Tesla has a lot of work to do on its Optimus robot Tesla headlined its AI Day 2022 event on Friday with the reveal of its "Optimus" robot prototype, showing just how much work was left to do on the project. While the demo was certainly more robotic than last year's dancer in a onesie, the lumbering mess of cables was far from the sleek and sexy design faithful Muskites might …
..sees humanities problems as not having someone cheap enough to water your plants for you.
Of course he doesn't see any problem in making the bottom third of the population redundant. When he says "you will have whatever you want in terms of products and services", of course he means only those people who still have jobs.
This post has been deleted by its author
The other thing is his perhaps less well known thoughts on the general decline in birth rates around the world - he even views it as a possible explanation for the Great Filter of the Fermi Paradox. i.e. unwitting population collapse.
As Japan has noted with an aging population and a falling birth rate they are going to need some means to physically look after increasing numbers of elderly, frail people with less young people and they are looking at robots to do this.
This is also true across many "western" countries too and many of those countries also have increasing opposition to immigration which is really their only source of cheaper labour who might be prepared to do this normally poorly paid work.
So, from this point of view, a humanoid form robot is a partial solution to this.
However, it's always going to be vastly cheaper to pay a carer a decent living wage than putting a horrifically expensive robot into someone's home to look after them, or a horribly cheap one that doesn't work and breaks down and requires constant servicing.
The problem isn't the absence of working humanoid robots to do cheap "unskilled" labour*, it's the lack of value given to humans doing what a robot can't do properly in the first place.
*There is, of course, no such thing as unskilled labour; all labour requires some sort of skill or other. What people mean when they say this is actually "undervalued labour".
A useful tool to help kickstart a colony on mars would be a tool that can answer the question; "Why exactly do we want to have a colony on mars again?"
Just being practical here, Mars doesn't exactly offer anything we really need. Wanna establish a colony? Why not the Moon? Its alot easier to reach, humans have actually been there and lived to tell of it, and it may even provide a valueable resource that isn't abundant on earth: Helium-3, which could be a key fuel component for fusion reactor technology.
" it may even provide a valueable resource that isn't abundant on earth: Helium-3, which could be a key fuel component for fusion reactor technology."
It's also the perfect place to do research on dangerous pathogens. It might be great for manufacturing advanced semiconductors or medical therapies that are easier to synthesize in low gravity fields.
I had a nice long chat with Charles Walker who was sent up on the shuttle three times to operate his protein electrophoresis experiment. Being able to run the tests in zero G lead to insight on how to perform the process on Earth. I think I have a recording somewhere of that conversation. It was a space convention and very few other people realized who he was so I was able to monopolize him for a good long time. That sort of return is one of the things that's great to see. If there were more opportunities by having a moon research station, great strides might be made in many disciplines.
Musk's latest fantasy won't make anyone redundant for many years, and probably never. It'll just be a curiosity a few rich people buy to impress their rich friends by telling it to bring them a drink or give them a laugh when it screws up.
Even the most menial jobs require a lot of understanding about the world that no AI has, and even where robots can find a niche (like cooking fast food orders or stocking shelves in a grocery) it isn't going to be one built like a human and walking around on two legs.
Every time people talk about humanoid robots that bring drinks, I'm drawn to remember the robot in Rocky IV and I have a good little chuckle to myself.
Because we don't even have that yet, and it was proper rubbish.
In case you'd forgotten, or are not familiar with the Rocky films, that film was made before the fall of the Soviet Union. I expect its makers like to claim some credit for that, in the same way that The Hoff was responsible for the Berlin Wall coming down...
"Every time people talk about humanoid robots that bring drinks, I'm drawn to remember the robot in Rocky IV"
It makes me think of the robot bartender in The Fifth Element. When I first saw the movie I had to wonder why they'd have an obviously robotic bartender using a common mixer dispenser rather than something even more robotic that had appendages for dispensing everything and a hand to place fruit and an umbrella. If the gag was to play to the priest pouring his soul out to the bartender, you'd think they want something a bit more humanoid.
