can actually go out for a walk on their own and have everything they might need carried alongside them as a mobile support system. That person just gained a little bit of extra freedom. At the enhancement end of the spectrum, an athlete could use this “support robot” to extend the possible range of their training runs.
I do wonder about the near term use cases for this sort of tech (at this price point) for people with disabilities of various kinds.
Often leading edge tech starts out in the realm of compensating for disabilities for a smaller audience before progressing to being looked upon as an “enhancement” for a wider audience (because it becomes affordable enough and useful enough to start to appeal to people who don’t need to compensate for any particular issue but see it as a net convenience worth paying for).
I'm seriously handicapped and have a dog phobia thanks to ugly events in early childhood. My handicap makes having actual animals in my home a bad idea anyway for cleanliness reasons.
While $2700 is way out of my price range -- handicapped people tend to have tight budgets -- I can absolutely see a tremendous market for companion animals for people allergic to animals, with a dog phobia etc.
Especially if they could add some fetching capabilities, this could be enormously useful to many people. Our population is aging and unlike other populations with physical limitations, people whose limitations are primarily age related sometimes have money. Sometimes lots of money in fact.
Companion animals are not pets. They are working animals.
Battery life has a tendency to change over time. For someone largely housebound, it wouldn't much matter. They could be on a tether plugged into a wall even and still do good things for many people.
I meant it’s not fit for a companion. It wouldn’t have agency, “keep you company” in any sense or understand your needs like animals can be trained to do. It’s a sharp, computerized tool.
I don't know a better term for what I am thinking of. I don't think it has to be able to serve that specific function in order to be useful to someone with significant limitations.
I can see where the miscommunication comes from, but I think my point stands. For handicapped people, a mechanical dog that can carry, fetch items for them, help them get up, light their way, etc is potentially very useful.
These things will be used by law enforcement. Deploying these and having remote cops surveil is cheaper than putting a cop on every corner.
The implications for a surveillance society are pretty intense though, but it can easily clean up tough neighborhoods where deterrents are necessary sadly.
I can also imagine pull over stops where the robot engages the passenger first, and is only equipped with a taser. You basically take the human out of the equation so there is no fear of getting shot by the officer or the passenger.
That’s 1/27th of Spot Mini. And probably only user programmable until manufacturer says but not you. My spine is frozen, don’t even feel the chill anymore.
So it takes a lot of R&D to find the general shape of a solution, but once that's been discovered, someone else can replicate similar results easily with cheaper components?
The Spot has an extremely impressive stabilization control system. If the Go1 had something close, they would have shown it in this ad.
The other thing to keep in mind is that this ad might be showing the best-case operation. On the other hand, the Spot has been extensively used/tested by customers already.
Not industrial espionange. This robot is heavily based off of the MIT mini cheetah, which most of the design for is opensource (e.g. [0], [1]).
Ben Katz has an amazing blog[2] which if you dig through you can read about a lot of the design decisions and testing for the mini cheetah.
Funny enough, after designing the mini cheetah for his master's thesis Ben went to work for Boston Dynamics (I think he started there two or three years ago).
I'm super interested in using this for my masters thesis. But their github docs are terse at best. I have no confidence that I'd be talented enough to figure out their api, and the mention of the broken website form further deters me. Too bad; I love the concept of quadrupeds for navigating stairs and other urban obstacles.
Bittle is obviously not of the same size as the Go1 so won't be suited to the same practical tasks but if you're looking at exploring the control aspects it would serve as a good platform to learn on (for a substantially lower initial outlay). Petoi has the OpenCat repo and there are plenty of ways to hook it up to additional sensors or a remote system for more advanced guidance etc.
Edit: Mine arrived on Friday. It's great, although putting it together had a few challenges (the springs are a tough fit, but people figured workarounds on the forum)
Interesting. If there is a market for these (over wheeled robots) then if the Unitree robot is in anyway capable compared with Boston dynamics there’s a good chance they will become the market leader in the same way that DJI ‘owns’ the drone market.
$2.7K is cheap enough for enough consumers to purchase to give them the economies of scale, profit and R&D capital, that you can’t get with selling $70K industrial units.
