Boston Dynamics robot Atlas goes hands on [video]
youtube.com
youtube.com
The task it is doing is undemanding. It's just moving things from one set of large slots to another. No need for precision placement, unstructured bin-picking, or object separation. If it could pick up engine covers from the messy pile seen atop one of the racks and slot them into the storage unit, that would be more impressive. It's cool to see this done with a humanoid, but off the shelf industrial robots could do that job. This is the same place where Rethink Robotics got stuck. They could do simple object movements in mostly-structured situations, but so can lots of other simpler approaches.
Amazon, despite substantial efforts, still doesn't have full robot picking. About two years ago, Amazon announced their "Sparrow" picking robot. But that seems to be experimental. It's not seen in videos of Amazon warehouses in 2024. Amazon is using the Agility humanoid, but only experimentally.[1]
This is how Amazon currently does picking.[2] Racks of product come to the picker on robotic platforms. The picking system projects a light square on the space in the rack from which the picker should take the product. The picker picks the item, waves it under a barcode scanner, and drops it in an outgoing bin. Repeat for 8 hours. The job requires no more than a room-temperature IQ. Machines should think. People should work.
Amazon keeps trying to automate that step.
Have you noticed how in second 13 it rotates the lower body clockwise while the upper body rotates counter-clockwise in order to optimize the movements to archive the goal?
All (bipedal/quadrupedal) robots I've seen behave either like an animal or a human, but this one is something else. Imagine the inverse kinematics required to perform such a movement, and how precise even the slight left foot's clockwise rotation is.
It does these very uncanny movements multiple times in the video, this video less about moving that thing from here to there than about how it does it. Also the recovery it had when it failed to insert one of the objects.
Could be a fancy optimization of course, but could as well be a side effect of a decoupled planning of the locomotion part and the upper body pick and place.
Algo could be: plan a path for a lower body motion from pose (4dof) A to pose B. And given pose B plan the upper body place of the tray in the target. If no constraints are broken plan both in parallel so it looks like one smooth behavior.
Obviously, I don’t know :)
Nevertheless, it fascinating how the legs/torso/head spin in different directions. It's scary even if you know it is physically weaker than its hydraulic cousins.
You mean the unnecessarily violent jerking motion it does for no great reason at 1:22 instead of gently trying again?
- Adam Smith
Because of actuator density- a humanoid has a high number of actuators. Also power density, and they did a lot of work in molding the hydraulic actuators directly into the arms and legs to make them compact. The bulk is all in the backpack pumps, power pack, etc.
So this move to all-electric represents real growth in Boston Dynamics being able to make compact motors that meet their needs of performance.
I am betting this new robot is less capable of the explosive dynamics of the old one. I am not expecting dance performances and backflips from this one right away. But I am expecting them to work toward that kind of performance.
Electrics win for energy efficiency and no-leaky-fluid, and also are improving pretty quick.
Then Schaft won the DARPA Robotic Challenge in 2013 with their all-electric humanoid robot with liquid-cooled motors. They beat out Boston Dynamics' Atlas. Then Google bought Schaft. Google got bored with robotics, couldn't find a buyer for them, and Schaft was shut down, along with several other Google robotics acquisitions.
Amazon could have gone in the other direction. The layout could be designed so that the picker sits on a swivel stool and never stands at all. That's what they do when they test robot picking - there's one human-sized robot arm on a turntable mount doing the picking.
Humans would get fat with that layout.
I don't think you'd be able to get this to move other objects or to other types of shelves without getting a programmer involved for a few days at least.
1: https://www.youtube.com/watch?v=zQeQWGqfFN0
Especially in food manufacturing, they need to add certain chemicals in the food ingredients so that the machines can process them and that's what separates the cheap mass produced food from artisan food.
A humanoid robot might be able to apply traditional processes and make artisan, additive free products cheap and available.
If we end up having practically unlimited energy(which we actually have, just not harnessed fully yet), we can also have small production centers on every corner that produce bespoke products using humanoid robots and traditional tools instead of having centralized and highly specialized and streamlined mass production.
Food additives are almost entirely about making products more visually appealing or shelf stable. Cellulose to keep cheese from sticking, nitrogen to keep meat from oxidizing, etc.
