Boston Dynamics has always built robots with agility few others could match. While great for attention-getting demos, from outside the company it hasn’t been clear how they’ll translate acrobatic skills into revenue. Now we’re getting a peek at a plan in an interview with IEEE Spectrum about their new robot Stretch.
Most Boston Dynamics robots have been research projects, too expensive and not designed for mass production. The closest we got to date was Spot, which was offered for sale and picked up a few high profile jobs like inspecting SpaceX test sites. But Spot was still pretty experimental without an explicit application. In contrast, Stretch has a laser-sharp focus made clear by its official product page: this robot will be looking for warehouse jobs. Specifically, Stretch is designed to handle boxes up to 50 lbs (23 kg). Loading and unloading them, to and from pallets, conveyer belts, trucks, or shipping containers. These jobs are repetitive and tedious back-breaking work with a high injury rate, a perfect opportunity for robots.
But warehouse logistics aren’t as tightly structured as factory automation, demanding more adaptability than typical industrial robots can offer. A niche Boston Dynamics learned it can fill after releasing an earlier demo video showing their research robot Atlas moving some boxes around: they started receiving inquiries into how much that would cost. Atlas is not a product, but wheels were set in motion leading to their Handle robot. Learning from what Handle did well (and not well) in a warehouse environment, the designed evolved to today’s Stretch. The ostrich-like Handle prototype is now relegated to further research into wheeled-legged robots and the occasional fun dance video.
The Stretch preproduction prototypes visible in these videos lacks acrobatic flair of its predecessors, but they still have the perception and planning smarts that made those robots possible. Those skills are just being applied to a narrower problem scope. Once production models are on the job, we look forward to reading some work performance reviews.
[via Ars Technica]
Welp, there goes my job.
At least five times cheaper than minimum wage.
It won’t be long before people start looking at “universal basic income” a bit more seriously.
There’s still the robot repair job everyone talks about when automation is discussed. And then there’s “programmer”, because, hey, if one can’t be a doctor, lawyer, etc, one can become a programmer. Almost as popular as seat belt tester.
How this comment reads to people with knowledge of economics:
“It won’t be long before people start looking at free energy a bit more seriously.”
Having robots do the jobs that can just as well be handled by people is less efficient.
The people still need to live and eat, and the robot costs money to keep, so you have two entities to maintain and yet the output is not doubled. Why, because the demand is limited to the people and the total amount of work done does not increase as long as the people do not consume more, especially people who are out of jobs and can’t pay for the work of the robot.
By the same logic, we should get rid of all of those fancy contraptions that farmers use these days and rehire all of the pickers and field hands that were put out of work by the “robots” of the last century. Or, go back only 40 or 50 years ago when companies where stuffed with hordes of clerks that were busy all day with shuffling/sorting/filing little bits of paper. When computers came to the workplace, we needed far fewer of those types of humans to run a typical office.
From a company’s perspective, when that human is no longer needed in that warehouse or clerk job, and the robot will do the tasks more cheaply and that human ISN’T drawing a salary from the company anymore, then the company saves money. When a company can produce it’s products more cheaply, they reduce the prices of their finished products and society as a whole benefits because we get goods at lower prices over time. (And, any company that saves lots of money and DOESN’T pass those savings along to consumers will find themselves out of business – someone else will come along and eat their lunch. That’s just how it works (and has always worked). )
Your implied assertion that a particular human being is tied to one job at one specific company and can do nothing other than that one job at that one company is flawed. Just as happened with the pickers and clerks of the last century (who, at the time, made up a huge portion of the US workforce), these few warehouse workers displaced by a robot that can stack boxes will find work elsewhere. And, over time, those new positions will be higher value and higher paying because of the constantly increasing efficiency gains such as the (potential) gains described here. That’s just how it works (and has always worked).
There may indeed be some temporary discomfort for some small number of people for some small amount of time, but that’s the cost of progress. The alternative is that we freeze the status quo and never allow improvements in the quality of life for everyone. Imagine if such a scheme was implemented a hundred years ago… or maybe three hundred years ago… :-)
>By the same logic, we should get rid of all of those fancy contraptions that farmers use
In principle, yes. In the past, when work was provided for welfare for the unemployed, governments barred the use of earth moving machinery around state worksites in favor of shovel gangs.