Musk said. "We've also designed it using the same discipline we use in designing the car, which is to say to design a form of manufacturing such that it is possible to make the robot in high volume at low cost with higher liability."
I hope this was a typo. Seems "liability" should have been"reliability"... Maybe foreshadowing?
"at low cost with higher liability"
Probably a phoned interview taken down in shorthand (or the electronic equivalent). Such "typos" have always been common under those circumstances. See Toseland M. A steroid hit the Earth, Portico 2008. He quotes a somewhat more impressive one from the telegraph: "The seaman, severely injured when the ship was three hours out, was taken to hospital and the hippopotamus removed".
"That's one expensive Roomba."
I tried Roomba a few times over the last 5 years and it did 80% of what I was hoping it would do and cost $500. At 80% I still had to vacuum and it would get stuck, and cleaning it was not much fun. So I felt it was a $500 experiment that ended up being worth nothing.
Not sure I'd pay $20K for a vacuum, but if it mowed and washed windows, and carried a few gallons of water to the end of the driveway for the flowers, oh and I suppose it would make pretty good security bot; at that point a few thousand a year might well be worth doing.
Oh- and if it can clear snow - then it would be well worth it!
Watching the film of the robot at work, I immediately thought "I've seen small-arm robots doing that far more efficiently".
While the idea is interesting, with the exception of the walking (which has been done better by others) this really looks to be not much better than we had before, decades ago.
The human body is the shape it is because of evolution; bipedalism has evolved the way it has because it evolved from quadrupedalism, there are inherent stability problems (which is why human infants can't walk at birth, whilst other animals, such as baby deer, can). In fact, due to the way the pelvis has evolved to both support bipedalism and encompass the birth canal, human infants can't do very much at all, as they have to do a lot of development post-partum.
Designing robots to look like humans might be necessary if you want to build a convincing sex-bot, but other than that, why not design the form to fit the function, and not worry about pleasing perverts.
If you wanted to design a robot which can walk efficiently (rather than one on wheels or tracks), why not give it four legs, which at least removes the need to design expensive and complex balancing systems to move the centre of mass about to avoid falling over when walking.
I mean, you wouldn't design a computer vision system to mimic the evolutionary mistakes of the human eye, such as putting the blood vessels on the inside, resulting in a blind spot, or using materials for a flexible lens that stiffen over time, resulting in the gradual drift of minimum focal length.
Having said that, maybe that does explain Tesla's computer vision and navigation systems.
The point isn't to build a robot that can walk efficiently or even effectively. The point is to pump up Tesla stocks and maybe con a few more investors, and this pile of junk that makes 1950's Disney animatronics look cutting edge is doing a very good job at what it was built for.
Watching a baby start walking on all fours, then gradually moving to semi-upright and eventually running around is a great illustration of how our evolution has happened... but will we eventually evolve to having a 3 year-old sit in a Tesla and telling it to drive across town? But I think that we will keep walking ... even if it's only going to the front door to pickup an Amazon delivery.
This post has been deleted by its author
Is it though?
The things humans interact with are designed to be interacted with by humans, yes, and in most cases, a human is going to be far cheaper than a robot to do things humans can do, on a human scale.
Plenty of things are not designed to be interacted with directly by humans, though. Take roads, for example; a motorway is designed for cars, not people. Pedestrians are prohibited. Yes, the cars are designed for humans, but they are notably not human-shaped.
Or how about industrial processes, where human operators need protective clothing, or other safety equipment? Much better to design robots that suit those environments in a suitable shape; not bipedal. Robots that service nuclear reactors, for instance, are not human-shaped, the clever ones that can get around through narrow gaps and so on are snake-shaped, for instance.
And how about space exploration (Elongated Muskrat's other money sink). Building spaceships that can contain one or more humans and keep them alive is much more expensive than building something that can propel a robot somewhere that is designed to withstand high G-forces, not need food, water, or air, can withstand large temperature changes, and so on. Show me all the humanoid Mars rovers?
