The tech is amazing. I cannot, however, understand how they expect widespread commercial use. I am an engineer. Among many things I do, one of them is robot automation for industrial applications. I would anticipate the real-world cost of these to be rather extreme. Much, much simpler articulating arm type units are already expensive, complex, and challenging to integrate into real applications. The more complex the system is, the more it will break. The more maintenance. The more specialized the system, the more expensive the parts and technicians are to keep it going. At the end of the day.....to be interesting it MUST make more money than it costs. The cost is the purchase price plus the real operating costs. The real operating costs will certainly be substantial.
I feel so sorry for Asimo. He's like a child star that's been forgotten in adulthood. I'll never forget you, Asimo. I'll always cherish the time that I spent with you at Disneyland 16 YEARS AGO.
Yes, there has been many good attempts for robot bodies. But they have all lacked a brain. That's what has changed now. We see it's now possible to get it one of those! To me these early attempts seemed to be pretty futile, at that time.
@@larsnystrom6698 Robots don't have a brain. They have software that selects the optimal procedure to perform a task. But still lacks the capabilities to improvise outside its software, once the working environment it operates, wasn't planned in its code. Don't fall for all the AI claims.
@@Roddy451 Imagine if the Tesla Bot had the same movement fluidity as capabilities, as past Honda and Toyota humanoid robots? Not a surprise the Stock would be in rush pumping, as delusional claims about their potential capabilities.
@@BillAnt Depends how free you need it to be. It could be just plugged in depending on the task. But I'd imagine these robots would be very mobile and go everywhere. We already have purpuse built assembly robots, that isn't the task these need to do. So what task would they do?
What has Tesla lied about? The Semi and Roadster are super late. People called the Model Y vaporware too until it because the worlds best selling car. Full Self Driving took a long time, but now it's better than a human driver. "Tesla, making the impossible merely late". Tesla and SpaceX have the worlds best engineers, period, end of story.
@@MichaelBushey full self driving still isn’t here, to name but one. Elon is a snake oil salesman, some see through his bullcrap (me) some don’t (you).
@@onetwothreefour-s1n Thank you for your concern. My family and I are doing quite well. What really helps me take my life back is FSD. I estimate I've driven over a million miles in my life and having the car drive now is a huge help. I asked Tesla for help, and they delivered.
@@somerandomperson4143 I think it was a dig at the Tesla robot that was folding a shirt, because that video was basically faked as a human was connected off camera folding the shirt and the robot merely mimicking the movements. Confirmed by Elon to not be autonomous. The usual scam like the original autopilot video that was faked.
@@lamsmiley1944 lol. You don't have FSD v12.3.4, do you? Do you not understand how/why the Waymo system architecture is utterly inferior to Tesla's AI? Waymo works OK in a couple of fully mapped out areas but cannot be scaled. Tesla system is AI and can drive pretty much anywhere. I love not needing to drive anymore, FSD is a BETTER driver than the average human.
@@MichaelBushey Can’t be scaled? They’re owned by a company that already has already mapped most of the world with street view. That was for a free mapping service.
@@MichaelBushey - Two different systems, Waymo is more accurate but cautious, while Tesla is not as accurate but can learn as it goes. No need to mock either system, unless you wanna come off as a cult member.
I would say this is not the best argument. I would say from an engineering standpoint of why they made the switch is very critical to having an understanding what really made Boston dynamics to make this change. A hydraulic system can give more force than compared to electric motors. As it is just a gas chamber being compressed back and forth. Although it can generate much more force it is extremely slow. An example would be an office chair, hydraulic systems are all the same as gas is compressed in a chamber and exchanged between two spaces. However change from out to in is not instant which can be shown from early versions of robots with clunky movement. Which is the computer doing calculations to bring the position back to origin to create the next set of calculations without error. So the question is why they hadn't used electric motors beforehand, its not like the company was trying to be more economically efficient but that electric motors was not a viable option for the previous model. Electric motors could not generate enough torque or force to support the old atlas. My proof comes from the completely different torso sizes. The old atlas housed a much more beefier processing compartment and database because of the physics models they render. My proof would be something like the new nvidia simulation training computer. They were pretty much running a monster machine at the time inside a computer; which answers the questions of why they would rely on a hydraulic system to support the robot. However new ai models have completely changed their workflow and made their physics models more seamless. No longer would they need to run specific calculations in real time with a beefy machine but are able to output information from their physics models. A huge benefit to the overall robotics space. As seen from previous generation of robots, calculations were being done on the robot itself having a huge strain of weight, energy, and time. However private ai models like even lama with a bit of tweaking with their old math models is more than enough to create great output that works well enough to interact with the world with significantly less real-time rendering involved. My experience would be with the spot robot last year, the quadrupedal movement was hard enough to program and calculate for, I am not even able to comprehend the amount of calculations and real time data the old atlas had to generate constantly to perform. As quadrupedal architecture has a larger degree of acceptable error as compared to a bipedal system. So in this aspect Boston dynamics might be miles ahead of other robotic companies as of this moment because they have and had so much more time to generate all the edge-cases and simulations for movement. However that is also overlooking on the whole goal of ai which is reduction in overall compute and more assumptions. Also, you think people mine coal with pickaxes?, we use machinery, maybe in those third world countries but what third world country mining facility would pay millions instead of a couple dollars for some labor. Also from an engineering design point what is the point of a robot versatile at a lot of things instead of just one. Such a wide net to have errors and problems, how could this thing beat a fried rice robot when the thing is literally made to make friend rice by spinning with a fraction of the cost, less maintenance, and a less chance of breaking or going wrong. Don't delete this time?
With hydraulics you just need one motor to drive the hydraulic pump. With electric you need one motor for each joint. Hydraulics are lighter for this application (for the same force).
Looks good - wonder when they will be used in Hyundai/Kia factories. US Dept of Defense no doubt is slavering over the prospect of future combat soldiers.
