Tesla promises to build robot you could beat up – or beat in a race During what felt like an hours-long job ad, Tesla on Thursday pledged to build a humaniform robot prototype. Screenshot of the Tesla bot presentation Behold, the Teslabot ... Click to enlarge CEO Elon Musk announced the project at the automaker's AI day, saying Tesla already does most of the things a robot needs – sensing, …
Surely, you know that "Semi" is the first step to achieve "Full". Just like the westworld thingy, they need time. It may be "Semi" bollocks now, but they will evolve into "Full" bollocks and conquer the world in a real bath of blood. I've seen it on television, so it must be true!
Of course the beauty of this term is that "semi" has an infinite range of vernacular definitions. It's strict meaning of "50%" is probably not even open to formal specification (which 50%?), so what we've been handed is a moveable PR feast (much as usual from the Musk(k)eg).
in a UK office the phrase "I have a semi" will have HR come after you
If you said that near me I'd ask if you what colour your half of the drainpipe was.
Musk's fantastic statements are not intended for the consumption of techno-savvy individuals, it is put out there to inspire the less well informed to part with their hard earned currency.
As far as his humaniform bot is concerned, conventional thinking may be more aligned with the idea of more or less built to purpose forms for robots but nature has already shown the human form to be a useful general purpose design.
Musk doesn't need millions of years to come up with something approximating a human as he has models to look at as a basis for applying existing tech to make something similar and then improve subsequent models.
So long as it can recognise pedestrians and trucks......
> nature has already shown the human form to be a useful general purpose design
Indeed, universality is the point here. Not to mention everything in our society is built to cater for human-shaped beings, so a human-shaped robot will literally "fit in", from furniture to using tools and man-machine (in this case machine-machine) interfaces.
Well, in an interview where he was walking around the new Starship build, he did say his deadlines were super optimistic, then he said "if I wasn't optimistic, then I wouldn't have tried to land boosters on barges, or build the biggest fully reusable orbital rocket. Everyone has repeatedly told me how impossible all that is."
However, I don't see how this robot "gets us to Mars"
They do sound a little weak for building jobs. The gravity is weaker on Mars, but still, what's the point of spending big money to send there a construction worker who can't lift reasonably heavy stuff?
I don't think this model is supposed to ever leave earth. Maybe a subsequent model built for defense contractors, and which has more impressive physical capacities? (Somebody has seen Sarah Connor lately?)
And while robots aren't totally immune to radiation damage, they are decidedly less fragile in many ways than real humans. Moreover, if things go badly, they are very unlikely to file a lawsuit re damages caused by their employer's callous disregard for worker safety.
However, that's probably a moot point as it is unlikely that Musk or anyone else will come up with a useful humaniform robot anytime soon. Consider that Tesla has been working on vehicle autonomy for a decade or so and so far all they have to show for the effort is an overhyped and none too reliable collision avoidance system, considerable bent metal, and a few dead bodies. Useful humaniform robots look to be a much more difficult problem. A problem for our great, great, ever so great grandkids most likely.
Humaniform, is just that, shaped like a human, I don't think there is a need for a fully aware and automomous robot just yet. All they need is to be relatively versatile and programmable for various tasks that require limited autonomy.
Positronic brains or whatever equivalent develops are not that necessary and way too complicated for what will amount to a tool.
I recently re-read the I robot series Asimov was way off on his ideas for robot psychology and cognition.
Well Boston Dynamics have a reasonable facsimile up and definitely running:
https://www.theregister.com/2021/08/18/boston_dynamics_parkour/
However they are rather quite about how long they can operate on internal battery power which is I suspect where the problems will arise for anybody considering robots of any kind.
"Robots don't need food, water, oxygen, etc for a 9-month trip to Mars."
They still need energy. If there were a safe portable power storage device to power these robots, we'd already have those exoskeleton suits that were in the Alien movies. To me, those make more sense. Without positronic brains, we aren't going to get R. Deneel Olivaw no matter how much pot Elon smokes.
It may not get us there, but could be useful when we are there. Extra pairs of hands that only need electricity, not food, air and water, could be quite handy.
Musk isn't just focused on getting to Mars, he's also thinking about living there, not just a "boots and flag and go home" mission that most governments would come up with. In that context a humanoid form could make more sense in terms of mass budget than a dozen different specialised robots.
Don't get me wrong, I'm not saying it's easy. But it's good that Musk is willing to have a real go at doing difficult things.