They don't specify the actual load-carrying capacity, but I would consider that the critical detail determining the practical utility of this product. Relieving you of the weight of a water bottle is a fun tech demo, but if it could carry a modest backpack's worth of stuff you'd otherwise be lugging yourself, that would have some real benefits in certain contexts.
The article says that the Edu version has a capacity of 5kg which is an upgrade of 2kg. This would imply a load-carrying capacity of 3kg for the base model. Not a lot, but not totally insignificant.
I don't think the intended use case is dog replacement. Human beings just can't help but anthropomorphize, thus it's a robot dog instead of a quadrupedal load-carrying robot.
How will this perform in rain and snow? What about operating temperatures? No mention of carrying capacity limits on this article or Unitree's website. How well does it handle variable terrain (ie, hiking)?
I can't see myself buying a $2700 piece of tech that can only be used in ideal conditions. Seems like a novelty item at this point.
Even in small numbers they would be great to patrol lesser used hiking trails once the battery is improved a bit.
It’s not uncommon for a hiker, particularly an elderly one, to stumble and slide downhill off the trail and be unable to make it back up due to an injury.
Just last Fall my wife and I came across a man whose horse had stumbled during a stream crossing and had fallen down a hill. The man had a torn ACL and couldn’t get out.
Had we and some passing trail runners not found this man, he would have certainly died that night as he had no gear, was soaked wet, and it was going to be well below freezing.
Imagine if the parks service just had a robot dog to do a final patrol of the trail at sundown, maybe with some thermal cameras or something.
If there are drones that have the battery life to fly ~15km under the tree canopy without hitting any trees or branches, etc, they would also be a great option.
I think the high end consumer drones have range in this ballpark (eg DJI Mavic 2 Pro can go 18km with no wind).
I’m not sure how much tree avoidance would be a problem, I’ve seen some pretty cool demos of autonomous forest drones.
My intuition is that the robot dogs would be slightly more robust since they can self-right if they stumble, whereas if a drone hits a small obstacle it can pretty easily come tumbling out of the air and get stuck on the ground. I think the dogs could also be made more weatherproof than drones.
I bet drones would be a much better option for lightly forested areas or desert areas, though, where they can generally fly above any obstacles.
There is actually a fair bit of research going on into using machine learning and image recognition to try detect bodies under tree canopies using thermal imaging
Nice. That's a reasonable price point for what's inside, and could get much cheaper with volume. Good robots have always suffered from a low-volume, high cost problem.
I imagine the project would be broken into hardware challenges (mechatronics, electronics, and manufacturing) and software (AI, programmability).
Most likely because everything you could need in the software world is free (eg Tensorflow, CUDA, etc). Everything to do with hardware is very expensive (Altium, Proteus, Simulink, any sort of CNC control).
Could work with simpler vermin, but smarter animals would surely learn that robot has no effective means of harming them (no way to bite or scratch, unlike a real dog; maybe punching, like a horse?), and reclassify it from a threat to a nuisance.
I wonder if someday when robot dogs are as affordable and intelligent as biological ones, will the market drop out for biologicals? All the selective breeding for hypoallergenic or good behavior is moot if it's programmable. Turn it to guard mode when you leave home, have a cuddle buddy when you're there.
Unlikely. There might be a drop in demand for the breeds used for home defense or herding, but most breeds are recreational and people enjoy the company of other living, breathing things.
>people enjoy the company of other living, breathing things
Can we say the same thing in the future? If a child is given a pet robot will they get bored of it like a toy or will they come to see it as a life long companion?
The likelihood is that industrialization will turn these toys into things that have short lifespans and are commonly replaced with new models, like phones. Children also have notoriously short attention spans. I think biological pets are more for the sanity of the parents than the children.
But these are things that you play with as opposed to what we hope to have in the future which is something that "plays with you". Aka a friend or a guardian of sorts.
For many people a dog is there to be a warm cuddly friend who is excited to see you and loves you. It’s a personal relationship. A robot that does exactly what you program it to do doesn’t hit the same notes.
