Shelf Actualization

If you are old enough, you may remember that, for a time, almost every year was the year that home video was going to take off. Except it never was, until VHS tape machines appeared. We saw something similar with personal computers. Nowadays, we keep hearing about the home robot, but it never seems to fully materialize or catch on. If you think about it, it could be a problem of expectations.

What we all want is C3PO or Rosie the Robot that can do all the things we don’t want to do. What we usually get is something far less than that. You either get something hideously expensive that does a few tasks or something cheap that is little more than a toy.

Labrador Systems is trying to hit the middle ground. While no one would confuse their Caddie and Retriever robots with C3PO, they are useful but also simple, presumably to keep the cost down which are expected to cost about $1,500. The robots have been described as “self-driving shelves.” You can watch a video about the devices below.

Next Stop

Essentially, the robot learns a number of stops and can move between stops using voice commands or an app. One of the robots can move up and down and pick up trays. The other just moves from point to point carrying whatever you put on it as long as it weighs less than 25 pounds.

The robots can avoid obstacles using IR and optical sensors, so they aren’t just running a preprogrammed course. They also take care of charging themselves.

It might seem like a first-world problem to want a robot to carry two cups of coffee from the kitchen to the living room. However, for the elderly or people with limited mobility due to injury or illness, having a robot helper could be a big boost to their quality of life. Sure, you can get a human helper, but that’s become an expensive proposition. A $1,500 robot that works all day, every day is a bargain.

Faster Food

Rolling shelf robots might really take off in the service industry. For example, people generally don’t want to work in fast food. That’s why there’s typically a huge turnover and you don’t really expect to see a motivated kid at the burger joint register who’s building a career. We’ve already seen robots that can work the grill. Now, rolling shelves are serving food to customers, too.

This isn’t a one-off stunt, either. That particular restaurant is in Boston, but they are doing something similar in Dallas, too.

Not C3PO

We have no doubt that we will get to C3PO if not Commander Data, but for today it does seem like the smart money is to pick something annoying but somewhat easy and focus on that. Industrial robot arms, for example, have been very successful. Robot vacuums are hardly brainy, but they do what they do well enough and are popular. You even see industrial versions plying the aisles at many big box stores. It wouldn’t be entirely strange to see the same thing bringing you your burger.

Sci-Fi Aside: What About the Jobs?

Science fiction has worried about what will happen to jobs when computers or robots take them all over. A lot of future fiction posits a system where money isn’t a thing or there is some universal basic income. Maybe robots will actually be what gets us there.

Imagine if corporations could not own robots legally. Instead, they would lease them from workers. When you are born you are assigned a robot. Not that you’d take possession of it any more than a futures trader that buys pork bellies expects to receive a truckload of bacon. The lease on the robot pays for maintenance, replacement insurance, and your basic income. You can save some money and buy more robots or more valuable robots to generate more income. It would be like the stock market, in a way.

On the other hand, maybe robots can do the jobs no one wants to do. Humans could be free to invent, create, solve problems, and be there for other people — all things robots are not great at doing, at least in the foreseeable future.

Or maybe that’s a lot of utopia for a rolling shelf to carry, but we’ve got to start somewhere. So what else is out there waiting for us to automate? Think small. Think ordinary. What’s stopping you?

54 thoughts on “Shelf Actualization

      1. That’s not wrong. I mean look at 3d printing. Scott Crump kept bragging about his dungeon but always refused to invite Dr. Adrian Boyer. 20 years later having a 3d printer in our own homes isn’t even a big deal!

    1. I’d be realy interested in your objections to “Robotic Sexual Assistants” (RSAs). to coin another euphemism.

      Call me a godless monster if you want but if they ( a female RSA) would give pleasure, help the economy and save women (usually) from what is apparently an unsavory task in providing sex for money, then why not? I think a male version might provide more difficult engineering challenges perhaps.

      1. It devalues men and women as partners, making the original problem of lonely people in need of a partner worse.

        Perhaps the only remedying factor about them is that they’ll probably be too expensive to maintain on a single salary.

        1. Plus, introducing automation at every level and niche of the jobs market is already pushing people to become dependents on limited incomes. The problem is that the social cost to maintain a robot AND a displaced person is greater than the cost of the person doing the work.

          Employing automation for the sake of automation (and private profits) does not increase general well-being in the society because it does not increase real productive output: real value production is still limited to consumption, and consumption is limited by income. Two things are happening instead: more people start to depend on welfare, and the rest come up with make-work and fake-work which is wasting real value in order to earn money.

    2. I sincerely hope it will soon work for robotic humanoids. And that multiple of these robots can cooperate on tasks. And that there will be housekeeping apps for these robots as kind of a screen saver.

      Why do I hope that?

      Last year I jumped in 4 months as a caretaker for a 82 years old engineer, who had a streak of bad luck: his wife passed 3 years ago, he had a stroke 2 years ago, and last year the handrail of the staircase came off and he fell down the staircase.

