Doesn’t look like a new design to me. Looks like a stagecoach from the late 1800’s. Minus the horses.
I would advise against using a concept to support your position. The vast majority of concepts never make it into production.
Here is another Nissan concept from 2005:
Another:
Meet the latest concept car from Nissan: The Teatro for Dayz . Don’t worry if you are a little confused about its crazy looks and design – chances are this car isn’t ever going to see any form of production.
3 Likes
Looks a lot like a Kia Soul to me. At least on the outside. Little shorter hood, a bit more square, but the same basic shape.
No it’s not. There is little about “humanoid” that makes sense. It’s overly complex for the majority of jobs, and better, easier, cheaper designs will cover nearly any (but not necessarily every) design.
A car doesn’t have to also be a truck, a van, and a motorcycle. It does one thing really well. Robots don’t have to do everything, and if they do they will be vulnerable to all kinds of maintenance and breakdown issues that a purpose built robot won’t.
I’m not resisting new ideas. I’m resisting stupid new ideas. Progress is doing things in a new, better, less costly way, I’ll admit. But chasing every dumb idea that someone has isn’t progress, it’s the antithesis of it.
You know how many money wasting, resource sucking terrible ideas there are? Most of them. The art of management is picking the good ideas , not saying yes to every crackpot that wanders the halls pontificating about “the future.”
I’m happy to admit that I have no idea what society will look like in 100 years, but neither does anyone else. 20 years up? I can guess pretty well, and I’m going to say it doesn’t involve toaster cars or complex robots that do “everything” poorly and “nothing” well.
5 Likes
This is nonsense and is demonstrably wrong. There are millions of jobs that do not require great intelligence or skill and are tedious, dirty, or even dangerous, that are being performed by “humanoid” humans instead of specialized nonhuman-shaped machines. Literally millions.
We’ve talked about meat processing. Amazon hired 250,000 humanoids this past holiday season to work its warehouses. We have a porous southern border because our economy doesn’t work unless we have hundreds of thousands of humanoids illegally enter the country to work in our farms, restaurants, and construction sites.
In fact, we are so dependent on “humanoid” workers that the decline of China’s humanoid workforce is stagnating the world’s number two economy.
There is far more real world evidence that a humanoid form with a semblance of human-like intelligence capable of generalized multipurpose functioning “makes sense” for the majority of jobs that are out there today. This is why even conservative companies like Honda, Toyota, Sony, Samsung, as well as the Chinese government are spending billions developing general purpose humanoid robots.
It might be prudent to at consider the possibility that these folks see something that you don’t.
Every idea that disrupts the status quo has been called stupid by those that make up the status quo. Twenty years ago it was the notion that online Amazon could challenge Barnes&Noble, let alone WalMart. Ten years ago that electric Tesla could take on GM and Ford. I have a feeling you considered all those ideas stupid at the time. Today it is the marketability of general purpose humanoid robots. The beat goes on.
1 Like
So, when bots take over and humans no longer exist, will they worship us as like a god:
Then humans said, “Let us make bots in our image, after our likeness. And let them have dominion over the fish of the sea and over the birds of the heavens and over the livestock and over all the earth and over every creeping thing that creeps on the earth.”
Just curious. 
4 Likes
If I am understanding this logic, citing demand for labor (in your examples: meat processing, Amazon hiring, immigration to US, China labor demographics) as evidence for humanoid robots being the best (or to be carefully nuanced, “makes sense”) robotic form for many jobs is off target because most of our labor today is available only in humanoid form. What choice do we have?
Perhaps when there is a selection of robotic forms available at scale as labor supply (humanoid as one), we can then measure what demand materializes for each form.
2 Likes
This is the future I see. Eventually each task will be completed by a form that is efficient for that task.
Here’s an example:
The insistence that there will only be one robot form … Is a strawman?
I’m in the camp for “there are and will be many tasks for which a humanoid form, human size, and GenAI will be the best fit”.
There are lots of companies vying for First Mover Advantage.
I see Tesla as having early leads in several aspects:
FSD and it’s associated GenAI development.
Experience training “AI enabled robots”.
In house design /development of specialized components.
Experience mass producing and scaling new products.
Production capacity.
Lots of money, lots of high talent people.
And Elon.
ralph
5 Likes
Exactly Ralph and all those arguing against a humanoid form have already been proven wrong. They are arguing against something that has already happened and they do not even realize it. That is what makes me laugh, it is like arguing that Tesla will never be a great auto company lol.
Andy
In point of fact, most of our labor today is being done by nonhumanoid machines. That’s the current industrial revolution and we have gotten very good at making machines of many different shapes that can do a specific task. Yet despite this technology, there are still many, many positions that do not require great skill and are tedious and often dangerous, but are still best done by humans.
Janitorial work for example is not intellectually taxing but involves multiple activities, from emptying trash, to changing light bulbs, to scrubbing toilets, and these have to be done in environments that are optimized for humans, with narrow hallways, steps, and doors. Seems to me that a general purpose humanoid robot is best suited to do such activities, as opposed to having multiple specialty machines each designed to do a single task.