> A robot that does exactly what you program it to do doesn’t hit the same notes.
Out of all animals, the one that most closely fits the bill of "does exactly what you tell it to do" is dogs, and I would argue that's exactly why it's so popular.
That’s true, but I’m not sure if obedience is the primary reason for their popularity? I would argue it’s primarily out of love. There are probably a small minority of dog owners who do it solely for the utility, or just to exercise dominance over another animal. Beats me, I’m a cat person. A robocat would be pointless.
One could make the case that the carefully cultivated selective breeding of wolves into domesticated dogs over several millennia was just the evolutionary equivalent of digital programming.
I dont know why everyone brings it up as if it preceded the tech. The episode was purely based on the designs coming out of Boston dynamics and co and while popular the episode is purely centered around hunting suspense and nothing more relevant.
But it's not like this tech device will have a lesser maintenance cost.
You have to replace your phone every few years, imaging that with a dog. Then think about the dollar and environmental cost of throw away tech devices.
Definitely an amazing tech given the tight budget but I struggle to see the appeal beyond pleasing my curious engineer mind. Any idea how such robot could be of any use for a “normal” private citizen (beyond carrying my water bottle while running)?
Excellent replacement for a security guard. If power consumption is low when not in motion, they can patrol, pause for a short time, watch their cameras, then move to a new position. Most aerial drones have to hover to patrol, which would make power consumption dramatically increase.
Knightscope has tried to deliver this (security guard mobile robot) for years, with decidedly mixed results:
>In December 2017, the San Francisco SPCA rented a Knightscope K5 robot to patrol the area next to their animal shelter, including the public sidewalk, to deter vandalism. The SPCA received complaints about using a robot on a public sidewalk where homeless people were encamped, and the robot was covered with a tarpaulin and smeared with barbecue sauce; the SPCA discontinued the contract.
https://en.wikipedia.org/wiki/Knightscope
> the robot was covered with a tarpaulin and smeared with barbecue sauce
A seasoned veteran.
This is going to be a problem everywhere as robots are rolled out - people don’t like them.
> This is going to be a problem everywhere as robots are rolled out - people don’t like them.
Yeah this is a big problem, especially for unaccompanied autonomous robots.
I’m very bullish on this stuff in technological progress terms, but I’d still dislike seeing these things on the street unaccompanied. At best you’re being remotely recorded/streamed/observed/stalked by some anonymous remote operator (and the internet has taught us a lot about what human behaviour is like when there is total anonymity).
At worst it’s packed with explosives.
I’m not sure how we deal with this issue. I guess there will be a big difference between how accompanied vs unaccompanied robots will be viewed.
A bit like a large dog wondering about on its own would be treated with a lot more caution than one accompanied by its owner!
Might work well as a patrol of private property in a “release the hounds” kind of way.
People don't like security guard robots that look like security guard robots. There's an easy answer to that: make it cute. I don't have any hard numbers, but I'd be willing to bet people are much less likely to vandalize a robot that looks like Wall-E than one that looks like ED-209, even if they do exactly the same thing.
Social ladder climbing for rich kids? Buy groceries and bring them home one pack at a time?
If it can self recharge It would probably make some sense for patrolling property and report suspicious movement. To do that however it should be 100% trustworthy, that is, no mysterious Internet connections to phone home, no closed phone apps, etc. and the user should be able to integrate it with any video surveillance system in place, program it extensively from the paths to follow to associating sensors signals to events, alerts etc. That way it could also become useable aside the fact that it is indeed an engineering marvel.
Another possible use is helping blind people to walk around, however to do that it should become good at leading by finding the right path, not just following an human.
A $2700 robot with a handle seems like a good thing to steal.
I'd be seriously interested in it as a mobility aid, as a "replacement" for a guide dog. Could be useful to some types of users. But yes, thats not really the average guy... I guess it is like with Alexa. First, there is the product, users will find usecases for it, and they will try fucking hard to legitimize their own investment.