Machines such as robots are simply tools, and things in society are designed for humans, because society is composed of humans, not machines. Pretending that society is for robots, and thus we should make robots look like people sounds a lot like pointless folly.
While I mostly agree with the arguments against making robots that walk similarly to humans, there is one lingering thought that remains a bit of an argument for. A robot that needs to navigate stairs (designed for humans) while reaching high enough (top shelves designed for humans, so a doggy won't do) and being not too bulky (so not a human-tall quadruped). Cf. Daleks.
Not sure the argument is all that strong, or that it was a part of the design, but it still makes a bit of sense.
Building spaceships that can contain one or more humans and keep them alive is much more expensive than building something that can propel a robot somewhere that is designed to withstand high G-forces, not need food, water, or air, can withstand large temperature changes, and so on.
That argument was used when Moon landings were being planned. The argument is absolutely correct from the engineering prospective. It was shot down by pragmatic politicians who wanted to push space exploration further but realised that they needed the public imagination behind "Let's put brave American boys on the Moon!" to get the votes for the budgets. "Efficient space exploration" would never (well, at least for some useful values of "never") get the political support for the (still astronomical) appropriations. It seems that Mr. Musk's marketing of colonisation of Mars draws on that lesson, too.
Marketing works.
It's far easier to build something that fits in to what we already have
Only if you are dumb enough to believe you can make a "general purpose" robot. Musk is a fool if he thinks he can.
There are a few categories of job that are probably ripe for replacement by robots in the not too distant future, cooking fast food orders and stocking shelves in a grocery store are a couple of examples I used above. Neither would be best served by a humanoid robot, and both have a large enough addressable market that it makes FAR more sense to design a robot specific to those tasks rather than try to adapt the human form to it.
Even if you want a human form why does it have only two arms? Obviously humans would be more productive if we had four, or they could reverse and allow us lift something behind us as easily as we can lift something in front of us. Why have only two legs, and make the job of balancing that much more difficult? Why have a "head", instead of a small stalk that raises out of the torso with sensors (eyes etc.) and can telescope up to 10' or more to allow seeing on high shelves in a warehouse?
Even if you keep the basic human form, there are obvious improvements that could be made.
This post has been deleted by its author
Vehicles? You think robots are going to be climbing in a car or forklift and driving themselves around? Utensils? You think you need two arms and two legs and a head to use a spatula?
You think robots are going to be wearing suits made for people and that's why they can only have two arms? What the fuck are you smoking?
This post has been deleted by its author
"You think robots are going to be wearing suits made for people and that's why they can only have two arms?"
Household robots, sure.
The whole point is to have *servants* but servants who will never rebel, never strike, never get tired of your shit and never, ever boff your daughters in any meaningful way. Slaves, if you prefer but slaves that are entirely free from ethical issues.
Household robots as general slaves *would* solve a ton of problems, eventualy, as with motorised vehicles, even for the moderately poor. If they could be made smart enough.
But it only works if they look like people but not *too* much so. And if some smart guy can solve the hacker problem.
Oh, and, yes, they need only two arms. Little girls will always love to play dress-up with their robot friends. Though spare manipulators could be stored inside the torso for emergencies.
The "environment is designed for humans" argument doesn't wash, really.
In the built environment, a tracked or tri-wheel machine is far simpler than a bipedal and a far more stable platform for whatever the task might be.
The ground is smooth and firm with occasional steps and stairs, both easily navigated by a relatively simple tracked machine.
In workplaces, the legal requirement for disabled access has already simplified away the steps and stairs.
So why bipedal? It's pointless in the built environment.
This post has been deleted by its author
This post has been deleted by its author
So don't make it out of a 1/4 ton of metals, use ceramics, plastics and other strong, light stuff instead.
They are not motor-cars, they don't need engine blocks. A human can do human scale stuff and weigh roughly as much as a human does. There are few physical or engineering principles that restric robots to not fitting into that scheme.
No, household robots do not need to be built by Volvo, they only need to be able to carry an unwell human upstairs and into its bed.
