Boston Dynamics releases video of Atlas robot doing parkour:
In the video[#1], two Atlas robots navigate jumps and steps of varying heights, vaulting over and running along a balance beam before ending the routine with two backflips a piece.
It's more than most humans can do, but the team at Boston Dynamics explain that the robots aren't intelligently navigating an unfamiliar course. They've been programmed and trained to complete it.
"It's not the robot just magically deciding to do parkour," chief technology officer Aaron Saunders said.
"It's kind of a choreographed routine, much like a skateboard video, or a parkour video, where it's an athlete that's practised these moves.
[...] The behind-the-scenes peek[#2] at how the often viral videos are made shows the stumbles and fumbles that often result when making a 175cm, 75kg robot do parkour.
[...] While the routines are rehearsed and pre-programmed, Atlas has become more capable at assessing its own environment and its movements now are much more based on perception than in previous videos, Boston Dynamics said.
Videos on YouTube #1 (1m6s) and #2 (5m50s).
(Score: 4, Insightful) by Rosco P. Coltrane on Thursday August 19 2021, @02:13AM (10 children)
The not-so-cool bit is that most of that cool tech is going to be used for military purposes. I wish there was the same kind of money available to develop advanced robots like that to help the disabled or the elderly, or simply do ordinary household chore for ordinary people, instead of helping spread destruction and misery.
(Score: 1, Interesting) by Anonymous Coward on Thursday August 19 2021, @02:40AM (1 child)
Sorry, they're still in Stealth mode.
Toyota is doing it. I asked them why they don't focus on agriculture. In my mind, we're automating the easy-to-automate stuff, and putting those people out of a job -- and telling them to go back to school and get smarter. Some of them can, some of them can't; and then there are the articles about how much of the middle-class salary has gone to the rich via automation.
The Toyota director's response was: sorry, we're focused on enabling the elderly.
It's a work in progress. Japan needs it -- desperately. I still feel that once we automate agriculture, that will free up so many people that care for the elderly won't be an issue -- nor will food for all the people that need to eat.
Toyota Research Institute, however, I don't hear as much about. They're doing it, details aside, but they're just... in perpetual stealth mode. They've got the funds, the means, the reasons, but alas. I just don't hear much more. (They're doing self-driving car development as well, but alas -- Toyota cars don't have self-driving yet. Just like Waymo. The principle in the Toyota case is: virtualize it. run it through 100 million virtual tests, and train the AI so that it doesn't have a problem in the conceived scenario, but alas -- ai isn't human yet, and coming from the self-driving industry, I haven't heard _anything_ about Toyota's self-driving tech.)
(Score: 1, Interesting) by Anonymous Coward on Thursday August 19 2021, @05:10AM
It might seem that way, but you'd be wrong. In developed countries actual agriculture uses about 1% of the labour force. If you include processing and delivery that goes up substantially, to about 10%.
If you include the entire food system; restaurants, fast food, supermarkets, street vendors, everything, you might be approaching 30%. But automating that lot is much more difficult than making a robot pick cabbages. Self-driving vehicles are going to eliminate at least ten times as many jobs as completely automating the entirety of agriculture would.
Don't get me wrong, I am very much in favour of automating away back-breaking labour, but you are not going to free up enough geriatric nurses to care for all the old japanese.
(Score: 0) by Anonymous Coward on Thursday August 19 2021, @02:44AM
The battery makers are the ones who can really transform the cool robots into killer robots. Double the energy density, double the killing potential.
(Score: 1, Insightful) by Anonymous Coward on Thursday August 19 2021, @03:04AM
Social workers never invent shit.
Just be glad somebody is innovating.
(Score: 2) by krishnoid on Thursday August 19 2021, @03:08AM
What do you think this is? The robots are being trained to do parkour, and now the elderly can shake their fists at them and yell at them for being punks while all the kids are inside on their phones and consoles.
(Score: 2) by RamiK on Thursday August 19 2021, @10:59AM
Not really. Capacity for robotic warfare is good enough to get robots building robots. And with solar power and free labor, you can desalinate and greenfly most of the world's deserts at lower-than-warfare costs. Well, disregarding the usual problem... [ucsd.edu].
Btw, Murphy's new 2021 (free) textbook is out: https://escholarship.org/uc/energy_ambitions [escholarship.org] https://escholarship.org/uc/item/9js5291m [escholarship.org]
compiling...
(Score: 0) by Anonymous Coward on Thursday August 19 2021, @11:16AM
household chores? it's still cheaper to hire human help.
(Score: 3, Interesting) by Immerman on Thursday August 19 2021, @03:47PM (2 children)
> I wish there was the same kind of money available to develop advanced robots like that to help the disabled or the elderly,
There sort of is - it's called military spending. A large, bulky, half-million dollar (a wild guess) robot isn't any good to ordinary people, but it lays the groundwork for what will eventually be refined into something that *is*.
Basically, the military has both a much deeper wallet, and a much lower minimum threshold for utility, than ordinary people. So a wide range of technologies are first developed for military applications.
Consider - take the robot as it is (assuming it didn't crash half the time it tried such a routine), and strap a gun and/or a bunch of cargo to it, and you've got something that could be useful in a wide range of military applications. But what would it be good for domestically? Giving piggy-back rides to the elderly? It's a still *long* way from having the dexterity to be able to clean house or play nursemaid. Or even move furniture more complex than a cubic footstool. And the first time it falls over... there goes the table full of family heirlooms.