Companionship with lower maintenance cost. The diversity of behavior is the problem but perhaps with the advances in neural nets we will get the behavior people desire in pets:
* unpredictable behavior within boundaries
* possible to alter behavior through repeated action (training)
* diversity of response to novel stimuli
Ultimately, what is it about a dog that makes it lovable? If we can replicate that and exploit the human desire to personify and love anything, we could create pets with intermediate intelligence and I think many people would enjoy that.
The actual product risk, I think, is making the robotic pet too useful. I suspect that would transform the pet into a tool. For instance, a robotic pet that reacts strangely and hides before the human owner detects a thunderstorm is far preferable to one which (using its link to the Internet) says “Thunderstorm approaching; batten down storm shutters”
The imperfection of understanding and the fact that the pet is a net cost center are two things that I think are crucial to mass market robotic pet adoption among those who desire companionship.
I personally would want companionship with something squishy and biologic. The idea that I'd be relying on something built and designed is extremely off putting.
It's masturbation vs sex, video games vs life, humanity vs borg-like collectivism.
If your friends are going to be machines, why not plug into a matrix like virtual reality where you can just live out your wildest dreams? Well, because it's fake.
To me, nothing has meaning without the constraints of reality and relationships to other biologic beings. That said, I suppose there may be a day when machines are even more sentient than I. Still, until I'm convinced there is a 'ghost in the shell', I just don't want it.
But then fooling humans into believing that about a robot dog is kinda your point I suppose. I just don't see the need other than some tech cos profit, and I find replacing a real dog with a contrived copy a bit cold.
Not sure why you’re being downvoted — many people would and indeed did find that useful vis a vis Sony’s Aibo. In fact, when those robotic dogs became unrepairable, some people would have a funeral for them: https://www.theguardian.com/world/2018/may/03/japan-robot-do...
Many of the dogs are accompanied by notes written by their former owners. “I feel relieved to know there will be a prayer for my Aibo,” one said. Another wrote: “Please help other Aibos. My eyes filled with tears when I decided to say goodbye.”
I think that if it can carry a water bottle, then it could carry a meal. The dog robots could automate restaurant food delivery for short range delivery in cities. They could possibly even climb stairs to reach walk up apartments.
In cities like NYC, I don’t think you’d see any buildings accept robots carrying unidentified packages in, at least within the next couple decades. Considering people also still often steal iPhones and other electronics to pawn off, I can imagine this device would be the equivalent of a completely unaccompanied several thousand dollar iPhone on legs to a thief.
Hmm, maybe you could have some anti-theft system if it can detect it's being "kidnapped", by recording video/audio of the thief, and transmitting GPS and phoning a security service. Although I guess you could still capture it by throwing a blanket over it and putting it in a faraday cage.
But you can still detect exactly where it was robbed, and if the robots of each restaurant chain have a distinctive shape people would know it was stolen, reducing the likelihood of reselling it.
Robot chop shops? There might be some value in reselling the batteries alone.
Make it all a custom single piece, so even the battery is no good for anything except for this model of robot.
I am now chillingly reminded of Battle Angel Alita. Also in the live action remake the character called McTeague was a Hunter with robot dogs... this is not a good future.
You could even automate the task with more robots.
It doesn't sound like they're equipped to autonomously navigate long distances. Not sure they understand things like street lights, crosswalks, construction manholes... my impression is they basically just stay "tethered" to your waist via an imaginary leash.
Regarding comments about theft... I think plain old vandalism would also be an issue.
it could take your meat dog out for a walk for you
A man's best friend's best friend.
Or chasing a toddler around the park?
> meat dog
I love that in the future this will be a distinction which will have to be made
That sort of thing (all terrain porter) could be a compelling use case, but with such a short range and payload, probably not for very many people.
It reminds me of the first home computers in the late 1970s. The product is not yet very useful, but attractive because home products like it soon will be.
it is missing a "mouth"/grabber.
Maybe you could use one of these to carry your purse dog around for you.
This looks really cool and I’m sure the tech can be used to do all kinds of interesting things, but I’m struggling to think of an actual use case for the product. From the marketing and price it seems to be directed towards consumers—is there anything this can do for me that a backpack can’t do? Still want one, anyway.