Even tiny, little nurses can manage that much.
> It's going to happen, and probably a lot sooner than you or I think. Only a fool would believe otherwise.
Why? Throughout history humans have gravitated towards making tools that are MORE specialized, not less, even in low tech solutions; We don't have one kind of hammer, we have claw hammers, ball-peen-hammers, sledgehammers, smith hammers, mechanics hammers, engineers hammers, brick-hammers, tack-hammers, rubber-hammers, prospectors-hammers, framing hammers, drywall hammers, etc. etc.
And that has proven to be a very beneficial approach. A smiths hammer is just not as efficent a tool for a carpenter as a tack hammer. And a general purpose robot, will never be as good at, eg. cleaning toilets, as one specifically designed for cleaning. For that matter, a robot designed for general cleaning won't be as efficient at cleaning toilets as one specifically designed to clean bathrooms.
This post has been deleted by its author
Mabe because that robots purpose is to use exactly that one type of hammer to do one very specific job in industry very precise, effective, and with a long service life.
If using a variety of hammers is beneficial for the purpose, we would build a robot that has a specialized appendage that can be fitted with said variety of hammers.
Why is it likely that would we do that instead of mimicking a human hand instead?
Because such an appendage could have qualities like being shock mounted, have specialized sensors that measure and record impact strength, be reinforced to have a long service life even when doing repeated hammer blows. Its actuators could be designed to perform hammer blows at angles and in space restrictions where having an elbow and lower arm is impractical. Or many other possible qualities that would either be diminuished or absent in a robot where the primary design principle of the appendage is _has to be modeled after a human hand_.
And by the way; general purpose robots are not even that commonplace in ScienceFiction. C-3PO is completely unsuitable to lift heavy things or do complex repairs. R2-D2 cannot even speak English. And, neither of them would be a very effective battle-automaton, compared to the likes of B2s or Driodekas. So even in an area where we have the creative freedom to have robots with unrealistic abilities, we tend to create Specialized Tools rather than general purpose robots.
This post has been deleted by its author
Much more likely to build a specialised robot to replace the hammer, and hammer whatever it is that needs hammering in a very precise and controlled manner.
A £5,000 hammering robot, as opposed to a £5,000,000 general purpose robot with a £5 hammer trying not to hit its own thumb.
This is what tool specialisation means. Unless you can find more than 1,000 different jobs that your robot costing 1,000 times as much can do, as effectively as the specialised robot, the economics don't add up.
You also have the problems that if you are trying to make a general purpose robot that is that good (and it will have to be better than a human, to be cost effective, because a human handyman with his own hammer will cost you £50 for that one-off hammering job), it is likely to cost far more than a mere 1,000 times as much as the specialised robot, which will have been designed to do the job as simply, efficiently, and cost effectively as possible. Just building the mechanical actuators and parts to be sufficiently resilient, flexible and strong for all possible uses is going to be horrifically expensive. Presumably you're going to need strong flexible corrosion-resistant alloys (human skin doesn't rust, but most metals corrode). Then let's talk about electronics, sensors (it's going to have to be covered in pressure sensors to start with, to match what human skin can do), and it starts to become apparent that you'll never actually be able to beat evolution at the job.
Fantasy general purpose humanoid robots are exactly that - fantasy, because reality finds the most efficient solution to a problem, and they ain't it. Unless the problem is "how to make a convincing humanoid sex-bot".
> it's going to have to be covered in pressure sensors to start with, to match what human skin can do
Even that wouldn't be nearly enough. Sensors in human skin measure both external and internal (elastic) pressure, thereby helping the cerebellum in figuring out what position our appendages are in. They also measure temperature, air currents via the skin hair follicles, and report damage via pain receptors.
Plus the whole thing is self repairing and doubles as a configurable heat exchange system. :D
The problem is that I am looking for a robot Jeeves. I earn $60,000 a year. I would need to hire 3 humans to perform that task 24/7. But you want any humans doing that to have a living wage. It is not economic. If I could get a $60,000 robot to perform those tasks (3X Musk's dream), I could afford that.