It's definitely impressive. But keep in mind that Boston Dynamics has always been more of a research lab and have never really been able to make any money. They continue to record huge losses every year while creating a lot of cool videos and demos that get a lot of social media attention. But it isn't an insignificant leap to go from cool demos to an actual scalable product that you can sell and support in the thousands or millions.
How long does Atlas last on a charge? Doesn’t seem like there’s a lot of space for batteries on it. And, what can it actually do, other than just ambulate?
You know, its funny you mentioned the (inevitable) possibility of these humanoid robots one day being developed as future partners for intimacy, speaking as a man myself, i will say that this will absolutely be the more popular application for this technology, and thats not a joke. I can wholeheartedly promise you that "lovebots" are gonna be a MASSIVE multi-billion dollar industry on its own, and i, for one, am looking forward to it.
Oh man, it looks so cool and creepy at the same time, Crool? Seems like it could be a tech horror concept, imagine that head lighting up in the darkness, walk towards you as it starts turning red, then you hear HAL9000's voice coming from it. Can't wait to see more advances by Boston Dynamics!
I can't wait for the future! Love all the hype around robots these days. Honda made a really excellent robot about two decades ago but it was too early given the limitation of the technology available. Key technologies improvements such as battery, motors, weightless components, manufacturing, llm, and major ability to compute like never before, makes the transition from cool gadget to a real application smooth. industrial robots have already changed manufacturing for the better, but 😊now humanoid robots that learn and are not task specific will change the world, slow at first and then with incredible pace.
I see the argument for robots like Spot and Stretch. Spot is an inspection bot designed to prevent the need for massive sensor arrays with multitudes of failure points in both hardware and software. Spot is a compact local solution and it is the only failure point. Stretch has less going for it, but fast, cheap deployment is a compelling argument for companies that don't already have big investments in automation, such as packing, moving, and storage companies. Atlas is weird. It's not really competing for the same customers as Optimus and Figure-1. It's bulky, harder to manufacture, and very powerful, unlike the other bots. This makes it more suited for heavy work, such as a factory, power plant or oil rig. An argument I would make for it for use in a factory is a similar argument I would make for Spot. Instead of having many job-specific robots that would add up in price maintenance cost over time, Atlas would be a general solution that could be reasonably good at most tasks. This would mean less points of failure, like with Spot. Of course, it wouldn't be good at all tasks, such as anything involving precision. If you were a company investing in automation but didn't want to fiddle around with a bunch of specialty machines with different methods of operation, software packages, and safety precautions, you could instead deploy a more versatile solution. Whether this is how it will turn out is another question entirely. Optimus and Figure-1 look to be geared towards more light-work tasks, like stocking, surveying, precision, and human assistance. Basically they are designed to be a lot closer to humans. More white-collar. They are easy to deploy AND operate (At least F-1, Optimus's method of operation is obscure and we've only seen evidence of live-mocap and preprogramming currently). They solve the ASIMO issue of needing an engineer around at all times tending to it. But I will say that ASIMO got something right that these models have failed to deliver. People are far more afraid of these robots than of ASIMO, simply because they are bigger and more human-like. ASIMO looks like furniture. Or a cartoon character. Or our childhood idea of what a robot would look like. Basically, its designed to be friendly-looking. Current models of humanoid robots don't really take this into consideration, and if they want to deploy these around people, that might put a damper on things. Overall, it's just a matter of waiting for the market to mature. We'll see if these are as applicable as these companies say they are.
Problem is that Atlas has a complete different purpose than Optimus. If anyone else then Elon would build Optimus, it would be interesting to see what they'll do in the future. But since its Elon we'll have to wait 10 years first to make any assumptioms.
I think they're going to have problems switching from hydraulic. Little electric motors could be limited by power output and overheating. On the other hand, no hydrolic hoses with hundreds or even thousands of psi of scalding hot oil.
BDs tech is amazing, but I don't see a large commercial need for "humanoid" robots. They may have a place, but there are already wheeled and tracked units that are filling the space. We shall see what the future brings. That headline is just Apple trying to stay relevant. I suppose they can burn the money that would have gone into their applemobile project, but I think it will have the same outcome. It's like Starship. There has to be a market, a need, and there isn't one.
The Tesla Bot was always a Stock pumping machine, with delusional capability claims. Plenty of other Automakers had robots in the past. And there are far more efficient robots in assembly lines without a humanoid design.
Yes automakers have industrial robots. But they also have 10'000 humans in each factory doing simple repetitive jobs. What do you think they should do about that? Nothing? It's not like a Tesla factory is less automated than others. But industrial robots can't only do everything, can they?
@@larsnystrom6698 Also the Bot for household applications is like hammering a nail with a laptop. And in terms of emotional appeal, it is like having a sex doll, instead of real Human affection.
@@larsnystrom6698 Toyota as Honda had in the past humanoid robots for technological demonstration. Being way more fluid in movement, as doing tasks than the Tesla Bot. But even so, it didn't pump their Stock with delusional claims.
They need to get robots for kids first... small robots that don't need too much energy and that can operate like r2d2. a little assistant that can also mentally stimulate the children and help them learn basic skills. from there robots can advance toward more serious roles such as caretakers for the elderly or disabled... but I'm still not sure robots will catch on that look more like c3po as opposed to r2d2. I mean walking on 2 legs is helpful on some terrain but I feel like it is not the most energy or cost efficient option. seems very complex compared to a four leg version or better a four leg version that also has wheels. combine wheels with legs is best.
Isn't Boston Dynamics already acquired by Google X (now Alpha X) in 2013? An IPO won't exist unless they spin it off as a separate company. The robot is impressive. I will be OK with getting one as the security guards for the home. :)
The Tesla robot is more or less Animatronics, someone is controlling it and it is just mimicking their movements. I'm not sure about the BD robot. I don't see any of these robots being owned by the public, they will be used by businesses to do things humans can't or probably shouldn't do, and they will be very expensive..