However, it’s not an argument that we should make people do hard backbreaking labor. We can choose to pay more to avoid that. It’s only that having robots do the work isn’t actually any cheaper or “economical” like many claim.
>and that human ISN’T drawing a salary from the company anymore, then the company saves money.
The company loses that money in increased taxation and other social expenses to pay the welfare that is necessary to support all the displaced workers. This again costs more, since it employs a middle man in the bureaucrat and state worker, and the state itself which is apt to squander the money on all sorts of “projects” instead of paying the poor their living.
“…having robots do the work isn’t actually any cheaper or ‘economical’ like many claim.”
It’s actually *measurably* cheaper (or “economical”, if you prefer). Your statements apply only to static closed economies and those only exist in Econ 101 textbooks. :-) You can include any social non-economic factors you wish, but on either the micro- or macroeconomic level, it’s always cheaper and “better” (ie: higher standard of living for everyone) to have technological progress that improves productivity.
The part that people always forget in this arguement is that you eventually cover the capital cost of the robots, therefore you have ownership and at the end of their life they have reclaim/scrap value – something humans don’t give a business.
A UBI decouples production from consumption, which means that people are no longer burdened by the need to “make ends meet” and population growth is left unchecked. As soon as you institute UBI, you also have to start looking to control prices and production, reproduction…
The basic flaw of the utopian visions of people having to do no work for their own upkeep is that most people don’t really have anything better to do. It’s either work, or drink and make babies.
And under UBI, the more babies you have, the more money you have at your control because each of them has their own UBI…
The whole success to using robots and avoiding the UBI is to never say you are using a robot in the first place – job done.
Fight For Fifteen, get replaced by a robot that can work all three shifts without taking a bathroom break.
so you’re view is either be a slave barely getting by working yourself to death or just let yourself be replaced without a fight?
that really shows the limitations of the capitalists mindset.
This describes pretty well what “knowledge in economics” actually means: taking some virtual world (the economist’s world) with money, property, spherical cows^H^H^H^Hconsumers and other stuff for the “real thing” (i.e. energy, matter, that stuff engineers and scientists delve in).
Economy is a “social science”, not a “hard science”. Which would be OK, because social sciences do have their place, except that economists insist in being a “hard science” and in inventing “laws” which they phantasise to be at the level of “laws of physics”, like, for example, the “Law of Conservation of Money” or something.
This, my friends, can definitely be characterised as psychosis
“Psychosis is a condition of the mind that results
in difficulties determining what is real and what
is not real.” [1]
[1] https://en.wikipedia.org/wiki/Psychosis
“spherical cows^H^H^H^Hconsumers” maybe your browser or localized language input put those “^H” in there.. or you typed them yourself, but it reads to me as a Max Headroom vocal glitch. :)
https://en.wikipedia.org/wiki/Backspace
“Terminals which did not have the backspace code mapped to the function of moving the cursor backwards and deleting the preceding character would display the symbols ^H”
Unloading trailers at 3 seconds per box, looks like they’ve met logistical standards. Glad I’m out of that industry.
Particularly 3 seconds per empty box (by the look of how the boxes float around when released)
It’s an interesting robot, and in my mind this is the Boston Dynamics robot with the most obvious use case. That said, fears of wholly replacing of warehouse workers are overblown here. Obviously a robot like this can have a marginal effect on the people who are doing work like this, but for now I see it more as the day laborer (temp worker) job in the warehouse where little training or knowledge is necessary.
Having spent a year and a half as a shipping clerk in a warehouse, moving product is a very analog experience. Can this robot load up a trailer that has uneven flooring (even holes in the floor the wheels can fall through)? Will it recognize the trailer is contaminated with leaks from previous shipping runs? Can it recognize when a box on one of the lower layers of the pallet is crushed? Can it pick up and repalletize boxes that have shifted/tipped over during shipping? Can it find the misplaced pallet in the warehouse necessary to finish the load ordered by the customer?