      I’m completely underwhelmed and underimpressed of the thing which is called senior citizen daycare. And me being as old as him now is only 33 years away.

      In any case I would prefer robots over those people, who often can not speak the local language (google and siri are way better), have a lot of time pressure, and are generally trimmed to only check billable boxes on task lists.

  1. “Robots” in the traditional sci-fi sense have come to substantial fruition because the traditional “robot” was based on the reality of the time; that the computer which operated it was incredibly expensive, large, complex, and hard to fabricate. So that “robot” was a necessarily flexible utilization of a “do everything” computer – an actuator/sensor platform for the single “brain”. This lends itself to great dramatic use (“The Evil Robot of Dr. X” or some such), but has little real world application, particularly given Moore’s Law that makes “the brain” ever cheaper and more flexible.

    In fact, robots are everywhere from your thermostat to your car engine, but the “brain” is so cheap, fast, and ubiquitous (at least when the supply chain is running well), that you can afford to have a separate robot for each job. For less mundane things such as a Roomba or an industrial robot they get a little more anthropomorphic, but a full-on humanoid (or canoid if you’re Boston Dynamics) robot has limited widespread advantage, even if they’re very cool.

    In industrial manufacturing “cobots” are beginning to flourish but they usually aren’t much past “select an item and do something with it” – more flexible than a dedicated processor and built with ease of training in mind but incredibly dull and repetitive. You can hire people to do the same thing but human error rates and absenteeism creates a market for flexible devices to do those awful jobs nobody usually wants.

    So you may see a fleet of industrial robot arms building cars, cobots sorting and filling packages and the like, but you likely won’t see a crowd of completely humanoid devices doing those same simple jobs any more than Rosie the Robot would be changing the room temperature setting with her “hand”.

    Outside of dramatic constructions, simpler and cheaper always wins the day. Sadly or perhaps fortunately, beyond “stabby Roomba” and its ilk, Rosie or the T-1000 Terminator isn’t likely to turn up any time soon.

    1. As someone who is a tad challenged in everyday tasks (tl;dr: hemiplegic) I’d welcome a personal assistant robot which can carry around stuff for me, e.g. while running errands.

      I’ve looked into products such as Gita [1], Spot and Unitree’s A1[2].

      Gita is impractical as it’s nothing more than a device which will fail once it will encounter a sidewalk’s curb or a small step.

      Boston Dynamics Spot is way out of my price range.

      A1… Not sure. It’s a Spot knock-off and probably not much more than a cheap development platform. Which is fine, but not a solution which would just work for me. Also I’m kind of suspicious about how much payload I’d be able to handle.

      What I basically want is a mule, the size and agility of a dog, which should be able to handle a payload of perhaps 10 kg and just some kind of ok battery life, let’s say 1-2 hours. It shall have a follow me function, and that’s pretty much it. (Am I asking for to much here?)

      Yes, I cod get me an assistant dog, but packing stuff on a dog just doesn’t feel right. Also I’m afraid I wouldn’t be able to give it the attention and care it needs.

      If someone were to provide such a device and be able to provide it for something in the range of perhaps $3000, even the most stubborn health insurance might cover part of the costs.

      That market is huge, but no manufacturer seems to take it serious.


      1. Seems like some sort of augmentation device, an exoskeleton, would better match your needs than a companion beast of burden. It would probably need less energy storage too, because the wearer could provide some.

        1. As the arthritis in my hands is here now (rather than Coming Soon) I would settle for something you can make now – and fairly cheap.
          How about a walking frame, with a crane attachment? Great for getting heavy shopping bags in and out of a trolley or car. One electric motor, a swing arm, and a few buttons. You could even put it on a mobility scooter.

    2. >In industrial manufacturing “cobots” are beginning to flourish

      Cobots are just hype right now. The same fundamental problems persist: the robot has such limited processing power and means of input of its surroundings that it’s essentially dumb, deaf, blind and all around senseless. It will put a screwdriver through your hand before it even recognizes that there’s something in the way because the way it recognizes the situation is by the increase in torque of one of its joints. Well, then it’s too late.

      There’s a threshold effect here, where the robot is almost useless until it reaches almost human levels of performance – and right now our AIs are worse than insects.

  2. VHS taking off was all about the price point, and that was based on licensing. We all knew that Betamax was superior, but it was owned by SONY, and SONY wanted to license the technology to all other vendors for a significant amount. JVC created VHS, and decided to make it royalty-free, so manufacturers didn’t have a licensing fee. JVC was willing to settle for not owning the whole pie to have a piece of the pie, SONY was not.