You haven't thought thing through! What we have seen from Optimus is just mostly testing the hardware. With exception of that sort8ng demo, which was a test of the learning meth9d. I think Tesla is now pretty fed up with copying them. So they won't show what they are doing to that degree. But Atlas isn't copying the mechanics. They are doing there own thing, except for going electric. That's something Tesla inspired them too do.
@@pdavidson33 the T shirt folding video was a fake. There's a human attached doing the movements and Tesla accidentally left him in frame so they had to admit it was not autonomous.
I think there are decades before any of this becomes a household product. But it is great that even so, some companies keep innovate and work on it, if not it will never happen. ...Tesla though, I feel like is developing the robot for other reasons. 1. like the cybertruck, the employees of tesla need to do that the stupid great leaders says, 2. they need new stuff to pump the stock. Which is also why you cannot trust any of the demos tesla show, they use a lot of trickery and rendering and cutting.
I think it's funny that everyone start doing electric humanoid robots after that Tesla presented their concept. I'm sure some of them will be profitable too. Maybe not as profitable as Tesla, though. Let's review this subject in 2 - 3 years!
Nobody said Tesla has de monopoly, Tesla are the first comers, and much of their products are copied or based on their ideas, Figure AI, Atlas 2, all are electric and no mhydrolics, also with Brains, and also all started after Tesla Optimus presentation, or at least, started taking humanoid robots seriously.
@@Voltaje_YT First, I was joking. But I was making a point that Tesla takes credit for things and people give them a pass. Honda introduced ASIMO in 2000 and it was a humanoid robot that used AI. Tesla copied and improved on ASIMO. You dig deeper, and you will find multiple papers on others integrating neural network and robotics before Tesla. Tesla was better at marketing their concept and stealing the limelight from other companies who have been working on this for decades.
@@larsnystrom6698 Look at the Wikipedia page on humanoid robots. There were plenty of companies using electric actuators and integrating AI years before Tesla. Tesla just had the marketing and the fanboys to get the message out to the rest of the world. But they were not the first. What's funny is how people are giving Tesla credit for things they copied. I am not mad at them because now billions are pouring into researching and developing this technology. Maybe we will have a consumer humanoid robot within 10 years.
@@larsnystrom6698Boston dynamics has been messing around with humanoid robots for years the first few versions of atlas’s was just research and development. And this current model clearly has been in the works for awhile
Also Boston D will not be making consumer robots at scale. They will be more for commercial and military use. Diffrent base. I think when it comes to warehousing, Teslas humanoid bot will be the winner, but BD will trash it when it comes to specific applications.
Why be so short sighted? This version would be perfect for Space exploration on the moom. Eploring the lava tubes for possible human occupation in the future. Whe can actually get early tests on human designed landing systems. They don't have to be fully human rated (cheaper to strart). They just have to perform successful on point landings. Tesla claimed their capsule could do a powered landing of a human crew back on earth. Put that claim to the test but on the moon. Once there they can prepare a landing field for future missions, deploy a "crane" for lifting up and down equipment into the lava tubes. Power can come from RTG's where the robots just plug in for a charge on a regular schedule. The same goes for Mars. They could assemble a rocket launch pad and build a "Sounding" rocket to launch our samples back from Mars. That would be truely reusable, our skycrane technology is very accurate. Rather than crashing them off in the distance. We can possibly land them and reuse some of the parts for building the return rocket. Less risk and way more ROI going that route vs. a factory of humanoid robots working 24/7 producing whatever for "humans" to buy? With what money? If even the factory jobs are fully automated like Elon wants. The one place the human race will always have a labor shortage is in space. These robot helpers do not have to be fully autonomous. Tell it what to do, it does it. Humans remotely control its actions. Sure we can create a library of "responses" or "actions" for repetitive work. Look beyond for their usefulness. We need space rated versions of this ASAP!
There's always been talk that we should just send robots and rovers for space exploration / colonization. But it doesn't really excite the public unless we send people. However, we could send people and leave the robots behind to do the real work :)
@@realchris Send them to set up the landing base and mining equipment. The pictures of them doing work is exciting for those that want to code. The hardware for those that love design. Do you see where I am going? These are our humanoid assistants not replacements. That is how it should be talked about for science. Elon Musk wants slaves, be they mechanical or not. He wants it all for himself. Money, fame, Cult leader devotion.
Yes but Teslabot at this stage is pretty lame. It can barely walk very slowly on flat smooth surface. is that the best that Tesla can do? Also remember there is minimum 3 to 22 minutes (round trip 6 to 44 minutes) time for radio communication between Earth and Mars. Tele-operation will not work on Mars where robot can fall down a cliff before a human operator has chance to react. Also, humanoid robot may not be best for space operation. Humans are limited to four limbs and five fingers, but a robot can have any kind of appendages and any number of them. Humanoid robots are primarily useful for navigating human settings. outer space uses will have vastly different requirements and robots with much more versatile design may be required.
@@springer-qb4dv I can see you are like Elon. Short sighted, and misinformed. I will just say this. Our current robotic probes all work "Tele-operation". We send commands, they reply with results. The ability to self right is shown with Atlas v.2 in this video. Tesla bot is vaporware imo, more like the puppet arm from the Terminator movies. Take the arm and hand for instance. In the scene when Arnold cut off the skin of his arm and hand. You see the actions of the hand and fingers, they are all wire and pulley operated by a group of puppeteers off screen. Today it can be operated by servos pulling the wires. Slow and weak. Just another Elon "Grift" for the gullible.
It has some moves! But I'm pretty sure they are scripted. But we see the inheritance from Atlas! It would be a better military robot than anything else we have seen. But that probably requires more endurance than is yet possible. So not a market for that yet. I would guess it will be pretty expensive to manufacture. So perhaps not for a mass market. It remains to see how good it's brain will be. Cooperation with someone? I'll guess Optimus will be more energy efficient, and much cheaper. We can be pretty sure they know how to get it a brain too. So, no it won't bankrupt Tesla. But maybe give them some healthy competition, so they have to strain themselves a little more. Isn't it funny that everyone now build electric humanoids? Tesla, or Musk, really drives much of the industry thinking!
