We are much closer to seeing those criteria at work in 2024 than you realize. Also, although the workspace for robotics is almost infinite, there are specific tasks that we require humanoid form factor for.
I like that you don't see this as being an either/or situation. I think it will take time for the transitions, and most likely during that time, people will be inventing new jobs which will eventually become undesirable and only robots will do them... until full automation occurs.
I think humanoid robots are useful at replacing a task already being done by humans currently, but it is almost definitely not the most optimal form in any specific task. As a field gets more automated, I'd think we would start to change out the humanoid robots for more efficient alternatives
In my opinion, the main advantage is the tech progress that we make with humanoids is transferable/applicable to other robots in real-world use cases. Not sure if we every use humanoids for industrial setting
Very good point - and one of the key reasons FOR humanoids - is that transfer learning from humans and videos of humans is much easier to humanoid robots.
They are certainly awesome, however, I worry that they will have a hard time solving even the simplest of tasks. Dedicated robots are strong competition for almost any task. Where humanoids may shine is where a job requires multiple, disparate tasks, which utilize the full spectra of possibilities offered by the humanoid compared to a limited spectrum offered by a dedicated robot. Picking up trash near a roadside is probably such a task, whereas depalletizing most likely will be solved with more specialized robots or even an automation. So, in my view, the humanoids need to be able to traverse very complicated terrain, travel far on a charge, and be close to AGI to respond with intuition whenever unexpected situations arise. In farming and fruit-picking, humanoid robots might revolutionize farming yet again, minimizing pesticides and compacitification the soil. To do so, they must be ruggedized for those conditions, and having much more durable, yet light-weight features. Will current actuators solve that job, being so heavy and inefficient?
Great post and discussion, Andra, thanks! I’m thinking about situations that can benefit from deploying humanoid robots that can be human-piloted, such as the general-purpose work robot Phoenix by Sanctuary AI. When running in human-piloted configuration, this robot’s sensor-laden hands and arms are directly controlled by the pilot’s hands and arms; the pilot can feel whatever the robot is touching.
One can anticipate needing a human-piloted humanoid robot, or even a fleet - for multi-robot operation or collaboration, all driven by a single human pilot who’s controlling and feeling the hand and arm actions of all those robots.
Nice Humanoid robot update letter, but you seen to be forgetting the ATOM 1.0 humanoid project from FutureBots, why is that? I have work harder than any of the big humanoid makers, I'm a team of one, they have 100's on there projects and deep pockets and was in the public eye years before them, is my project less worthy than those big funded projects, ATOM is a very highly developed humanoid with features these others are just looking at? Mine has similar locomotion as the others, but better in some ways than Tesla and Figure, Please I just want a fair shout out as you do for the others, right now I'm trying to finish up upgrades to ATOM and get it walking this year, but money flow is low, the bills come first, thanks Andra for any help, you have a great newsletter, working by yourself is hard. Merry Christmas Dan
Dan, I only mentioned a handful of companies that are meeting COMMERCIAL criteria. I put ATOM on the general humanoids list and if I get time to update the post I'll give you a shout out -- but realistically speaking you are not competing on the same playing field.
If that’s the criteria for success then we have a LONG way to go. But I have to ask the fundamental question. Why? Humanoid robots are not (IMO) the correct solution for most industrial problems. I discuss it at length here. https://automationnavigator.substack.com/p/to-humanoid-is-to-err
There is a simple job - currently being done by Agility Robotics - that Amazon can not find people to do. It doesn't even require much in the way of hands, which is great because hands break!. One simple reason that humanoid robots are doing this job is reaching high and reaching low, then moving across a varied length distance.
Absolutely basic skills (and incredibly boring) for people. Also too fiddly for fixed automation or any other robotics solution.
Maybe Andra I'm not on the same playing field because of the media never seems to take the ATOM project seriously, it day-viewed in 2010 in Servo and Robot and Engadget, the only thing that separates me from them is Money, they have the deep pocket connections in industry and academia, they seem to put there hands out and there are millions for good and bad projects, I have been trying for years to find VC's, Angels, I have talked to Intel, Microchip, and few VC's and nothing yet. I may not be on there playing field but I was on a playing field many years before any of them even thought of building humanoids. I can just imagine Microsoft or Apple today trying to get started in there garage they would I bet be in the same boat like me, not being taken seriously. the work goes on regardless of money or help, this is my passion building robots, it's not theirs, they get money they build something and in most cases it fails, then they go to the next funding project. I have seen so many entities get millions of funding and then go bust with in a year, my ATOM has been a 20 year project on my own with little money, but I have ideas they don't, none of the Tesla's of Figures can say they have passion to build, to them it's just another funded project. I build my ATOM series of humanoids to truly help people in a car-taking role, I was my Mom, my Brother and Dad's caretaker when they were sick before there passing, now I want robots to do that hard work, Thanks for the input. Regards Dan FutureBots.