Your 'bespoke artisan' products are mostly improved marketing and improved staging (better lighting, less crowded stores, more personal service, etc)
Same idea with individual production. A toyota factory builds you one exact copy of a car in any colour you want (as long as it's black). But we might go back to a more individualised default in manufacturing, where for example tailor made clothes become affordable to most people again.
That's still too much "inside the box" thinking.
That baker doesn't need "automate everything", in fact, no-one does.
What "that baker" needs, is to rank work on effort, cross sectioned with "costs to automate". And then automate the top item or items only.
The baker needs a machine to knead the dough. Special tools to cut the dough of twenty cookies at once with minimal waste. Trays and tools that can be cleaned in a dishwasher. A bread slicing machine. A bakery that can be mopped in one go, rather than a fully automated humanoid mopping machine.
This is an analogy for all automation: we don't need to cover 100%, be fully autonomous, 100% flexible. We need to automate the hardest part, even if that's boring tech. Then the next hardest part. And so on.
An omnitasker that can emulate a significant chunk of "all the stuff you just give up and have a human spend an hour on" is precisely what the baker (and countless others) need after hitting that threshold.
Toddlers are a transient disruption to the household logistics operation. They induce novel challenges. Their clothes are have different form factors, and may be dirtied in novel and interesting ways. Their nutritional demands are only partially-overlapping with the preexisting household diet. They introduce hundreds of small toylike objects into the environment. Even their dishware is smaller, flimsier and more numerous, disrupting any existing dish-washing workflows. And the development (or purchase) of permanent solutions to these problems is in some sense wasteful because the toddlers will stop being toddlers before too long, and you will have to pivot again.
The holy trinity of washer, dryer, dishwasher solves a relatively high percentage of household labor but the tail is very long. The purpose of the above digression on toddlers is to illustrate that some household labor disruptions are transient and really not worth developing specific solutions to permanently resolve.
This household example is both a true example of the problem and a metaphor for this type of situation which occurs in all human environments. There is always a long tail. There are always changes to processes with knock-on effects that leave gaps in preexisting workflows. And the solution of final resort has always been human manual labor. There are always transient solutions that require transient workarounds. And this long tail is where it will be enormously helpful to have humanoid form factor robots. This is true for exactly the same reasons that a young parent would appreciate having a humanoid robot in the house, to pick up clothes off the floor and fold clean ones, gather up toys and put them in their appropriate storage bin, and clean the table and floor and nearby walls (and possibly ceiling) after meals.
Actually that's getting interesting. For non-physical jobs we already got significant boost from LLMs. Robots will be another wave when they get cheap enough. For robot like on the video $20K price for mechanics looks achievable.
This is huge for the industry. Smarter Every Day visited a frisbee factory and they had automated a bunch of things. However, every automation point was extremely protected (fenced off) so that a bumbling human couldn’t walk somewhere and get a limb ripped off. If I remember correctly they joked that it was OSHA or something, which it turned out to be.
The effort required to change a process shouldn't be underestimated.
Especially considered that industrial environments are already (a) automated for lowest-hanging fruit things (e.g. moving stuff around at human height) & (b) optimized around human capabilities for the remaining things. Substituting a not-humanlike robot would require reconfiguring a lot of existing automation around it.
If you have "like a human, but costs less" that can be plugged into any existing still-human process? You can literally swap them in.
Eventually we'll get to hyperoptimized machines, but an easier sales story to say "We automate your existing human processes."
Yes there will be use cases but it just seems like a problem looking for a solution most of the time.
Not saying the bot isn’t useful, just that it seems more of a gimmick than a real need in any serious factory / plant.
Cranes, carts, lorries, conveyor belts (with vision), my robot vacuum cleaner, my bread baking machine, a car wash, a dishwasher, the ticket gates at the underground, the coffee machine at our office and so on.
There is value to the human form, our versatility and adaptability.
A machine that replicated a human would have incredible economic value (though not for the people whose jobs it replaced). A machine that exceeded a human in versatility, e.g. by having more arms, even more so.
Yes, there are. But that's not caused by lack of humanoid robots, but because these humans are rediculously cheap.
There's absolutely no way that automation and tech can undercut (effective) slave labor.Though already this is happening: your tshirts and socks are probably not hand-knitted by "Bangladeshi children" or Chinese "prisoners", but by machines tailor-made (pun intended) for t-shirt and sock-knitting. This is happening slowly, and in the supply chain there's still loads of manual labor - cotton picked and processed for these socks is more and more automated, but still requires armies of cheap farm-hands.