Do you have a video where you outline your personal feelings on Tesla? Everything that I've seen appears negative, or maybe it's just Elon you take issue with. It's really no different than SMR and his constant pumping no matter what. Thanks
I watched this video on X last night and couldn't tell if it was real or AI - if real, why did they make it stand up in the creepiest way possible? Just to troll Elmo?
Yep, your video is about Tesla, or how? Why is Tesla in danger, and not all others doing the same thing? Clickbait, BS, and a lot of ....BS! You people cannot see that this robot is just a technological demonstrator with no real-life meaning (seems like Boston Dynamics is in panic and made this)? Who needs those joints, and for what activity/purpose? PS: Logic, why should someone buy an acrobatic robot with no dexterity in fingers (Is it just me, or has this robot already arthritis, as hands are fixed in that characteristic position!)? Is its purpose to climb a rope, or play in a Circus?
@@Xanthopteryx tesla is not betting much on the most amazing humanoid.. Its focus is on affordability. A strategy that has helped their car business.. Making cars affordable by scalable tech and removal of nkn essential parts..
It appears the hydraulic version was never built to scale into production whereas the electric now has this in mind. As to the humanoid form being developed I believe the rational behind that is the ability to offer something commercially that can be integrated into a environment which is currently suited for humans but moreover to pique public interest in order to further the endeavor. It would seem to me if your ambition is to replace humans for commercial applications, in many instances you would have full control over designing the environment the robot operates in and build task specific robots as part of the facilities automated eco system. This brings into question the importance of humanoid form to commercial requirements and to what degree these humanoid demonstrations are for the purpose of gaining interest. I would say the objective is in seeing how adaptable robots are in operating in complex environments without requiring someone to teach it each and every step of a task as well the ability to analyze situations which don't cleanly fit into their expectations and having them take the appropriate actions within the context of what they were assigned to carry out. BD's short demo captures this concept well. See it as human but also see it as something other than human adapting in a way a humans can't.
Yes....no one asks who is going to build them. The jump between prototypes and mass production is something many underestimate. Tesla will pump them out in the millions.....
Hyundai owns 80% of Boston Dynamics, so Hyundai pretty much has total control of BD. Mechanics of robots is hard, but mechanics is trivial compared to designing robot AI which will allow a robot to independently engage and work with dynamic changing environment full of people and other machines. Tesla is nowhere with it's optimus Teslabot. Musk's touting of optimus as serious product this or next year is another one of his whopper lies. Musk should have bid for BD and bought it while it was available but apparently Musk didn't realize he needed Teslabot back in 2021! LOL
The tech is amazing. I cannot, however, understand how they expect widespread commercial use. I am an engineer. Among many things I do, one of them is robot automation for industrial applications.
I would anticipate the real-world cost of these to be rather extreme. Much, much simpler articulating arm type units are already expensive, complex, and challenging to integrate into real applications. The more complex the system is, the more it will break. The more maintenance. The more specialized the system, the more expensive the parts and technicians are to keep it going.
At the end of the day.....to be interesting it MUST make more money than it costs. The cost is the purchase price plus the real operating costs. The real operating costs will certainly be substantial.
Per Elon they can fold towels with some one guiding it.
I feel so sorry for Asimo. He's like a child star that's been forgotten in adulthood. I'll never forget you, Asimo. I'll always cherish the time that I spent with you at Disneyland 16 YEARS AGO.
I have not forgotten Asimo from Honda, and I post about it as much as I can. But you are correct; it tends to be lost in the limelight.
Asimo walks like baby with its diaper full of sh*t. ha-ha-ha
Yes, there has been many good attempts for robot bodies. But they have all lacked a brain.
That's what has changed now. We see it's now possible to get it one of those!
To me these early attempts seemed to be pretty futile, at that time.
@@larsnystrom6698 Robots don't have a brain. They have software that selects the optimal procedure to perform a task. But still lacks the capabilities to improvise outside its software, once the working environment it operates, wasn't planned in its code. Don't fall for all the AI claims.
@@Roddy451 Imagine if the Tesla Bot had the same movement fluidity as capabilities, as past Honda and Toyota humanoid robots? Not a surprise the Stock would be in rush pumping, as delusional claims about their potential capabilities.
I like it because it utilises a range of movement that we can't. It doesn't just want to imitate human movement.
Freakish looking/moving for sure. The big questions is how long can it powered before recharging, strength etc.
@@BillAnt Depends how free you need it to be. It could be just plugged in depending on the task.
But I'd imagine these robots would be very mobile and go everywhere. We already have purpuse built assembly robots, that isn't the task these need to do.
So what task would they do?
The humanoid f9rm from "first principles". Not restricted to just simulate humans, but to emulate them.
BD makes the Trashlabot look like a typewriter.
The thing is with Tesla is it’s after getting to the stage that you doubt what they say as they’ve lied so many times before.
What has Tesla lied about? The Semi and Roadster are super late. People called the Model Y vaporware too until it because the worlds best selling car. Full Self Driving took a long time, but now it's better than a human driver. "Tesla, making the impossible merely late". Tesla and SpaceX have the worlds best engineers, period, end of story.
@@MichaelBushey full self driving still isn’t here, to name but one. Elon is a snake oil salesman, some see through his bullcrap (me) some don’t (you).
@@MichaelBushey I hope you can get the help you need. There are many support groups that can help you take back your life. You just need to ask.
@@onetwothreefour-s1n Thank you for your concern. My family and I are doing quite well. What really helps me take my life back is FSD. I estimate I've driven over a million miles in my life and having the car drive now is a huge help. I asked Tesla for help, and they delivered.