In that context a humanoid form could make more sense in terms of mass budget than a dozen different specialised robots.
Everybody knows that sending humans to Mars, or even to the Moon, is practically pointless, because robots can do ten times the work for a tenth of the price. Landing humans elsewhere is political grandstanding and that's it. So, given all that, there is no need to make robots for Mars look humanoid. As the current ones crawling around show.
"Everybody knows that sending humans to Mars, or even to the Moon, is practically pointless, because robots can do ten times the work for a tenth of the price."
Steve Squyers' book "Roving Mars" on the Mars Exploration Rovers, he was the Principal Investigator, makes a good argument for going with robots for some things and humans for others. A geologist can do more exploratory work and identify more things of interest in a hour than one of the Mars rovers could do in a week. The book is very fascinating and worth a read. It's not pointless to send humans to the moon, it's just good to optimize what they are doing that robots can't do or can't do very well. I do think it's pointless to send people to Mars right now as the trip is too long for people to arrive in good health. If nuclear propulsion becomes a reality, the trip time could be cut way down and most of the trip done under acceleration to avoid the wasting effect of zero G on the body. Until then, robots are the best option.
I'd like to see some workable approaches to putting sun shades around Venus.
... is a standard prefix to a marketing lie. The people who do not know are supposed to shut up to hide their ignorance. Next time you could try "Your neighbours have already signed up...", "Limited time offer..." , "All inclusive price" or "The pea is under one of these three shells".
"Don't get me wrong, I'm not saying it's easy. But it's good that Musk is willing to have a real go at doing difficult things."
It's not that it's difficult, it doesn't look possible at the current state of engineering art. You can't railroad until it's time to railroad. The other consideration is "what is the end goal" for this sort of robot. With a big one-armed industrial robot, it's easy to see that you can attach a whole assortment of things to the end of the arm and have it do work. It's not problem for one of those to swing around a spot welder weighing 100kg. Do that with something human sized and it's welder one way and robot the other in a simple mass ratio mechanics problem. The big robots have better strength, longer reach, more precision and the ability to get stuck into jobs that are hazardous for people since we aren't as predictable.
"one thing he is consistent with are his delivery estimates. "Optimistic" is a massive understatement. "Fantasy" might be a better word."
Looking at his Wikipedia entry, I'd say he's been responsible for 7 deliveries which were probably not far off the predicted lead time. Of course, it helps that it only needed his direct input at project initiation and the rest was left to whoever was Mrs Musk/Musk-Muse at the time to take to completion and delivery
I fully expect :
- that this will take a lot longer than Musk says to hit the market (if it ever does)
- that the bot will not be able to do anything more useful than carry something or stand guard
- that it will have to be plugged in all night, else it will run out of power
- that you'd better not put it in a crystal shop
We don't have the technology for a fully-autonomous, humanoid robot. Hell, we can hardly make a dog-like robot that doesn't scream its presence to everyone within 100 meters.
This is a pipe dream. Not going to happen.
Boston Dynamics robot spot dog,has been used by SpaceX for months,who nick named it Zeus.
It's not a great leap of tech from dog shape to human shape.
Yes Musks timelines are always optimistic, but he is the epitome of shoot for the moon and even if you miss you'll land among the stars, he pushes people to think nothing is impossible, and sometimes they do achieve impossible things.
Which is alot better than the if we cant do it litigious style blocking approach his competitors favour.
Humaniform? Autonomous? The current state of robotics is "we're struggling to make something walk bipedally when physically strapped to power and processing capacity", and Elon reckons he'll have a prototype next year?
What's the use case for this? We already have robots that do "repetitive, dangerous tasks", but there's no need for them to be humaniform. Butlerbot maybe????
Hmm - maybe it'll be an autonomous chauffeurbot for Tesla owners...?
Well, a humaniform robot might be suitable for search and rescue, firefighting, stocking/fetching goods in a store with a diverse inventory with widely varying packaging, etc. But I have to think that in each case, a purpose built device with suitable attachments would likely be cheaper and more reliable.
This simply confirms my theory that Elon Musk read far too many "Tom Swift Jr" books as a child.
Tesla Cars? Tom Swift and his Triphibian Atomicar.
SpaceX? Tom Swift and his Rocket Ship
The Boring Company? Tom Swift and his Sub-Ocean Geotron
Hyperloop? Tom Swift and his Repelatron Skyway
and so now he has reached Tom Swift and his Giant Robot.