I thought a dead simple military application would be sentries. Have them patrolling around FOBs, return and charge when they need to, fresh ones go out. Unarmed of course.
This is done in Black Mirror, Season 4 episode 5 Metalhead. It’s terrifying. The dogs were armed.
I like Black Mirror. Like an extension of the argument that "guns don't shoot people, people shoot people", Black Mirror displays the equation, XYZcoolTech+people = horrifying.
Apply that here and you can bet that "florida robot dog does stupid thing" will be a headline in the future.
"The street finds its own uses for things." - William Gibson
https://williamgibson.fandom.com/wiki/Slamhound
> Unarmed of course.
I admire your optimism!
How many autonomous armed robots does the US currently use?
Is that a trick question? Do flying ones count?
The grandparent said autonomous armed robots that are currently in use (which excludes under development).
As far as we know, the answer to that question is zero, unless you might count a missile as an autonomous flying robot (but even that still has to be launched by human).
Zero
> is there anything this can do for me that a backpack can’t do?
I agree it’s early enough that the tangible uses for this type of tech are still a bit blurry to us. But I also think this is a bit like saying “is there anything a car can do for me that my legs can’t do?”
It depends on what criteria you evaluate it on. Yes you can move from A to B on your own steam, of course. But automation is incremental. These things aren’t a significant obvious benefit, until they are.
Scalability is one thing. Comparing writing notes in my iPad to writing them on paper, at first glance doesn’t look much different, apart from that the iPad can effectively store a lifetime worth of notes, easily searchable, shareable, linkable etc. Its scalability potential (for this use case) is almost infinite.
When evaluating these sorts of things it’s worth bearing in mind the “leverage” that the tech might give us, that we didn’t have before. This scalability/leverage benefit is often not immediately obvious until you look a little bit below the surface.
This robot dog might start off by carrying your water bottle but later it might be able to also carry a foldable chair, a waking stick, food and water, and a communications device, suddenly meaning that someone with limited mobility can actually go out for a walk on their own and have everything they might need carried alongside them as a mobile support system. That person just gained a little bit of extra freedom. At the enhancement end of the spectrum, an athlete could use this “support robot” to extend the possible range of their training runs.
I do wonder about the near term use cases for this sort of tech (at this price point) for people with disabilities of various kinds.
Often leading edge tech starts out in the realm of compensating for disabilities for a smaller audience before progressing to being looked upon as an “enhancement” for a wider audience (because it becomes affordable enough and useful enough to start to appeal to people who don’t need to compensate for any particular issue but see it as a net convenience worth paying for).
It could probably be programmed to fetch me a beer. Or walk around the yard serving appetizers to guests during a BBQ.
It could also be an effective scarecrow I imagine. When planting certain things it would be nice to reliably keep the birds and varmint away.
I'd quite like it to just walk next to me and have a power cable to a drone that hovers and holds my coffee. #winning
I 100% buy into this vision.
Carrying groceries 10 blocks in NYC isn't the easiest thing. This would help.
But if you live in car dominated sprawl then idk.
I want one of these. However, they say the payload is only 2kg. Won't carry that many groceries.
The obvious solution to carrying 20kg would be to have 10 of these robots following you.
I think there was a lost opportunity with the latest 101 dalmatians remake.
Perhaps it can pull a trailer?
I'm seriously handicapped and have a dog phobia thanks to ugly events in early childhood. My handicap makes having actual animals in my home a bad idea anyway for cleanliness reasons.
While $2700 is way out of my price range -- handicapped people tend to have tight budgets -- I can absolutely see a tremendous market for companion animals for people allergic to animals, with a dog phobia etc.
Especially if they could add some fetching capabilities, this could be enormously useful to many people. Our population is aging and unlike other populations with physical limitations, people whose limitations are primarily age related sometimes have money. Sometimes lots of money in fact.
This is much more of an industrial robot than a pet. And 20 minute battery life.
Companion animals are not pets. They are working animals.
Battery life has a tendency to change over time. For someone largely housebound, it wouldn't much matter. They could be on a tether plugged into a wall even and still do good things for many people.