Basically almost everything at some point .
The big question is, can they fold a towel unaided?
😂😆
Probably because it has hands and we are talking about Boston dynamics here .
@@somerandomperson4143 I think it was a dig at the Tesla robot that was folding a shirt, because that video was basically faked as a human was connected off camera folding the shirt and the robot merely mimicking the movements. Confirmed by Elon to not be autonomous. The usual scam like the original autopilot video that was faked.
Exactly what I was going to say. If this thing can do laundry I'll take out a mortgage to buy it.
@@bccsivxx-xxivvii 😂😆
Boston Dynamics is the real deal.
One of the ex Google moonshot projects outperforming Tesla. Waymo is also years ahead of Tesla “autopilot”.
Of course, with major DOD funding it's easier to built cool robots. ;)
@@lamsmiley1944 lol. You don't have FSD v12.3.4, do you? Do you not understand how/why the Waymo system architecture is utterly inferior to Tesla's AI? Waymo works OK in a couple of fully mapped out areas but cannot be scaled. Tesla system is AI and can drive pretty much anywhere. I love not needing to drive anymore, FSD is a BETTER driver than the average human.
@@MichaelBushey Can’t be scaled? They’re owned by a company that already has already mapped most of the world with street view. That was for a free mapping service.
@@MichaelBushey - Two different systems, Waymo is more accurate but cautious, while Tesla is not as accurate but can learn as it goes. No need to mock either system, unless you wanna come off as a cult member.
If somebody built Bender from Futurama I'd buy one as that would be a lot of fun 😊
I would say this is not the best argument.
I would say from an engineering standpoint of why they made the switch is very critical to having an understanding what really made Boston dynamics to make this change.
A hydraulic system can give more force than compared to electric motors. As it is just a gas chamber being compressed back and forth. Although it can generate much more force it is extremely slow. An example would be an office chair, hydraulic systems are all the same as gas is compressed in a chamber and exchanged between two spaces. However change from out to in is not instant which can be shown from early versions of robots with clunky movement. Which is the computer doing calculations to bring the position back to origin to create the next set of calculations without error.
So the question is why they hadn't used electric motors beforehand, its not like the company was trying to be more economically efficient but that electric motors was not a viable option for the previous model. Electric motors could not generate enough torque or force to support the old atlas. My proof comes from the completely different torso sizes. The old atlas housed a much more beefier processing compartment and database because of the physics models they render. My proof would be something like the new nvidia simulation training computer. They were pretty much running a monster machine at the time inside a computer; which answers the questions of why they would rely on a hydraulic system to support the robot. However new ai models have completely changed their workflow and made their physics models more seamless. No longer would they need to run specific calculations in real time with a beefy machine but are able to output information from their physics models. A huge benefit to the overall robotics space. As seen from previous generation of robots, calculations were being done on the robot itself having a huge strain of weight, energy, and time. However private ai models like even lama with a bit of tweaking with their old math models is more than enough to create great output that works well enough to interact with the world with significantly less real-time rendering involved.
My experience would be with the spot robot last year, the quadrupedal movement was hard enough to program and calculate for, I am not even able to comprehend the amount of calculations and real time data the old atlas had to generate constantly to perform. As quadrupedal architecture has a larger degree of acceptable error as compared to a bipedal system. So in this aspect Boston dynamics might be miles ahead of other robotic companies as of this moment because they have and had so much more time to generate all the edge-cases and simulations for movement. However that is also overlooking on the whole goal of ai which is reduction in overall compute and more assumptions.
Also, you think people mine coal with pickaxes?, we use machinery, maybe in those third world countries but what third world country mining facility would pay millions instead of a couple dollars for some labor.
Also from an engineering design point what is the point of a robot versatile at a lot of things instead of just one. Such a wide net to have errors and problems, how could this thing beat a fried rice robot when the thing is literally made to make friend rice by spinning with a fraction of the cost, less maintenance, and a less chance of breaking or going wrong.
Don't delete this time?
With hydraulics you just need one motor to drive the hydraulic pump. With electric you need one motor for each joint. Hydraulics are lighter for this application (for the same force).
I like your calm demeanor. Great microphone btw. You earned a new sub today.
Welcome!!! :)
Looks good - wonder when they will be used in Hyundai/Kia factories.
US Dept of Defense no doubt is slavering over the prospect of future combat soldiers.
It's definitely impressive. But keep in mind that Boston Dynamics has always been more of a research lab and have never really been able to make any money. They continue to record huge losses every year while creating a lot of cool videos and demos that get a lot of social media attention. But it isn't an insignificant leap to go from cool demos to an actual scalable product that you can sell and support in the thousands or millions.
I am going to miss hydraulic Atlas, but I am excited for this new Atlas
How long does Atlas last on a charge? Doesn’t seem like there’s a lot of space for batteries on it. And, what can it actually do, other than just ambulate?
"Do the dishes, laundry, vacuuming please."
I can see the market for that, but I can get a human based service for that at a pretty reasonable price.
You know, its funny you mentioned the (inevitable) possibility of these humanoid robots one day being developed as future partners for intimacy, speaking as a man myself, i will say that this will absolutely be the more popular application for this technology, and thats not a joke. I can wholeheartedly promise you that "lovebots" are gonna be a MASSIVE multi-billion dollar industry on its own, and i, for one, am looking forward to it.
Oh man, it looks so cool and creepy at the same time, Crool? Seems like it could be a tech horror concept, imagine that head lighting up in the darkness, walk towards you as it starts turning red, then you hear HAL9000's voice coming from it. Can't wait to see more advances by Boston Dynamics!
I can't wait for the future! Love all the hype around robots these days. Honda made a really excellent robot about two decades ago but it was too early given the limitation of the technology available. Key technologies improvements such as battery, motors, weightless components, manufacturing, llm, and major ability to compute like never before, makes the transition from cool gadget to a real application smooth. industrial robots have already changed manufacturing for the better, but 😊now humanoid robots that learn and are not task specific will change the world, slow at first and then with incredible pace.