One of the robots most often seen in manufacturing is the palletizing robot that takes crisp new boxes off the assembly line, stacks them on a pallet, and shrink wraps them. It’s a great job of automation since it’s hard on a human body to do this lifting and the parameters are strictly controlled. Yet a lot of the steps leading up to that (assembly, placing in retail box, etc.) are often still done by humans because they can be so variable. Warehousing becomes infinitely variable the further you get away from the manufacturing source, so transferring and distributing won’t be nearly as controlled.
This robot is like that palettizing robot on wheels. Robots can and do move product around warehouses already. But it’s not yet a solved problem to cut the human out of the equation. My hope is that we’ll find a medium where the warehouse jobs are still necessary because of the problem solving needed, but the physicality of the work is eased by automation so people don’t have to wear out their bodies so much in these roles.
“Can this robot load up a trailer that has uneven flooring (even holes in the floor the wheels can fall through)? Will it recognize the trailer is contaminated with leaks from previous shipping runs? Can it recognize when a box on one of the lower layers of the pallet is crushed? Can it pick up and repalletize boxes that have shifted/tipped over during shipping? Can it find the misplaced pallet in the warehouse necessary to finish the load ordered by the customer?”
These are great questions and they need not be rhetorical – they have relatively simple answers. :-) The short answer is “no”. The slightly longer answer is “not right now”. The way progress works is rarely that someone has an idea and the next day the world is transformed. It will take time for this particular product (and the ones that come after it) to mature and fit existing needs or for existing processes to change to the new tools available. Perhaps today, the only thing being replaced is that one guy (we’ll call him “Fred”) that physically pulls boxes off a pallet and puts them on the feeder into the truck. By the end of the year, Fred’s position will be eliminated. But since there was another guy in the shipping department that was retiring anyway, Fred will move into his job. Over time, the shipping department will be reduced from 12 people to 9. And then, later, to 6.
At the same time, sensors and AI will improve and more of the answers to the questions will become “yes”. The robots WILL be aware of the conditions proposed. And over time, the shipping department is reduced all the way down to 3 people. But that might not happen for 10 years. But when it does happen, those 3 positions spend their time on fixing problems and implementing improvements, NOT mindless moving of boxes from one place to another. That sounds like a win-win to me. :-)
You make good points but I would argue your timeline is way too short. I’m constantly skeptical about the pace at which computer vision/recognition systems are advancing.
>That sounds like a win-win to me. :-)
It also sounds like there will be a horde of people who don’t have the skillsets suitable for higher and higher management jobs, kicking beer cans down the street, reduced to living on a government allowance with nothing better to do than make mischief.
After all, you have to remember that half the people out there have and IQ below 100 and an education to match that.
It’s a normal distribution with the median and mode at 100. Most people have an IQ within a standard deviation of 100, and that will always be the case because we define IQ from the average. It’s a relative scale not an absolute one.
From your comments, you seem to have a dim (and condescending) view of the general populace. That’s unfortunate because people are generally good and productive and self-motivated. (Have a view of history for billions of examples. :-) )
Interesting, but I’m sceptical on how it’ll work on real-life boxes. There’s no way the boxes it moved at the start contained spots. They were empty.
A lot of the boxes which reach me would fall apart if lifted by one side.
How’s its detection of broken boxes, or boxes tearing whilst it lifts them? Seems like you’d need a lot of humans keeping an eye out for problems. – much better job than lifting though!
Boston Dynamics (now Hyundai) is hastening a bleak future. “Oh, but it’s happened before!” No it hasn’t.
“Humans Need Not Apply” – CPG Grey
“Manna” – Marshall Brain [sic] (Stop when the Australians show up. It’s just an ad after that.)
UBI will never happen. Automated leisure and luxury for the rich, Terrafoam cells for the rest of us.
I can not imagine the overgoods/leakers from poorly designed packaging.
Just yesterday I had a 50lb box full of trailer hitches burst out of the single walled box they had put it in.
Put that machine in a real world UPS trailer and you will see some serious redesign coming around.