We are much closer to seeing those criteria at work in 2024 than you realize. Also, although the workspace for robotics is almost infinite, there are specific tasks that we require humanoid form factor for.
I like that you don't see this as being an either/or situation. I think it will take time for the transitions, and most likely during that time, people will be inventing new jobs which will eventually become undesirable and only robots will do them... until full automation occurs.
I think humanoid robots are useful at replacing a task already being done by humans currently, but it is almost definitely not the most optimal form in any specific task. As a field gets more automated, I'd think we would start to change out the humanoid robots for more efficient alternatives
Nice article Sean.
In my opinion, the main advantage is the tech progress that we make with humanoids is transferable/applicable to other robots in real-world use cases. Not sure if we every use humanoids for industrial setting
Very good point - and one of the key reasons FOR humanoids - is that transfer learning from humans and videos of humans is much easier to humanoid robots.
They are certainly awesome, however, I worry that they will have a hard time solving even the simplest of tasks. Dedicated robots are strong competition for almost any task. Where humanoids may shine is where a job requires multiple, disparate tasks, which utilize the full spectra of possibilities offered by the humanoid compared to a limited spectrum offered by a dedicated robot. Picking up trash near a roadside is probably such a task, whereas depalletizing most likely will be solved with more specialized robots or even an automation. So, in my view, the humanoids need to be able to traverse very complicated terrain, travel far on a charge, and be close to AGI to respond with intuition whenever unexpected situations arise. In farming and fruit-picking, humanoid robots might revolutionize farming yet again, minimizing pesticides and compacitification the soil. To do so, they must be ruggedized for those conditions, and having much more durable, yet light-weight features. Will current actuators solve that job, being so heavy and inefficient?
Great post and discussion, Andra, thanks! I’m thinking about situations that can benefit from deploying humanoid robots that can be human-piloted, such as the general-purpose work robot Phoenix by Sanctuary AI. When running in human-piloted configuration, this robot’s sensor-laden hands and arms are directly controlled by the pilot’s hands and arms; the pilot can feel whatever the robot is touching.
One can anticipate needing a human-piloted humanoid robot, or even a fleet - for multi-robot operation or collaboration, all driven by a single human pilot who’s controlling and feeling the hand and arm actions of all those robots.
Woot woot
Hello Andra,
Nice Humanoid robot update letter, but you seen to be forgetting the ATOM 1.0 humanoid project from FutureBots, why is that? I have work harder than any of the big humanoid makers, I'm a team of one, they have 100's on there projects and deep pockets and was in the public eye years before them, is my project less worthy than those big funded projects, ATOM is a very highly developed humanoid with features these others are just looking at? Mine has similar locomotion as the others, but better in some ways than Tesla and Figure, Please I just want a fair shout out as you do for the others, right now I'm trying to finish up upgrades to ATOM and get it walking this year, but money flow is low, the bills come first, thanks Andra for any help, you have a great newsletter, working by yourself is hard. Merry Christmas Dan
Dan, I only mentioned a handful of companies that are meeting COMMERCIAL criteria. I put ATOM on the general humanoids list and if I get time to update the post I'll give you a shout out -- but realistically speaking you are not competing on the same playing field.
They shouldnt break (often)..
If that’s the criteria for success then we have a LONG way to go. But I have to ask the fundamental question. Why? Humanoid robots are not (IMO) the correct solution for most industrial problems. I discuss it at length here. https://automationnavigator.substack.com/p/to-humanoid-is-to-err
It's going to be very interested to see more work in the areas of multi-robot teleoperation.
There is a simple job - currently being done by Agility Robotics - that Amazon can not find people to do. It doesn't even require much in the way of hands, which is great because hands break!. One simple reason that humanoid robots are doing this job is reaching high and reaching low, then moving across a varied length distance.
Absolutely basic skills (and incredibly boring) for people. Also too fiddly for fixed automation or any other robotics solution.
Maybe Andra I'm not on the same playing field because of the media never seems to take the ATOM project seriously, it day-viewed in 2010 in Servo and Robot and Engadget, the only thing that separates me from them is Money, they have the deep pocket connections in industry and academia, they seem to put there hands out and there are millions for good and bad projects, I have been trying for years to find VC's, Angels, I have talked to Intel, Microchip, and few VC's and nothing yet. I may not be on there playing field but I was on a playing field many years before any of them even thought of building humanoids. I can just imagine Microsoft or Apple today trying to get started in there garage they would I bet be in the same boat like me, not being taken seriously. the work goes on regardless of money or help, this is my passion building robots, it's not theirs, they get money they build something and in most cases it fails, then they go to the next funding project. I have seen so many entities get millions of funding and then go bust with in a year, my ATOM has been a 20 year project on my own with little money, but I have ideas they don't, none of the Tesla's of Figures can say they have passion to build, to them it's just another funded project. I build my ATOM series of humanoids to truly help people in a car-taking role, I was my Mom, my Brother and Dad's caretaker when they were sick before there passing, now I want robots to do that hard work, Thanks for the input. Regards Dan FutureBots.