Your tile made me laugh for a long time.🤣🤣
I see the argument for robots like Spot and Stretch. Spot is an inspection bot designed to prevent the need for massive sensor arrays with multitudes of failure points in both hardware and software. Spot is a compact local solution and it is the only failure point. Stretch has less going for it, but fast, cheap deployment is a compelling argument for companies that don't already have big investments in automation, such as packing, moving, and storage companies.
Atlas is weird. It's not really competing for the same customers as Optimus and Figure-1. It's bulky, harder to manufacture, and very powerful, unlike the other bots. This makes it more suited for heavy work, such as a factory, power plant or oil rig. An argument I would make for it for use in a factory is a similar argument I would make for Spot. Instead of having many job-specific robots that would add up in price maintenance cost over time, Atlas would be a general solution that could be reasonably good at most tasks. This would mean less points of failure, like with Spot. Of course, it wouldn't be good at all tasks, such as anything involving precision. If you were a company investing in automation but didn't want to fiddle around with a bunch of specialty machines with different methods of operation, software packages, and safety precautions, you could instead deploy a more versatile solution. Whether this is how it will turn out is another question entirely.
Optimus and Figure-1 look to be geared towards more light-work tasks, like stocking, surveying, precision, and human assistance. Basically they are designed to be a lot closer to humans. More white-collar. They are easy to deploy AND operate (At least F-1, Optimus's method of operation is obscure and we've only seen evidence of live-mocap and preprogramming currently). They solve the ASIMO issue of needing an engineer around at all times tending to it. But I will say that ASIMO got something right that these models have failed to deliver. People are far more afraid of these robots than of ASIMO, simply because they are bigger and more human-like. ASIMO looks like furniture. Or a cartoon character. Or our childhood idea of what a robot would look like. Basically, its designed to be friendly-looking. Current models of humanoid robots don't really take this into consideration, and if they want to deploy these around people, that might put a damper on things.
Overall, it's just a matter of waiting for the market to mature. We'll see if these are as applicable as these companies say they are.
It's called the Canny Valley and that's not creepy that's full-on Exorcist.
Wow. You're really on top of it! Awesome. You're my new favorite channel.
Appreciate you coming back! You should check out my book also :)
🎉🎉 very cool, I always learn new stuff here
Reminds me of Nightscope, robot, and AyeEye drone maker. They teamed up with Dragonfly.
It has that atomic Fallout 4 feel to it. like what advance robots would look like in the 80s if they had the tech.
For unknown terrains, walking on 4 limbs seems more flexible than 2 limbs.
I want a spot dog so badly it's nuts.
Articulation Chris! Impressive, nevermind Elon versus Angry Venezuelan dude, let's see Optimus get torn limb from limb by The Boston Bad Bot 🤣
Problem is that Atlas has a complete different purpose than Optimus. If anyone else then Elon would build Optimus, it would be interesting to see what they'll do in the future. But since its Elon we'll have to wait 10 years first to make any assumptioms.
They've had a 15 year advantage in robotics but now there is a lot more competition. They had like 100 chances to take over but they failed.
the turn of the upper body part reminds me of ig-11 in mandalorian :D
I hadn't thought of that, but yes it can move like that :)
I think they're going to have problems switching from hydraulic. Little electric motors could be limited by power output and overheating. On the other hand, no hydrolic hoses with hundreds or even thousands of psi of scalding hot oil.
my knees felt that when the atlas broke, good to know tehy can blow out their knees too! haha
If it can build iPhone without jumping or lifting by 30%, that’s a milestone.
BDs tech is amazing, but I don't see a large commercial need for "humanoid" robots. They may have a place, but there are already wheeled and tracked units that are filling the space. We shall see what the future brings. That headline is just Apple trying to stay relevant. I suppose they can burn the money that would have gone into their applemobile project, but I think it will have the same outcome. It's like Starship. There has to be a market, a need, and there isn't one.
The Tesla Bot was always a Stock pumping machine, with delusional capability claims. Plenty of other Automakers had robots in the past. And there are far more efficient robots in assembly lines without a humanoid design.
Tesla is a meme stock and Musk just keeps announcing new ventures to stop the price from crashing. I think the house of cards is about to all crumble.
Yes automakers have industrial robots. But they also have 10'000 humans in each factory doing simple repetitive jobs.
What do you think they should do about that? Nothing?
It's not like a Tesla factory is less automated than others. But industrial robots can't only do everything, can they?
@@larsnystrom6698 Humans are versatile, humanoid robots not so much. Way more efficient as cost effective to design a robot focusing on the task.
@@larsnystrom6698 Also the Bot for household applications is like hammering a nail with a laptop. And in terms of emotional appeal, it is like having a sex doll, instead of real Human affection.
@@larsnystrom6698 Toyota as Honda had in the past humanoid robots for technological demonstration. Being way more fluid in movement, as doing tasks than the Tesla Bot. But even so, it didn't pump their Stock with delusional claims.
They need to get robots for kids first... small robots that don't need too much energy and that can operate like r2d2. a little assistant that can also mentally stimulate the children and help them learn basic skills. from there robots can advance toward more serious roles such as caretakers for the elderly or disabled... but I'm still not sure robots will catch on that look more like c3po as opposed to r2d2. I mean walking on 2 legs is helpful on some terrain but I feel like it is not the most energy or cost efficient option. seems very complex compared to a four leg version or better a four leg version that also has wheels. combine wheels with legs is best.
Robot walk like an old man looks like it was about to fall over
It's light years ahead of the Tesla bot. Great company all work and no hype unlike Tesla.