I meant it’s not fit for a companion. It wouldn’t have agency, “keep you company” in any sense or understand your needs like animals can be trained to do. It’s a sharp, computerized tool.
I don't know a better term for what I am thinking of. I don't think it has to be able to serve that specific function in order to be useful to someone with significant limitations.
I can see where the miscommunication comes from, but I think my point stands. For handicapped people, a mechanical dog that can carry, fetch items for them, help them get up, light their way, etc is potentially very useful.
IIRC it can jog for an hour.
The clearest specifications I could find on their site is here: https://www.unitree.com/uploads/%E5%A4%A9%E7%8C%AB-%E8%8B%B1...
In the marketing material they say that the processing power compares favorably to a TX2, but it looks like their $8.5k model is using Nano and/or NX.
Wondering if this will fill the market segment that bought the Sony Aibo.
These things will be used by law enforcement. Deploying these and having remote cops surveil is cheaper than putting a cop on every corner.
The implications for a surveillance society are pretty intense though, but it can easily clean up tough neighborhoods where deterrents are necessary sadly.
I can also imagine pull over stops where the robot engages the passenger first, and is only equipped with a taser. You basically take the human out of the equation so there is no fear of getting shot by the officer or the passenger.
That’s 1/27th of Spot Mini. And probably only user programmable until manufacturer says but not you. My spine is frozen, don’t even feel the chill anymore.
Why does the Spot Mini cost so much and this so little?
This isn't industrial espionage is it?
Is Boston Robotics trying to recover R&D for everything else they do?
Spot mini was just the first. When 3D printers came out, their price for a decent one was high for both FDM and resin. Now most of them are sub-$1000.
So it takes a lot of R&D to find the general shape of a solution, but once that's been discovered, someone else can replicate similar results easily with cheaper components?
Iterative improvement marches down the price...
Yes! Certainly no shortage of cheap electric cars and scooters now in China.
The Spot has an extremely impressive stabilization control system. If the Go1 had something close, they would have shown it in this ad.
The other thing to keep in mind is that this ad might be showing the best-case operation. On the other hand, the Spot has been extensively used/tested by customers already.
Not industrial espionange. This robot is heavily based off of the MIT mini cheetah, which most of the design for is opensource (e.g. [0], [1]).
Ben Katz has an amazing blog[2] which if you dig through you can read about a lot of the design decisions and testing for the mini cheetah.
Funny enough, after designing the mini cheetah for his master's thesis Ben went to work for Boston Dynamics (I think he started there two or three years ago).
0: https://github.com/mit-biomimetics/Cheetah-Software
1: https://dspace.mit.edu/handle/1721.1/118671
2: https://build-its-inprogress.blogspot.com/
Put a four foot mast camera on it and I can see influencers eating it up.
I'm super interested in using this for my masters thesis. But their github docs are terse at best. I have no confidence that I'd be talented enough to figure out their api, and the mention of the broken website form further deters me. Too bad; I love the concept of quadrupeds for navigating stairs and other urban obstacles.
Then it's an attractive point for hacking: Use an existing system and repurpose it to serve your own goal.
You may already be aware of it, but the Bittle from Petoi may be of interest:
https://www.petoi.com/
Bittle is obviously not of the same size as the Go1 so won't be suited to the same practical tasks but if you're looking at exploring the control aspects it would serve as a good platform to learn on (for a substantially lower initial outlay). Petoi has the OpenCat repo and there are plenty of ways to hook it up to additional sensors or a remote system for more advanced guidance etc.
Edit: Mine arrived on Friday. It's great, although putting it together had a few challenges (the springs are a tough fit, but people figured workarounds on the forum)
Interesting. I didn't realize these were shipping yet. Seems like a good option for exploration!
Interesting. If there is a market for these (over wheeled robots) then if the Unitree robot is in anyway capable compared with Boston dynamics there’s a good chance they will become the market leader in the same way that DJI ‘owns’ the drone market.
$2.7K is cheap enough for enough consumers to purchase to give them the economies of scale, profit and R&D capital, that you can’t get with selling $70K industrial units.