"This simply confirms my theory that Elon Musk read far too many "Tom Swift Jr" books as a child.
Tesla Cars? Tom Swift and his Triphibian Atomicar.
SpaceX? Tom Swift and his Rocket Ship
The Boring Company? Tom Swift and his Sub-Ocean Geotron
Hyperloop? Tom Swift and his Repelatron Skyway
and so now he has reached Tom Swift and his Giant Robot."
Tesla cars: Detroit Electric, Baker electric, GM EV-1, 97-98 Ford Ranger, Toyota Rav4 EV
SpaceX: NASA, ULA, Armadillo Aerospace, Masten Space Systems, North American Aviation, Boeing
Boring Company: Brunel and a world full of subways, sewers, the London underground mail line
Hyperloop: Patented years ago as the Vactrain (1945, Robert Goddard. Yes, that Robert Goddard)
The best that can be said is Elon has popularized the EV. Building the Supercharger network went a long way to allaying fears of not being able to charge a Tesla but kneecapped itself by going with a non-standard proprietary interface and plug. As far as SpaceX goes, landing on a barge is very impressive, but everything else had already been done by multiple entities long before. See "Rocket Propulsion Elements" by Gary Sutton for a good primer. My copy is autographed.
For very critical looks at Elon's projects, see "Adam Something" on YT or CommonSenseSkeptic. They even show the math.
I see money from governments going to WAY LESS BENEFICIAL things all of the time. Might as well do something FUN with the money, other than empowering politicians and paying off political debts...
Sure. Let's build a Sci Fi world based on Elon Musk's (sometimes crazy) ideas. Why not? MUCH better than "clown world", I bet. [and it could be a jobs bonanza]
optimism sometimes has interesting positive effects on EVERYONE. I mean if YOU saw someone in a jet pack flying by, wouldn't YOU be inspired? OK I stole that line but seriously...
Many of us work in development so we're used to mentally correcting management schedules to approximate realistic time. So the fact that Musk's projects are often behind schedule doesn't bother me, especially given what his companies are trying to do. The fact is that when a company like SpaceX delivers it really does deliver. Its now got an incredible capability, reliably launching payloads at a fraction of the cost of other launch vendors. Tesla not only delivered viable electric vehicles but also the entire infrastructure needed to build and manage them. Much of what is done by Musk's companies is truly leading edge so I modulate my expectations accordingly -- after all, to quote Musk, "Self driving vehicles are hard" so only the truly ignorant would switch on Autopilot and expect it to reliably take them home while they napped in the back seat.
As for Mars, the more we know about the place the less it looks like a viable place for humans to visit, much less try to live there. I live not far from the Mojave desert and desolate as this might be its positively verdant compared to Mars. We humans might be uniquely well adapted for thriving on Earth but we're not going to be able to survive there without a whole lot of technology, so much so that it might be better to just send the machines and leave the wetware at home.
"reliably launching payloads at a fraction of the cost of other launch vendors. "
The big question is if he's doing it at a profit. SX raises billions a year in funding drives. The math on Starlink looks really dicey and may wind up being something that will never pay back unless they can flog it off in an IPO since any private investment company will be able and will run the numbers. Starship is a long way from being viable. They have only got one empty shell to not crash. I've got the trigger ready to pull on a long telephoto lens rental and a bomb squad suit to go watch the full stack from South Padre Island if there is enough prior notice and I can spare the time.
I do see some merit in letting the rocket body assume it's natural falling orientation rather than fighting to keep it upright, but being upright means the fuel/oxidizer will move to the bottom of the tanks with a bit of persuasion and there isn't a radical maneuver right at the end. Something that's not going to work with people on board and certain won't work at Mars. In fact, decelerating and landing on Mars will take a whole different approach and so will landing on the moon.
To hell with a sex robot. Sex is so easy that most little boys do it all by themselves.
What I want is a household robot which can pilot my flying car, dig the spuds and mow the lawns (and take care of the equipment required), do the dishes and the laundry (including properly putting away both), fetch me the snail-mail and a beer, and change the sprog's nappies ... all in one unit.
Has he not looked at Honda's Asimo. That was truely great. Had noticed not heard in mentioned in a while and it would appear because Honda completely canned it in 2018. So if they'd gotten that far and canned Asimo, then what is Elon thinking? Can't help but feel he talks bollocks just to get more investment.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