Wow. So advanced.
lol, it looks like it can be very compact for shipping, say to a war zone
😂😆
Isn't Boston Dynamics already acquired by Google X (now Alpha X) in 2013? An IPO won't exist unless they spin it off as a separate company. The robot is impressive. I will be OK with getting one as the security guards for the home. :)
The Tesla robot is more or less Animatronics, someone is controlling it and it is just mimicking their movements. I'm not sure about the BD robot. I don't see any of these robots being owned by the public, they will be used by businesses to do things humans can't or probably shouldn't do, and they will be very expensive..
That's not true
@@pdavidson33 you can see the guy controlling it in some of the videos Tesla put out.
@@LordandGodofRUclips But Musk told him it was really advanced, and we all know that he never lies. Right? It’s not like he’s a compulsive liar.
You haven't thought thing through!
What we have seen from Optimus is just mostly testing the hardware. With exception of that sort8ng demo, which was a test of the learning meth9d.
I think Tesla is now pretty fed up with copying them. So they won't show what they are doing to that degree.
But Atlas isn't copying the mechanics. They are doing there own thing, except for going electric. That's something Tesla inspired them too do.
@@pdavidson33 the T shirt folding video was a fake. There's a human attached doing the movements and Tesla accidentally left him in frame so they had to admit it was not autonomous.
I think there are decades before any of this becomes a household product. But it is great that even so, some companies keep innovate and work on it, if not it will never happen. ...Tesla though, I feel like is developing the robot for other reasons. 1. like the cybertruck, the employees of tesla need to do that the stupid great leaders says, 2. they need new stuff to pump the stock. Which is also why you cannot trust any of the demos tesla show, they use a lot of trickery and rendering and cutting.
I hope they can still do parkour
You mean Elon doesn't have the monopoly on humanoid robots? What will his fanboys say now about Tesla's future stock prices?
I think it's funny that everyone start doing electric humanoid robots after that Tesla presented their concept.
I'm sure some of them will be profitable too. Maybe not as profitable as Tesla, though.
Let's review this subject in 2 - 3 years!
Nobody said Tesla has de monopoly, Tesla are the first comers, and much of their products are copied or based on their ideas, Figure AI, Atlas 2, all are electric and no mhydrolics, also with Brains, and also all started after Tesla Optimus presentation, or at least, started taking humanoid robots seriously.
@@Voltaje_YT First, I was joking. But I was making a point that Tesla takes credit for things and people give them a pass. Honda introduced ASIMO in 2000 and it was a humanoid robot that used AI. Tesla copied and improved on ASIMO. You dig deeper, and you will find multiple papers on others integrating neural network and robotics before Tesla. Tesla was better at marketing their concept and stealing the limelight from other companies who have been working on this for decades.
@@larsnystrom6698 Look at the Wikipedia page on humanoid robots. There were plenty of companies using electric actuators and integrating AI years before Tesla. Tesla just had the marketing and the fanboys to get the message out to the rest of the world. But they were not the first. What's funny is how people are giving Tesla credit for things they copied. I am not mad at them because now billions are pouring into researching and developing this technology. Maybe we will have a consumer humanoid robot within 10 years.
@@larsnystrom6698Boston dynamics has been messing around with humanoid robots for years the first few versions of atlas’s was just research and development. And this current model clearly has been in the works for awhile
Also Boston D will not be making consumer robots at scale. They will be more for commercial and military use.
Diffrent base. I think when it comes to warehousing, Teslas humanoid bot will be the winner, but BD will trash it when it comes to specific applications.
Bankrupt Tesla? We don’t need a robot that stand up this way. It serves no fxxking purposes.
If not, Tesla will be bankrupted by Elon. The future looks bleak for Tesla.
And by the way, we do want robots that have no silly artificial limitations.
They need to make him look like baymax!
Why be so short sighted? This version would be perfect for Space exploration on the moom. Eploring the lava tubes for possible human occupation in the future. Whe can actually get early tests on human designed landing systems. They don't have to be fully human rated (cheaper to strart). They just have to perform successful on point landings. Tesla claimed their capsule could do a powered landing of a human crew back on earth. Put that claim to the test but on the moon. Once there they can prepare a landing field for future missions, deploy a "crane" for lifting up and down equipment into the lava tubes. Power can come from RTG's where the robots just plug in for a charge on a regular schedule. The same goes for Mars. They could assemble a rocket launch pad and build a "Sounding" rocket to launch our samples back from Mars. That would be truely reusable, our skycrane technology is very accurate. Rather than crashing them off in the distance. We can possibly land them and reuse some of the parts for building the return rocket. Less risk and way more ROI going that route vs. a factory of humanoid robots working 24/7 producing whatever for "humans" to buy? With what money? If even the factory jobs are fully automated like Elon wants. The one place the human race will always have a labor shortage is in space. These robot helpers do not have to be fully autonomous. Tell it what to do, it does it. Humans remotely control its actions. Sure we can create a library of "responses" or "actions" for repetitive work.
Look beyond for their usefulness. We need space rated versions of this ASAP!
There's always been talk that we should just send robots and rovers for space exploration / colonization. But it doesn't really excite the public unless we send people. However, we could send people and leave the robots behind to do the real work :)
And just an FYI my first real break was working with Bill Nye the Science Guy :) I'm all for space science!
@@realchris Send them to set up the landing base and mining equipment. The pictures of them doing work is exciting for those that want to code. The hardware for those that love design. Do you see where I am going? These are our humanoid assistants not replacements. That is how it should be talked about for science. Elon Musk wants slaves, be they mechanical or not. He wants it all for himself. Money, fame, Cult leader devotion.
Yes but Teslabot at this stage is pretty lame. It can barely walk very slowly on flat smooth surface. is that the best that Tesla can do? Also remember there is minimum 3 to 22 minutes (round trip 6 to 44 minutes) time for radio communication between Earth and Mars. Tele-operation will not work on Mars where robot can fall down a cliff before a human operator has chance to react. Also, humanoid robot may not be best for space operation. Humans are limited to four limbs and five fingers, but a robot can have any kind of appendages and any number of them. Humanoid robots are primarily useful for navigating human settings. outer space uses will have vastly different requirements and robots with much more versatile design may be required.