And then it's really hard to undercut the price of cheap "western" labor with full automation in many sectors. Part of that is due to some form of 90/10 rule: the last 10% of automation is magnitudes more expensive. And many automation leads to a higher TCO, as the people programming, maintaining and optimizing the robots are far more expensive than an "army" of low-wage workers.
Humanoid robots may be an answer in some future, but currently and in near future will certainly not solve these economics. If ever.
An assembly line with robotic arms has been standard for a long time now. And having many such arms working at the same time is normal. And each robotic arm will be doing one extremely narrowly defined task.
Anything involving autonomous judgment and mobility introduces uncertainty.
Seems at the very least it could have little quad wheel things so it doesn't have to worry about balance as much.
You actually end up running a warehouse with less robots because they can easily be repurposed for other duties.
It looks complicated but is a rather simple mechanical design. There are several motors in the base pulling on cables. The arm itself is just segments, disks, and cables. So all the complexity is in the base, and you can surround that with a metal or concrete box to prevent damage. Arm replacement wouldn't be too expensive.
Tesla is now touting wireless car charging, but that's a lot of power to transmit through air.
Though the demo is awesome, the use case is bad. This "problem" of sorting and moving parts around, has long been "solved". It can be forever optimized, sure. But humanoid robots are definitely not that optimization.
Racking 480V three-phase breakers: https://youtu.be/Rytjdqj_Img
There is a risk of arc flash when you disconnect the bus fingers and also when you reconnect them to the bus, and a robot is easier to replace than a human.
Why does it need to turn its head or it's torso?
Maybe just to make it more humanoid?
It’s the same reason why companies are throwing money at a self driving car which can coexist with manually driven cars, rather than building roads specifically for autonomy
How much do you benefit from 360 vision vs front facing camera.
You also get some relative directional information as well with the current design.
Specialized automation is there, proven, efficient. It can be improved, sure (easier (un)loading?), but I doubt a generalized, humanoid robot is the best way towards these improvements.
I wouldn't really care if the robot was up all night doing dishes as long as they were done in the morning. And, you know, the robot did their work quietly.
Yes, human-looking gait is nice, but it isn't worth anything if it can't translate to real-world settings.
But IMO manipulation is harder than locomotion, both in hardware and software, and neither company is convincingly ahead there. I think the uncut laundry demo from Physical Intelligence a few days ago is better than anything shown by Unitree or BD for manipulation. https://www.physicalintelligence.company/blog/pi0#:~:text=Af....
Haven't the last century of sci-fi books and movies taught you this?
Maybe it's just me..
However, the guy in the video is attacking the robot in the wrong way. He should be using a spray paint can instead of a baseball bat.
Because it's obviously useless and a regression from the kinds of robots we already have.
No. It means it's not impressive.
PS: Isn't that the Lord of the Lithium, the Guardian of the Tunnels has the robot prototypes that not only serve drinks but do the small talk with full self driving AGI already?
It seems like when the robot thought the object was put away (or some safety feedback mechanism of receiving too much force feedback activates), it "relaxes" its actuators and goes limp, say.
Then the stored force from the spring is released. What we're seeing with the jump is the robot rebalancing itself in order to remain upright.
Not a chance. The compartments are clearly just hanging there and not under the kind of tension that could cause something like that.
In this state it doesn't take much counter force at all to imbalance the system. You're also ignoring the momentum of the arm being pushed back. Then it is the supervisory control loop (balancing the robot) that "over reacts" to maintain current position. It is this active control that is responsible for the jolt we see, not the spring itself.
I think a more impressive demo would be last-mile package delivery, since it can climb stairs and operate elevators.
This helps make sure the customer gets what they purchased and helps for QC and recalls. If one batch is bad then this system allows them to pinpoint down to the exact VINs that were affected.
So that's the competition.
Which, humans do at least once a day, among tending to other biological needs as well. So the question is what's a useful run time if there's a bank of swappable batteries it can run back to refill from? Even if it had to go and swap batteries every two hours, for a factory robot that wouldn't be insurmountable.
How many naps are you taking? Normal uptime is at least 16 hours. In special circumstances, much more is possible.