Only after the basic technology is mature - e.g. the robot is relatively inexpensive and can reliably navigate through the chaos of a typical household for months on end without damaging anyone or anything, *then* people can start focusing on how to use it for domestic services. Until then it'd mostly just be wasted effort. Kind of like SpaceX is currently mostly ignoring orbital refueling, ocean-going launch facilities, and even cargo-bay doors for Starship. Those are all long-term problems, and there's a whole lot of developing and testing still to do just to get the thing to orbit and back reliably before they're worth focussing on. There may even be major design changes between now and then that would render premature development a complete waste of time.
And in point of fact, there *are* other companies working on robots with more immediate domestic applications - for example cooking robots, at least one of which is currently being marketed to fast food joints that can offer a well-organized commercial kitchen, and for whom a cooking robot can potentially replace multiple shifts of workers, saving them several tens of thousands of dollars a year, even at minimum wage. But at present, even if it is versatile enough to handle a home kitchen, there would still be vanishingly few people for whom it would be a cost-effective option. Far cheaper to hire someone to come by a few times a day to cook a meal, or just order delivery, than pay for a robot that will sit idle most of the time.
For domestic robots to become remotely cost effective, they probably need to be a drop-in replacement for a human servant at a much lower cost. That probably means human-like range of motion, dexterity, and versatility. And we're a long, long way away from that. In fact, I'd be willing to wager that virtually all low to moderate skilled jobs will be automated out of existence long before anyone is able to develop a home robot that's remotely versatile and inexpensive enough to be anything more than a novelty toy for rich people.
After all, any robot that can handle a wide enough range of household chores to be useful should be able to handle all the same "chores" for a business, replacing several shifts of workers and, assuming service life is relatively high, freeing up dozens of people-years worth of salary and benefits to pay for its purchase and maintenance while still saving money. Meanwhile very few people can afford even one or two years worth of salary for a domestic servant, even if they hadn't already lost their job to a similar robot.
(Score: 1, Flamebait) by Rosco P. Coltrane on Friday August 20 2021, @10:25AM (1 child)
Okay, let me rephrase this:
I wish massive amounts of public money was spent developing socially-useful products, as opposed to massive amounts of public money being spent developing tools to kill, maim and make entire populations' lives a living hell, hoping some of it will trickle down into socially-useful products.
(Score: 2) by Immerman on Friday August 20 2021, @09:22PM
You and me both. Sadly, socially useful products are mostly much less profitable for our "betters" that are making the decisions. After all, the purpose of war spending isn't ultimately to kill, maim, and make people miserable - that's just a stepping stone on the way to the pillaging and oppression that enriches and empowers the select few who "matter"
And less spitefully - there's just far less demand for cutting-edge socially useful products. The people that could afford the first few generations of technology already have human servants that do the work for them anyway, better and with with far more ego-gratification than any clumsy robot (or whatever) could provide. So even if we dumped tax dollars into developing such technologies - the amount of money and time would have to be vastly greater to reach the point where the market could begin to take over. How many generations of nursemaid robot would have to sit basically unused, or maybe raffled off to get user feedback from people who couldn't hope to afford them, before they matured to the point that the upper-middle class would consider them a worthwhile investment and start really driving cost savings from large economies of scale.
High technology just really isn't something within reach of the mass market - it's only after the underlying technologies have matured and cheapened to the point that the middle-class can afford products based on them that they have an opportunity to take off.
(Score: -1, Offtopic) by Anonymous Coward on Thursday August 19 2021, @03:23AM
Hey dog... DOG! Did YOU see the size of that chicken?
(Score: -1, Spam) by Anonymous Coward on Thursday August 19 2021, @03:30AM
Feed me all night long........ Heh, THAT'S RIGHT BOY!
You can do it!!!!!! FEED ME SEYMOUR feed me all night long!
Cuz if you feed me seymour, I can grow up, BIG, AND STRONG
(Score: -1, Troll) by Anonymous Coward on Thursday August 19 2021, @03:34AM
"I'm hot and I don't even have to try."
That's sexy.
Boston sucks - too many patties and guidos - just as inbred as alabama.
(Score: -1, Troll) by Anonymous Coward on Thursday August 19 2021, @03:49AM (1 child)
You put up bullshit post, you get bullshit comments.
(Score: -1, Troll) by Anonymous Coward on Thursday August 19 2021, @04:35AM
Still salty, Eth?
(Score: 0) by Anonymous Coward on Thursday August 19 2021, @12:46PM (1 child)
Parkour is stupid enough when humans do it.
(Score: 2) by pkrasimirov on Thursday August 19 2021, @10:50PM
You can say the same for playing DOTA2 but it is impressive achievement for machines.
(Score: 0) by Anonymous Coward on Thursday August 19 2021, @12:53PM (1 child)
The mechanical/servo/micro_goal_sw is coming along nicely.
(And a good marketing person keeping it in the news.)
Kind of like a self driving car can do this little stuff pretty well.
(Start, stop, stay in lane, park, avoid mostly.)
The high level goal stuff is very much a work in progress.
(Like for the emergency vehicle parked and hit by a self driving car.)
These bots have the power to throw quite a bit of mass around quickly.
Don't think I would want to be physically anywhere near one that's powered up.
Think what will happen when the AI goal director gets as good as the car driving goal director.
It will be around and safe mostly.
(Kind of like todays 'self' driving cars.)
As a transitional plan, supposedly, the car has a driver with a stop button.
What's the plan for the bot's stop before is hurts someone button?
(Score: 4, Insightful) by crafoo on Thursday August 19 2021, @01:29PM
I don’t think the people with the bots and the power care all that much about individual human life. Consider a world that is all Africa hellholes, Islam caliphates, and flavors of China 2.0 fascism. Freedom and wealth have been slowly suffocated out of the majority, who now live like obedient herd animals in their pigpens, “masked-up” of course for safety.