A similar argument can be made in many industries. For example, one way to make hospitals more efficient is to use humanoid bots to do the more menial but still important tasks, like emptying bed pans, changing bedding, and overall cleaning. One can partially automate hospitals without having to redesign them.
One reason we have huge industrial farms is because automating agriculture with single purpose machines is best done on a large scale. Humanoid bots are one way around that. Family farms can partially automate through general purpose smart robots that can perform multiple tasks using human tools.
This is just a long-winded agreement with what rainphakir posted: “…there are and will be many tasks for which a humanoid form, human size, and GenAI will be the best fit”.
2 Likes
@btresist
You wrote that A implies B (or at least that’s how I understood it, maybe I misinterpreted, which is not uncommon is these discussions).
Where A is
demand for labor (in your examples: meat processing, Amazon hiring, immigration to US, China labor demographics)
and B is
humanoid form is the “makes sense” (me: ?best, ?nearly best) robotic solution
Above I am paraphrasing you and mostly quoting me.
I am saying A does not imply B. Nothing you wrote in prior post addresses A implying (or not) B.
A implies B would be weird because:
- All of that labor demand most likely does not have meaningful supply of robotic alternatives today, so there is no substantial choice other than humans.
- You just cited cases today in which non humanoid has replaced humans. Is there reason to believe this doesn’t continue, at least some, perhaps a lot?
Are you still claiming A implies B?
Do you disagree with 1 or 2?
Side note: perhaps we need to distinguish between labor and capital, with the former being human and the latter machine (or I am open to other useful definitions). Perhaps we can use the concept of work, which can be done by labor and capital, mostly by combining the two.
Side note 2: I am not making any claims about the future state of robotics, I have no claim about form, or use cases, or probably much else robotwise, other than they will continue to happen.
1 Like
Yup. It’s like this. During the past two centuries non-humanoid single purpose machines have replaced humans in a great many jobs. Yet humans are still needed in a large number of low skilled jobs. We don’t need as many people to dig holes but instead we need them to clean hotels, care for the sick and elderly, put up dry wall, and pick produce. This illustrates the limitations of current methods of automation. Single function machines are cost effective when used frequently on a large scale, but as the task gets more varied or infrequent they become less effective and too expensive.
So yes, I believe the current demand for labor in low skill jobs reflects the limitations of current methods of automation and the potential benefits for the new paradigm of automation using general purpose humanoid robots.
Here’s the logic. We know humans can do these jobs, therefore we can be reasonably certain that humanoid robots mimicking humans can also do these jobs. Furthermore, there is no reason why that humanoid robot can’t do a multiplicity of jobs, which one would think would be a great advantage for the owner of the robot. It is just a matter of technology and cost of production.
Don’t get me wrong, there is nothing wrong with hiring humans if they want to do low skilled labor, but if you look at demographic trends, there soon won’t be enough humans to fill the need. That’s bad news for economic growth.
2 Likes
We dig exponentially more holes than we used to, but we use robots to do it. Excavators, steam shovels, boring machines, and yes, landscapers with post hole diggers. Not a lot of jobs there for humanoid robots, I don’t think, at least until they are fantastically advanced from anything on even the most optimistic drawing boards today.
Well, I’m reasonably certain of no such thing. Perhaps in 50 years? 75? The kind of physical dexterity coupled with the learning and motive power required to do “all these things” is beyond imagining, unless you whittle down the jobs to a handful of specialized applications. Which, of course, is exactly what we’ve done: we have robot arms, but without feet or brains. We have robot brains (AI) but without fingers or legs.
Yes. Just like “full self driving”, which looked easy when it first started leading to wildly optimistic predictions, and got harder and harder as the technology developed, to the point where some wonder if it can ever be accomplished, at least to the same degree as “human self driving.”
Elon Musk's Talked of a Self-Driving Tesla for Years. It Still Doesn't Exist..
Meanwhile, I have no doubt that robotics will continue to progress and that there will be substantial investment in it going forward, and success for many players. I also am quite sure that 98.3% of that will be for special purpose robots, everything from floor vacuums and ice makers to industrial scale special purpose machines and custom tool applications, like in chip fabs and elsewhere.
No, they haven’t happened in any meaningful sense. There are cars that float in the water and cars that can convert to airplanes, so those have “happened”, but nobody can make an economic case for them. That’s what we’re talking about with “humanoid robots”, at least in the short or medium term. That doesn’t mean people shouldn’t still tinker with them, but there’s going to be a lot of expensive tinkering before they’re ready for prime time.
Again with the dreamy, unrealistic nonsense. You’re going to build a $250,000 robot to clean toilets and bedpans? OK, and good luck to your sales team.