They don't specify the actual load-carrying capacity, but I would consider that the critical detail determining the practical utility of this product. Relieving you of the weight of a water bottle is a fun tech demo, but if it could carry a modest backpack's worth of stuff you'd otherwise be lugging yourself, that would have some real benefits in certain contexts.
The article says that the Edu version has a capacity of 5kg which is an upgrade of 2kg. This would imply a load-carrying capacity of 3kg for the base model. Not a lot, but not totally insignificant.
The owner looked so lonely while walking in the park with the robot dog. Isn't companionship a big reason to have a (bio) dog.
My washing machine is a great companion. Barely needs play and just obeys my commands. Plus keeps my clothes smelling fresh.
I don't think the intended use case is dog replacement. Human beings just can't help but anthropomorphize, thus it's a robot dog instead of a quadrupedal load-carrying robot.
> I don't think the intended use case is dog replacement.
Maybe not in this iteration, but why not? There are already robotic seals[0] for companionship, so a robotic dog doesn't seem too far off.
[0] https://en.wikipedia.org/wiki/Paro_(robot)
That just seems strange to me, but hey, I wouldn't begrudge a pet robot if it makes anyone happy.
Maybe I'll look back at this post in 30 years and chuckle at how little foresight I had when they're incredibly popular.
Think cameras, gps and microphones instead.
How will this perform in rain and snow? What about operating temperatures? No mention of carrying capacity limits on this article or Unitree's website. How well does it handle variable terrain (ie, hiking)?
I can't see myself buying a $2700 piece of tech that can only be used in ideal conditions. Seems like a novelty item at this point.
Has anyone been able to actually order one of these? The Captcha on the company's order form appears to be broken.
I got it to work
These would be great in large quantities for search and rescue of large areas, perhaps.
Even in small numbers they would be great to patrol lesser used hiking trails once the battery is improved a bit.
It’s not uncommon for a hiker, particularly an elderly one, to stumble and slide downhill off the trail and be unable to make it back up due to an injury.
Just last Fall my wife and I came across a man whose horse had stumbled during a stream crossing and had fallen down a hill. The man had a torn ACL and couldn’t get out.
Had we and some passing trail runners not found this man, he would have certainly died that night as he had no gear, was soaked wet, and it was going to be well below freezing.
Imagine if the parks service just had a robot dog to do a final patrol of the trail at sundown, maybe with some thermal cameras or something.
How would these work better than drones with thermal cameras.
If there are drones that have the battery life to fly ~15km under the tree canopy without hitting any trees or branches, etc, they would also be a great option.
I think the high end consumer drones have range in this ballpark (eg DJI Mavic 2 Pro can go 18km with no wind).
I’m not sure how much tree avoidance would be a problem, I’ve seen some pretty cool demos of autonomous forest drones.
My intuition is that the robot dogs would be slightly more robust since they can self-right if they stumble, whereas if a drone hits a small obstacle it can pretty easily come tumbling out of the air and get stuck on the ground. I think the dogs could also be made more weatherproof than drones.
I bet drones would be a much better option for lightly forested areas or desert areas, though, where they can generally fly above any obstacles.
There is actually a fair bit of research going on into using machine learning and image recognition to try detect bodies under tree canopies using thermal imaging
Doing drones out of visual line of sight is quite dangerous. Worst thing if a robot dog “crashes” is, you pick it up next morning to recharge it.
Nice. That's a reasonable price point for what's inside, and could get much cheaper with volume. Good robots have always suffered from a low-volume, high cost problem.
> Our company actually spent more manpower and money than software on the hardware level such as machinery.
I would like to know what was meant here -- software was the biggest cost?
I imagine the project would be broken into hardware challenges (mechatronics, electronics, and manufacturing) and software (AI, programmability).
Most likely because everything you could need in the software world is free (eg Tensorflow, CUDA, etc). Everything to do with hardware is very expensive (Altium, Proteus, Simulink, any sort of CNC control).
This could be really cool to protect a garden or bird feeder.
Could work with simpler vermin, but smarter animals would surely learn that robot has no effective means of harming them (no way to bite or scratch, unlike a real dog; maybe punching, like a horse?), and reclassify it from a threat to a nuisance.