@@springer-qb4dv I can see you are like Elon. Short sighted, and misinformed. I will just say this. Our current robotic probes all work "Tele-operation". We send commands, they reply with results. The ability to self right is shown with Atlas v.2 in this video.
Tesla bot is vaporware imo, more like the puppet arm from the Terminator movies. Take the arm and hand for instance. In the scene when Arnold cut off the skin of his arm and hand. You see the actions of the hand and fingers, they are all wire and pulley operated by a group of puppeteers off screen. Today it can be operated by servos pulling the wires. Slow and weak. Just another Elon "Grift" for the gullible.
It has some moves! But I'm pretty sure they are scripted. But we see the inheritance from Atlas!
It would be a better military robot than anything else we have seen. But that probably requires more endurance than is yet possible. So not a market for that yet.
I would guess it will be pretty expensive to manufacture. So perhaps not for a mass market.
It remains to see how good it's brain will be. Cooperation with someone?
I'll guess Optimus will be more energy efficient, and much cheaper. We can be pretty sure they know how to get it a brain too.
So, no it won't bankrupt Tesla. But maybe give them some healthy competition, so they have to strain themselves a little more.
Isn't it funny that everyone now build electric humanoids? Tesla, or Musk, really drives much of the industry thinking!
Not impress, can it do it on spandex body suit?
Ditto. They are 20 years ahead of Tesla.
30 years....lol.
🏠 Start Investing! www.chrisnorlund.com/invest
Do you have a video where you outline your personal feelings on Tesla?
Everything that I've seen appears negative, or maybe it's just Elon you take issue with.
It's really no different than SMR and his constant pumping no matter what.
Thanks
Look cool but my guess is typical Boston Dynamics it will cost $1000000 each.
I watched this video on X last night and couldn't tell if it was real or AI - if real, why did they make it stand up in the creepiest way possible? Just to troll Elmo?
Yep, your video is about Tesla, or how? Why is Tesla in danger, and not all others doing the same thing? Clickbait, BS, and a lot of ....BS!
You people cannot see that this robot is just a technological demonstrator with no real-life meaning (seems like Boston Dynamics is in panic and made this)? Who needs those joints, and for what activity/purpose?
PS: Logic, why should someone buy an acrobatic robot with no dexterity in fingers (Is it just me, or has this robot already arthritis, as hands are fixed in that characteristic position!)?
Is its purpose to climb a rope, or play in a Circus?
So tesla robots are already behind?
Of course. What Tesla have, students have done years ago.
Tesla bots were always a fraud.
@@Xanthopteryx
That student's product will sell, though. Bringing in trillions!
@@Xanthopteryx tesla is not betting much on the most amazing humanoid.. Its focus is on affordability. A strategy that has helped their car business.. Making cars affordable by scalable tech and removal of nkn essential parts..
@@larsnystrom6698 Not to mention: It will actually work!
It appears the hydraulic version was never built to scale into production whereas the electric now has this in mind. As to the humanoid form being developed I believe the rational behind that is the ability to offer something commercially that can be integrated into a environment which is currently suited for humans but moreover to pique public interest in order to further the endeavor. It would seem to me if your ambition is to replace humans for commercial applications, in many instances you would have full control over designing the environment the robot operates in and build task specific robots as part of the facilities automated eco system. This brings into question the importance of humanoid form to commercial requirements and to what degree these humanoid demonstrations are for the purpose of gaining interest. I would say the objective is in seeing how adaptable robots are in operating in complex environments without requiring someone to teach it each and every step of a task as well the ability to analyze situations which don't cleanly fit into their expectations and having them take the appropriate actions within the context of what they were assigned to carry out. BD's short demo captures this concept well. See it as human but also see it as something other than human adapting in a way a humans can't.
Bankrupt Tesla? That's delusional
If anyone can do it it’s Musk
Yes....no one asks who is going to build them. The jump between prototypes and mass production is something many underestimate. Tesla will pump them out in the millions.....
@@lamsmiley1944He's doing a great job of losing money with X .
@@lamsmiley1944He's doing a great job of losing money with X .
I think the word youre looking for is "articulation". This is mad creepy and the last thing somebody will see in the near future.
Yes articulation :) When I make my videos it's live and a single take.
It's not about the "body", it's about the brain....How far is BD ahead in AI?
You need to stop calling them Boston Dynamics. Is just Hyundai now.
Dude you are obsessed with Tesla let it be.
Nah its fine. Too many Teslainvestors pumping, we need an opposite side as well.
Tesla's was remote controlled. More musk bs
They didn't say how much or what kind of ai it will use. Hyperbole
Atlas will bankrupt Tesla 😂. That’s got to be one of the dumbest things I’ve ever heard.
You are funnier than Kim Jong Un😂🤡
it is definitely coming, the companion/sex robots.
Have fun stuffing that cold metal p*ssy or penis for that matter.. lol
Your sexual preferences are noted!
Ayoo?
Klaatu Barada Nickto
Boston dynamics does not have the manufacturing capability of tesla, not even close. Even if they are first to market, tesla will dominate
Curious. Where did you hear the manufacturing capability thing?
Algorithm
Something tells me you have a wierd hate for Tesla. Lol, hope you find peace in your mind and start believing Live and let Live. 😂😂😂😂
Hyundai owns 80% of Boston Dynamics, so Hyundai pretty much has total control of BD. Mechanics of robots is hard, but mechanics is trivial compared to designing robot AI which will allow a robot to independently engage and work with dynamic changing environment full of people and other machines. Tesla is nowhere with it's optimus Teslabot. Musk's touting of optimus as serious product this or next year is another one of his whopper lies. Musk should have bid for BD and bought it while it was available but apparently Musk didn't realize he needed Teslabot back in 2021! LOL
Bullshit
Looks like CGI.