4 Likes
Ahh so now the bar has been raised to can they be profitable? Well I know many companies that have been going for over 20 years that are still not profitable but they do have FCF to keep their companies going. I do not think anyone can argue about this because you would first have to know what their business model looks like and since I know you haven’t studied that it is all a guess. I am fine with winning the argument over whether they would produce a humanoid robot or not. Now that we have made it past that I will wait to see how the companies that are producing them proceed.
Andy
That’s what I’m thinking. The jobs are being done by humans because humans are cheaper than machines. If (when?) the machines get cheaper than humans, machines will start doing the jobs. And if that happens, the machines will be designed to do the job with no regard for making them look like humans.
—Peter
Yup. Most of robotics to this point has been to replace high cost workers, like UAW guys on the line making $50/hr, not CNA nurses making $12. Changes the value proposition entirely.
The other segments are doing little jobs which nobody wants (vacuuming your floors, washing your clothes), or which can make vast productivity improvements (shaking fruit trees to drop the fruit, moving merchandise stacks around in Amazon warehouses). The idea of androgynous robots doesn’t seem to work in any of these scenarios given their high costs and the low value work being replaced. Fast food workers? Janitorial? And how does a hospital robot emptying bedpans disinfect itself between each patient?
No, that has been the argument from the beginning. You can do almost anything if you pour all your resources (and more) into it, but that’s a bad way to run a company or make an investment thesis. Nobody - except you - is talking about making a wildly uneconomic product which cannot be viably supported in some way.
Yes and no, or rather no and yes. I was already an AOL investor when Amazon debuted. I thought Bezos case for books was good, but it didn’t look as lucrative as it turned out to be. (Amazon didn’t start selling other products for 5-6 years, IIRC.)
And while I didn’t see Tesla’s success, I did advocate for “electric vehicles” back in the late 90’s. But admittedly that was for upgraded golf cart (vehicles), owing to my experience with my parents at a retirement community where car transport was “too much” and walking was “too far” (and my own use of carts on the golf links, of course.) Not the same as predicting Tesla, certainly, but I wanted a better, weather protected electric vehicle but obviously never took it anywhere close to “road vehicle.” I never dissed the idea of a road EV, and looked at the early models (Bolt, etc.). I’m skeptical of FSD, but that is a different thing. You see?
I see no need, and a sliver market for humanoid robots, and haven’t changed that opinion even with all the glittery promises made in this thread.
7 Likes
Don’t dismiss it until you run the numbers! If it costs $250,000 and you finance it, let’s say with a mortgage-like instrument that amortizes over 30 years, then the monthly payment is only about $1750 a month. That, included with maintenance and electricity, may well be less than paying someone $2500 a month “to clean toilets and bedpans”.
It might work!
3 Likes
Looking at the questions I posed, @btresist
You responded to this question:
And for these two that I asked,
You basically agree with (1) here,
saying that there is no current supply of robots to substitute for current labor demand.
As far as “A” (demand for labor) implying “B” (humanoid robot makes sense), you have instead provided in your latest answer a somewhat different precondition (call it “A2”) and statement, something like:
A2 (demand for labor + some argument/speculation) might result in B (humanoid robot makes sense)
where the “some argument/speculation” you explain here:
But the above is speculation, as noted here, for example:
You have not shown A implies B, instead you just make a claim that A might lead to B. There’s a difference.
Your argument is theoretical, in my view.
My argument is empirical, I am talking about data that exists now. These things are true statements:
- All of that labor demand does not have meaningful supply of robotic alternatives today, so there is no substantial choice other than humans. And hence we can not measure/observe which use cases will actually be supplied by human vs humanoid robots vs non-humanoid robots. (if we could run this trial now, we could just look at the results and stop debating about the unknown future)
- You just cited cases today in which non-humanoid robots have replaced humans. And this is happening now, every day, and through the history of automation, perhaps because they fit the use case and perhaps also because there is no meaningful amount of alternative supply of humanoid robots. And hence there is lots of data (evidence) showing non-humanoid robots can replace humans, but we don’t have this same data for humanoid alternatives.
My example 2 gives data where [A] results in [not B], which is counter to the claim [A] implies [B].
Again, for example 2, non-humanoid robots replacing some humans, is there reason to believe this doesn’t continue, at least some, perhaps a lot?
There will be advances in non-humanoid and humanoid robots, and I would speculate some portion of the labor demand you cite will be replaced by each, almost certainly some non-humanoid because that is happening now. How much by non-humanoid versus humanoid remains to be seen.
As I said, I have no strong view of humanoid vs non-humanoid or other details on the future of robots, I was just interested in whether the claim A implies B is firmly supported, or if it is just a speculation about the future. I think it’s clear that it is a speculation about the future, and most people, I imagine, would agree that this is a speculation versus a stronger IF A THEN B conclusion which is how I saw it presented in the context of the original post, and also when you answered “Yup” to “Are you still claiming A implies B?”
By the way, National Robotics Week takes place from April 6 to 13 this year, for all you enthusiasts.
1 Like
I doubt it. Would you?
That is the point. Tesla expects over the first decade of implementation a savings of $25 billion. Who cares how wrong that figure might be? It is a savings.