Looks like one of the alien robot assassin dogs on BBC TV series War of the Worlds?
I wonder if someday when robot dogs are as affordable and intelligent as biological ones, will the market drop out for biologicals? All the selective breeding for hypoallergenic or good behavior is moot if it's programmable. Turn it to guard mode when you leave home, have a cuddle buddy when you're there.
Also, no pooping or vet bills.
But awesome security implications.
Just the occasional mech bill after some parts replacement.
Unlikely. There might be a drop in demand for the breeds used for home defense or herding, but most breeds are recreational and people enjoy the company of other living, breathing things.
>people enjoy the company of other living, breathing things
Can we say the same thing in the future? If a child is given a pet robot will they get bored of it like a toy or will they come to see it as a life long companion?
Most Tamagotchis and Furbys had a short lifespan.
https://www.vice.com/en/article/dy8axx/the-tamagotchi-cemete...
The likelihood is that industrialization will turn these toys into things that have short lifespans and are commonly replaced with new models, like phones. Children also have notoriously short attention spans. I think biological pets are more for the sanity of the parents than the children.
But these are things that you play with as opposed to what we hope to have in the future which is something that "plays with you". Aka a friend or a guardian of sorts.
When robot dogs become legit intelligent it means it’s time for us to pass the torch to our successors
For many people a dog is there to be a warm cuddly friend who is excited to see you and loves you. It’s a personal relationship. A robot that does exactly what you program it to do doesn’t hit the same notes.
> A robot that does exactly what you program it to do doesn’t hit the same notes.
Out of all animals, the one that most closely fits the bill of "does exactly what you tell it to do" is dogs, and I would argue that's exactly why it's so popular.
That’s true, but I’m not sure if obedience is the primary reason for their popularity? I would argue it’s primarily out of love. There are probably a small minority of dog owners who do it solely for the utility, or just to exercise dominance over another animal. Beats me, I’m a cat person. A robocat would be pointless.
I wonder where the line is though, where a sufficiently advanced simulation of this behaviour would actually be enough for some people.
One could make the case that the carefully cultivated selective breeding of wolves into domesticated dogs over several millennia was just the evolutionary equivalent of digital programming.
Oh no, you're right! Humanity is monstrous. Maybe we should start putting the wolf back into the dogs.
My brother bought a dog (pure bred) for about €1500. Add food & veterinary costs over 10ish years, I’m pretty sure this robot dog is cheaper.
Why are these not being marketed to do something useful.
I mean a $3,000 water bottle holder...
They are. That video shows reconnaissance applications. Who do you think is in the market for 10,000 of these?
Its job is to be a lobster on a leash. To bring attention to itself and its owner.
They background music makes it hard to know the noise it really makes.
This is a really good point, my first thought was "I bet you it's super loud annoying"
That's an expensive way to carry your one bottle of shopping.
There is a Black Mirror about this:
https://en.m.wikipedia.org/wiki/Metalhead_(Black_Mirror)
I dont know why everyone brings it up as if it preceded the tech. The episode was purely based on the designs coming out of Boston dynamics and co and while popular the episode is purely centered around hunting suspense and nothing more relevant.
Real Dog from your local shelter Looks Pretty Great, Costs Just $50
A real dog costs much more than $50 after recurring long term food and medical expenses like annual vet visits.
But it's not like this tech device will have a lesser maintenance cost.
You have to replace your phone every few years, imaging that with a dog. Then think about the dollar and environmental cost of throw away tech devices.
My honest uninformed sentiment: The device has very little utility so it will have no maintenance cost. ;)
My feeling is that there is no overlap whatsoever between use cases for a real dog vs. a robot dog…
It looks like it can carry water. But I need it to be able to jog for over an hour!
Ah, another other model Aliengo up to 4.5 hours of battery life.
Very slick production video but the grammar / spelling errors instantly tank credibility.
Serious question - where is the disconnect with Chinese OEMs not being able to do basic spell / grammar checking?
I suspect it is just local availability of English speakers.
Western copy writers and copy editors are expensive by Chinese standards and access to them is hard for a Chinese firm.