The problem with specialized robots is that you need so many of them. I don't have space in my home for a cooking robot and a vacuuming robot and a toilet cleaning robot and a laundry robot and a dish washing robot and a handyman robot and a gardening robot and a shower cleaning robot and a window washing robot and a nurse robot. Specialized robots work well on an assembly line where they do the same thing over and over. There are very few tasks in the home that fit that description.
By the time you add up all the specialized robots you need, the cost is going to exceed that of of a robot Jeeves. It will also use much less resources. This is why you will want a humanoid robot
Good luck ever being able to buy a robot that is as functional as a human, even at a restricted set of tasks, for a mere $60,000. You'd probably need to stick a couple of extra zeroes onto that. You'll also need to pay someone to service and repair it, and its useful lifetime will probably be measured in a small number of years before it starts to wear out and become uneconomical to repair (compare to a car, which is relatively simple in comparison, the average life span of which is probably a little over a decade)
"I don't have space in my home for a cooking robot and a vacuuming robot and a toilet cleaning robot and a laundry robot and a dish washing robot and a handyman robot and a gardening robot and a shower cleaning robot and a window washing robot and a nurse robot."
What would be the point? If you had a robot that could do all of those tasks, would you work double shifts, spend more time on the golf course, read more books?
You could subscribe to a meal service that delivers your meals. You could hire a housecleaning service to do all of your cleaning. You could have a maid come in a few times a week and change your linens and wash your clothes for you. Or, if you were making a good salary, just have a gardening service do a general service every week or two and do everything else yourself.
I've made my own bed ever since I can remember. Laundry is some time every couple of weeks (the trick is to have enough smalls to go a fair time between). I do a big cook a couple of times a week with menu selections that keep well and can be tasty when reheated/thawed.
To pay for a robot that will do things I am perfectly capable of doing myself and don't take a whole lot of my time to accomplish is spending money I could use elsewhere. Even after a $60,000 robot is paid for, one service call would wipe out all of the money put away for that year's holiday.
It's going to happen, and probably a lot sooner than you or I think. Only a fool would believe otherwise.
Yes, dear, of course it is, and it will be controlled by that Artificial General Intelligence that has been coming Any Day Now™ for the last 50 years.
Now, have you taken all your pills? Show me under your tongue, good...
There's also the psychological aspect of people reacting to robots in everyday society. I'm guessing humanoid robots will be accepted far more easily than four-armed, six legged, stalk headed freaks.
I guess you've never heard of the "uncanny valley" then, where studies have shown the exact opposite of your guess.
Things that look almost human, but are not, freak people out. The "uncanny valley" refers to the range where something is human-enough looking to look human-ish, but not human-looking enough to pass for actually being human. Either side of that is fine; convincing enough to look human, which is actually very hard, because we have all sorts of wired-in brain functions to recognise human faces, or not human-looking like a roughly humanoid robot, or something with six arms, or on caterpillar tracks, or nothing like a human at all, such as an autonomous vehicle with a robotic arm on it.
All existing vehicles, tools, utensils, weapons, etc are designed to be used by humans.
Except for:
- Autonomous vehicles, such as planetary rovers, drones, automated forklifts in robotic warehouses...
- All the other tools that aren't. For example, I'm not very good at spinning a drill bit round really fast with my hands. Robots themselves fit the definition of "tool", as do any number of electronic devices that run perfectly well without a monitor or keyboard attached. The bit about tools only holds true if you define a tool as something that can be used by a human, which is a bit of a circular argument to say the least. In formal logic terms, that type of argument is known as a closure; it demonstrates nothing other than its own definition.
- I have plenty of utensils in my kitchen that are not designed for use by humans, but for use by other utensils; for example, a hotplate requires tongs or similar to place food on/take food off to avoid burning your hands. This is a parallel to a utensil that is usable by a robot, and "utensil" is literally a synonym for tool in any case, so this is a restatement of the previous item.