I'd be amazed if the robots are more than 2. Slightly surprised if they're over 1.
It's not exactly comparable.
Oh, and humans have traditionally banded together to punish those who ask for more than 8 hours a day or 5 days a week, and have even historically gotten very murdery over the subject. Buyer beware!
How much PTO does the robot use? Are they a culture fit? Do they have any insights for process improvement?
If robots can replace humans for the most monotonous tasks, I'm all for it, as long as it doesn't hurt living conditions for those humans.
But all these robots are only shown to us in very carefully controlled conditions. If a human was doing what we saw this robot doing in a factory, they'd probably get disciplined unless they increased their speed by 400%. This is state of the art, I presume.
Maybe this kind of thing will be useful some day, but it feels like that day is a long way off.
As far as speed, it depends on what the limitations are. We know electric motors in general are capable of moving faster, so presumably the limits are with computation, to which I'd point out that Moore's law is still hanging on. Compared to Honda's Asimo in 2000 BD's Atlas robot in 2016 was nothing short of phenomenal progress, but that took 16 years.
Hopefully it doesn't take another 16, but we'll see. Once there are useful commercially available humanoid robots on the market, progress is likely to accelerate.
Literally how would it not? Losing your job hurts your living condition, it's not like the warehouse workers will just be given cushier tasks.
Yep, you’re not allowed to own them, only rent them. And there’s no right to repair if you have a defective unit.
Not that that would be a bummer here nor the baseline, but humans can be properly amazing and we keep stretching the possible further and further.
A company might happily have 4 hour shifts to facilitate charging the robot workforce, if that means they dont need to deal with unionizing, smoke-breaking, hung over, late coming, slow working and generally unpleasant humans in their workforce.
It's probably stopping to finish each verse.
The companies buy the robots, and just shoot the employees in the head as they leave the robot store as an act of mercy.
No, the fifth industrial revolution won't leave space for us kowly individuals to participate in I'm afraid.
Let us consider a spectrum of splitting differences, and their extremes:
Consider a factory that runs 3 human shifts, and N full-time labour positions.
In one extreme the company buys for each 3-human-shift position a single robot price.
In another extreme, some workers decide to buy a robot on their own, and hope to send it to the factory.
Now consider 3 robots are bought per 3-human shift position, i.e. one robot corresponding to each human worker. If the robot follows the worker home, it would spend 2 / 3rds of the time at home, and one third of the time at work. So a company could pay a third of each robot for 3 shifts. What is won by whom?
The companies:
1) paid the same amount of money
2) don't need space to store a large stock of backup robots (suppose your robot broke down and needs a few days to get repaired, the company can ask your colleagues from other shifts if they want to rent out one of their robots, so instead of having 2 robots in the household, some will only have 1 for a few days and they will get compensated by whoever was responsible for the damage, if it occurred in job related conditions or leisure conditions)
From the perspective of the employees:
1) Instead of paying 100% of the robot price, you only pay 66.6%
2) Instead of blindly hoping to find an employer where you can send your robot for work, your employer is organizing this for you
From the perspective of the robot makers:
1) Sales triple
2) gather real world data 3 times faster than robot makers not participating in this type of scheme
3) just like the factories, having a larger base of robots means the shot noise on the stream of incoming robots needing a repair is buffered, so capex on the repair centers decrease
From the perspective of jurisdictions, power blocs, and robot warfare:
1) Your economy learns 3 times as fast compared to blind ideological jurisdictions that believe the optimum must lie at the extremes: either company buys 1 robot, or 3 employees each buy their own robot
From a different perspective one can argue much simpler: we already see this in action with company cars, robots for human transportation operated for work part of the time and operated for leisure in the rest of the time, complete with myriad of rules and regulations to determine who pays for repairs and how or when the damage was incurred.
Boston dyns may be building actual products for the US army, but none of their videos is interesting anymore. It s like watching Hollywood stunts
You might not be able to see it if you don't know what to look for, just like how some people seem unable to notice when that ugly motion interpolation is turned on their TV ¯\_(ツ)_/¯
I tend to believe both videos probably are not fake. With speech recognition, text to speech AGI and the advancements in machine learning as applied to robotics, it is not impossible for many or any committed enough group of engineers with financial means to make some notable progress on that front.