- Most sophisticated weapons are not directly usable by a human (and anyway cf the fact that utensils are tools, and so are weapons). For example, a pointy stick is a weapon that is designed for use by a human, as is a spear, but a spear thrower is a tool that makes the spear more effective. Taking this further, you could throw a pebble at someone and maybe hurt them slightly, but a sling makes the pebble a weapon. Taken to the logical extreme, nobody would argue that a nuclear warhead is not a weapon, but it is decidedly not designed for human use (directly, in any case), it must have a detonator, and a delivery system to be of use.
So your argument may be that all tools are ultimately designed for use by humans, but that doesn't hold true, either, because other animals, such as chimps, and crows, use tools.
I can, just about, go along with the fast food bit, providing the range of fast food is severely limited (say just flipping burgers but not making custom sandwiches) but I think it will be a while before shelf stackers can be replaced. There's a lot of variation there which would be difficult for a robot to cope with.
"...but not making custom sandwiches ..."
Many, many fast-food joints can't make custom sandwiches. Many, many shops, garages, supermarkets and McD's can't even make them without mayo and those are (allegedly) staffed by and supplied by humans. There is not a single physical "law" that limits a food-bot to being unable to supply a burger-in-a-bun without mayo. All it would take is a couple of lines of code and humans seem to be adept at supplying millions of those.
True, there are always going to be the accent and dialect issues but I have asked fully trained humans who supposedly apeak my language for a sandwiche wiht butter and chicken *only* and received margerine, mayo and ham so that type of error isn't unique to machinery.
Oh, and "Sorry, but it is now ten seconds after eleven and the breakfast menu is closed, how about a nice burger instead?" is an idiotic paradigm created by marketing droids with the express purpose of annoying us. It's amenable to a bug-fix.
With genetic engineering and "medichanics", there are many, many obvious improvements that could be made to humans, too. You mentioned more arms, but having a microhand, with tool-ends such as screwdrivers and very strong grippers (spanner-like bits) and perhaps their own optical inputs would be nice. Not having dislocatable joints would be cool and knees are a daft design - far too fragile and complex.
The human eye, as you also noted, is a poor design, too. Sure, it works and it has allowed the species to do some truly amazing stuff but it is incredibly limited.
Human skin is far too delicate. Bones, too and the circulatory system is a joke.
The human brain *definitely* needs some upgrades. And the comunications systems could do with some tweaks. Current computers have "telepathy", there isn't any reason why a "better" human couldn't join them.
I'm not entirely sure what a "perfect" child of humanity would look like but it would not closely resemble the mal-adapted mokeys we commonly see at present.
Maybe it would more be like one of those doggy-style robots, but with better manipulators?
"If you wanted to design a robot which can walk efficiently (rather than one on wheels or tracks), why not give it four legs"
Let's go with something like a Puppeteer that has 3 legs and two head/arms with the brain in the midriff. If the development project needed some funding, they could lease out what they have for a Ringworld movie.
"This means a future of abundance,"
Abundance of what? Jobs?
"implying that he hopes Tesla's robots might one day replace humans on production lines"
Oh, maybe not.
"A future where there is no poverty"
How will that work? Will the robots be dumped in a hole in the Congo to mine minerals for their own construction?
This post has been deleted by its author
This post has been deleted by its author
And here I thought a tech audience would wet their pants at the mere possibility of a humanoid robot built by the same person that was told EVs were inviable and reusable rockets nothing short of impossible.
But don't stop on my account and do carry on, I never miss a chance to learn from robotics and business experts :D
Electric vehicles were a thing when I was a PFY. They were called milk floats. Fifty years of explosive improvements in batteries plus steady improvements in motors meant that EVs were a matter of when and not if. The people saying reusable rockets were impossible were guided by motivated reasoning. Rockets used to be affordable - the launch provider checked your budget and decided what you could afford. Reusable rockets were not as inevitable as EVs but a shake up of the launch business was clearly an opportunity for someone with enough money to jump the first hurdles.