So it is not just skills but money too.
Also, I didn't say what they demoed was impossible (other than the unnaturally smooth motion). There's probably good reason they went with a cgi video, but we don't know.
From what I've observed, (so preface the following with "It seems to me that:") BD obviously has a well-honed "playback" tool. They have the ability to command the robot to operate in a new environment, record its own movements to sub-millimeter accuracy, and record the environment (it has a live 3d point map after all.). Then, they load in the recorded environment+motion data, and play it back as a 3D scene. Naturally, this would be needed to analyze the performance and make software and hardware iterations. However, this data is likely also used to re-create a digital scene for the purposes of cgi enhancement of a performance that's intended to be recorded and released to Youtube. Similar to how Favreau uses a giant screen to approximate the lighting on the chroma-keyed subjects[1] as they perform, this data -- a giant shiny rectangle goes here, a window and a flashing blue light bulb there, these steps over here -- combined with accurate camera motion data, is then used to create a digital model with the exact shadows that one would need to create a more realistic "fake video" in the first place. The trick is that Boston Dynamics then takes the original take (or iterations of takes of it running the exact same sequence), and iterate a cgi-enhanced version of their "cool dance video" that the marketing team then signs off on. "Ohh yeah, that's what our vision is. We don't want them to see shaky appendages, or micro-stutters. Add in some extra scuff marks, too! Perfect. Upload it."
[1] https://illumin.usc.edu/the-volume-how-the-mandalorian-revol....
I can see that. Because you're posting incoherent scenarios with no evidence.
Meanwhile we have Spot being deployed in Australian mines for surveillance where the environment is constantly changing. Not sure how they would accomplish this if everything they do is CGI.
Now not so much and I am glad BD isn’t under Tesla umbrella.
I think that if Tesla had bought BD, it may have been more trouble than it's worth. They might have had to fire a lot of the staff, and then I'm not sure what they would have been left with. Tesla is trying to optimize for production cost and mass manufacturing whereas BD doesn't care about cost. It also seems to me like BD hasn't done much in the AI realm. They've focused a lot more on the locomotion and have been using classic robotics techniques.
They need to rock and sock each other already though.
unfortunately the world is full of sooooo many much smarter slaves to exploit
https://bostondynamics.com/industry/manufacturing
And it's one thing to have a cute demo showcased under ideal scenarios and another to have it deployed in mission critical environments around real people.
We have had robots stamping parts for years. There is no need for a humanoid style configuration. Meanwhile what is needed and far more complex are robots that can interact with unpredictable people in unpredictable environments.
Boston Dynamics has already demonstrated they can do this.
Machine presses still have human operators somewhere, even though theres machines and robot arms involved. eg the end of that figure one video.
Sure, a dog sized object autonomously not running into objects, some of which can move is an achievement, but it's still just a remote control camera.
On what do you base that?
Tesla is so far behind BD, who are behind the Chinese
My money is on the Chinese
* From a business point of view. This is an incredible techinal achievement, I don't want to sound like this is not impressive. But it seems that every new development seems to focus on how they can replace humans or be better at or do things that we usually do.
They take long bathroom breaks (and think you don't notice, because they think you're stupid.) They steal. They fight with each other and need managing aka children's therapist for their bullshit. Which never stops. You can stop wasting your time dealing with the "human touch" and get back to what you really want - making more widgets so you can sell more widgets so you can make enough money for that kitchen remodel/winter/summer home/yacht/European vacation/jet.
Fire them and replace them with RobotWorker. It doesn't get drunk and cause fights or HR incidents because it can't keep it in it's pants. They'll work all through the night and through every holiday, without the same trouble of having a second and third shift. You don't need to follow OSHA with these things, though you still don't want to damage them - there's a support contract but that's unnecessary downtime for you, and you don't want that. Still, you can just replace a robot's arm. Just imagine the lawsuits when that happens to a human employee.
blahblahblah. I'm sure you can come up with more.
Reliable production numbers are also a thing.
Might you not prefer working with a colleague who brings skills and knowledge to the work environment, understands the business, is motivated to improve the operation, and can respectfully discuss the challenges that you face as a team?
Someone who, if treated fairly and with respect, may help build success as your co-worker and perhaps even community as your neighbour?
In exchange for a decent wage and an affordable sandwich, of course.