The problem here is AI (artificial stupidity). There are a bunch of programmers here and they have very low expectations for the capabilities of AI for the foreseeable future (pre economic fusion power). Take a look at a similar example from the same source: Full Self Driving. FSD has been pre-sold for years as an upgrade that will be ready real soon now. Different people see it differently: some early adopters are furious that they payed a lot of money years ago for something that will not be ready real soon now. Some are dead because they did not realise that full self driving currently means you have to pay attention continuously and be ready to take over at a moment's notice (or with no notice at all). Some are very enthusiastic about their investment and are joyfully providing training data for free so that the dream will come true - any decade now.
I have been holding back, waiting for the actual pre-sales, but others are ready to assume the likely is already inevitable: there will be pre-sales. The robots will be very late. When they finally arrive the functionality will be crap. Despite that, there will be fans joyfully paying tens of thousands for the privilege providing free training data for a product that is going to be less useful than the cheapest unskilled labour.
https://www.youtube.com/watch?v=2NoFVNku01M
Fiat sold BEV 40 years ago, the problem at the time was that lead acid technology wasn't suitable to use on cars, in the advert above is said that NiCd batteries could be installed instead of lead acid.
https://www.youtube.com/watch?v=PKBUXkNdYag
As you can see Fiat also built electric vehicles in 1949, that didn't have any batteries inside, and actually didn't have any electronic device inside. This one, restored to be again rail-worthy has some electronics systems inside, because the tickets now use NFC chips instead of becing a piece of cardboard that is punched like in the 50s.
"Fiat sold BEV 40 years ago, the problem at the time was that lead acid technology wasn't suitable to use on cars, in the advert above is said that NiCd batteries could be installed instead of lead acid."
At the turn of the last century, EV's were popular since they were simple to operate, easy to recharge (if the town had electricity) and were suitable for use by a lady. It wasn't until the invention of the electric starter that petrol cars really got going. Women didn't have to stoop over to crank the car (very improper) and it eliminated the injuries and deaths from kick backs of the crank. The inventor was motivated by the accidental death of a friend hand cranking a car.
The requirements were much different when electric cars were first sold. Travel distances were much shorter and there wasn't the miles of paved motorways we have now. There's no way somebody would think of driving from London to Southhampton like they would today. Once there was a train, that was it. Nobody thought of taking a coach once there was a train either.
With longer travel distances, fewer public transport options and battery technology caught up to the point where it's affordable enough, EV's start making sense again. There's still a long way to go, but cheap supplies of crude oil are a thing of the past and geopolitical issues are much more of a factor.
And here I thought a tech audience would wet their pants at the mere possibility of a humanoid robot built by the same person that was told EVs were inviable and reusable rockets nothing short of impossible.
Electric vehicles were not (except for a few specialised applications) viable with lead-acid batteries, and anyone who said reusable rockets were impossible had clearly never heard of the space shuttle.
> Electric vehicles were not (except for a few specialised applications) viable with lead-acid batteries,
> and anyone who said reusable rockets were impossible had clearly never heard of the space shuttle.
That's my childhood, inverted.
We had milk floats. Everyone had milk floats. They were everywhere. Pure electric vehicles powered by lead-acid batteries.
We also had the Apollo program. It was both amazing and just part of our normality. No one had ever heard of the space shuttle. Rockets were definitely not reusable, and no one suggested that they ought to be.
I'm afraid that I don't get your point.
-A.
I guess there's some value to re-using the engines (although I'm willing to bet they need completely stripping down and rebuilding after each launch). I also reckon that most of the cost of a launch is the fuel, because, in simplest terms, your average rocket is a thin shell of metal around an awful lot of fuel with some engines on the bottom, and a relatively tiny payload on top. Basically one big fuel tank with a (usually, fingers crossed) controlled explosion at the bottom.
"I also reckon that most of the cost of a launch is the fuel,"
Not at all. Fuel is pretty cheap in comparison to all of the other costs and the payload.
The rocket and its fuel are the easiest to calculate but aren't the whole picture or even the biggest line items. There is also the cost to keep the factory that builds them staffed and open. The R&D costs for that rocket has to be amortized and enough built into profit to pay for ongoing improvements and design of the next model to replace it. The launch site charges fees both ongoing and per launch. The launch pad facility is expensive to build and has to be kept maintained in a coastal environment. Legal, admin, compliance, insurance, blah blah blah. The launch market hasn't really grown that much in the last couple of decades. Much of what SpaceX is doing is Starlink launches which is an internal project and unlikely to have been done out of house. I think Pressure-Fed Astronaut has done some adding up of tonnage sent into space over time while analyzing if there is any work for Starship as a rocket for hire. His conclusion is no, but it's an education in all of the reasons why to have a good understanding of why "no" is the answer.
"Out of interest does anyone know what percentage in value terms of a rocket is reusable?"
Mainly it's been about getting the booster back and discarding the second stage that would need thermal protection to get back through the atmosphere. You can have as much as you like back. Anything you want to use over and over needs to have enough margin built in so any wear is negligible. There's a fuel and capability cost for landing legs and spare fuel to be able to run the engines again. The engines need to be designed to be restarted which adds mass.
You could have a car that will travel 800 miles without needing to be re-fuelled, but the car would be much larger to have all of the fuel onboard that it needs and would be far less efficient since you would be hauling a larger proportion of the petrol for a long distance. If you know where all of your intermediate stops are for meals and restroom breaks, you can be much more efficient if your fuel load only needs to take you that far plus a bit of margin. The same sort of thinking applies to rockets. If you build a rocket to push a certain amount of mass up to a certain speed, you can optimize the rocket to be as efficient as possible if you only need it to work for those couple of minutes and no more.
The space shuttle was not even close to being "reusable"! It was as reusable as a Formula 1 racing car. It had to be stripped down and many, many parts replaced between each use. This took many weeks - the record was 54 days but that was unusual and normally it was much longer.
In this context it is quite clear that "reusable" means something closer to an airliner. The Falcon 9 is not there yet but it is closer than anything else that has ever existed in the rocket world - the turn around record is three weeks (and most turn arounds are close to this where as the 54 day shuttle was a clear outlier).
If all the aspirations come true, then the Starship will do it.
If all the aspirations come true, then the Starship will do it.
It hasn't even made it to orbit yet, or played chopsticks. But remember Starship will allow us to book a cheap, suborbital flight from LA to Sydney, and the 'starports' will be able to offer multiple flights per day. That implies a turnaround time similar or better to regular passenger jets at a similar per-passenger cost. So they've got a very long way to go to get turnarounds down to hours, and deliver anything close to what they've promised.
But that's normal for Musk. Putting an airhockey table inside a vacuum tube really isn't that difficult, and you can see this amazing technology in action in Vegas. Where else would a Tesla in a storm drain be considered a miracle of technology?
"and reusable rockets nothing short of impossible."
They've been reusable for a very long time from a technical standpoint. It was the financial aspect that didn't work. SpaceX is building a rocket that is 50% more capable than needed for the bulk of the missions they fly to gain that reusability. Is that a good return? It's hard to tell as SpaceX is privately held and doesn't publish its financials.
I was working with (not just "on") VTVL rockets long before Elon. We heard a rumor that Elon "got the idea" after seeing reports about a NASA Prize for Lunar landers. (Northrup Grumman Lunar Lander Challenge). It's easy to look that up and SpaceX's Grasshopper program. DC-X predates that.
Which begs the question, why didn't you make a video of the one with everything, rather than the rough development one?
Of course, we all know the answer to that is that your claim to have anything better is total bullshit, because only a complete idiot would film the shit robot instead of the good one and show it at their "AI Day event".
The simple answer is to start training AI in how to design better robots - ways to make them more energy efficient, more powerful, more graceful, and more intelligent Then hook that AI to machinery that can actually build those better robots. Then Musk will get the robots he's dreaming of.
Which will promptly kill him and begin a war against humanity.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