  3. The biggest problem with the “lease from the worker” idea is that it will be billed as socialism — and it is.
    But without that, we will see more and more of industry monopolized by the value of capital: If a company can leverage capital to buy robots, versus paying people, they will.
    I’m most concerned about automated vehicles, specifically trucks. The conversion or replacement of existing long-haul cabs is likely to be too expensive for individual owner/operators, and will almost undoubtedly go to large corporations. If you need to belong to a conglomerate to get your self-driving big rig fueled/recharged and serviced (a neo-truck stop), that’s another barrier for the individual o/o. What should be happening today is cooperatives forming that are owned by the operators, not by public shareholders, to enable those rigs to get serviced on the road, provide loans for upgrades/replacements, etc.

    1. The problem with things owned by workers/members (like mutual funds) is that they tend to all vote to sell out for the offer of short-term gain.
      Most people are too short sighted to invest for the future. See also credit card debt, etc.

    2. What is the point of cheaply manufacturing (or servicing) everything through humanless operations if most none will have an income to pay for it? You need a finite ratio of what-you-get to what-others-get to be above them. It is a sort of Ohm’s law – if current is not flowing, there is no voltage drop, no circuit, just isolated noncommunicating points.


    Can we please retire this line of idiocy?

    I drove to the lumberyard a few days ago.. about 40 miles each way because I live out in the sticks. I bought a pickup-load of plywood and 2×4, and drove home. In doing so I deprived maybe a dozen laborers of a few weeks work carrying the lumber cross-country by hand. Or maybe I shut down a livery with horses, wagons, feed, and so on. The EEEEEVUL internal combustion engine in my truck indisputably cost people jobs.

    Of course, the increased speed and reduced cost of carrying lumber has created far more construction jobs than the number of teamsters the existence of pickup trucks displaced, but let’s not let that interfere with an opportunity to flail our arms like an air-dancer at a used car lot.

    The economics of automation are well established: machines only replace people when they reduce costs and/or increase efficiency. Those savings make new kinds of business (and new kinds of jobs) cost-effective. And in general, the jobs that can be replaced by automation tend to be the lowest-skilled, lowest-paying, and least desirable ones in the market. The cost of products and servivces, and the migration of human labor, moves from simple and massively fungible things like material handling to human activities that can’t be automated, like decision-making.

    Jobs become more skilled, but not at ditch-digger to rocket-scientist scale. it’s more like ‘hand-loading ingots of pig iron onto rail cars’ to ‘forklift driver’. And when one forklift replaces a dozen hand-loaders, twenty companies can suddenly afford new operations that require a forklift driver.

    1. This is typically true for most of humanity, and if history is any indication of the future you’d be right. But we’re pouring a TON of resources into mixing that “decisions” part of the human element by letting computers infer a decision based off those historical decisions we like to call “context.” If the AI can infer the right decision 95% of the time, and the other 5% you can right off in the lawsuit, then we can get rid of a ton of jobs. Sure, some jobs will be augmented, but others will take on multiple jobs that used to be with different people. Soon, you’ll have two camps, people who use the AI as a coprocessor, and people who make the AI. In both cases, we’re looking at 20% of the world being able to comprehend these jobs, the other 80% you’ve got to figure out what to do with. And unlike before, where we had PLENTY of time to figure out what’s next, our world of expediency will move too fast for adaptation. The best we can hope for is the current supply chain issues we’re feeling can buy some time.

      What I can’t wait for is our interfacing computers like we see in tv and films today. My wife and I make fun of how far fetch sci-fi computer usage is but if we think about our computers running AIs all the time and our interactions are simply mechanism to control them, it doesn’t seem so far fetch. How did the computer know what graph to use to show the data? Well, the AI inferred the right visualization based on the data, opening the person to interpret the data vs loading a spreadsheet up, organizing the data in columns and rows, highlight the data, add a chart, adjust settings until it made sense.

    2. Spot on. People severely underestimate the amount of mostly high labour jobs that became obsolete due to technological advancments. There also seems to be a neverending stigma of automation completely removing humans out of the “service line” so to speak. I think it’s mostly a result of science fiction creating a false image of automation being either a solution to all our problems or the downfall of humanity as a whole.

      1. Makes you want to write a children’s story:

        Ten men want to build ten small sheds in their gardens, so they went to town for the materials. There they started thinking: “How do we get the wood from the lumber yard?” – so the men each bought themselves trucks to carry the wood, except one who bought a wheelbarrow. The ten men thought, “How do we cut the wood?”, so they each bought cordless circular saws, except one who bought a humble panel saw. The ten men pondered, “We need some way to nail the wood together”, so they each bought nailguns and compressors, except one who bought a hammer. Next the ten men needed to dig a trench for the foundation, so they rented diggers, except one who bought a shovel.

        The only thing left was to gather everything up and start working. To their surprise they found out they didn’t have any money left to buy wood at the lumber yard, or cement, or nails at the hardware store, because they had spent everything on their trucks, their power saws, their nailguns and compressors, and renting a digger for the weekend. Except one guy; he took his shovel, his hammer and his panel saw, couple bags of cement and some wood, strapped them to his wheelbarrow and went home to build his shed.

        Moral of the story: If you’re trying to save labor by spending money on technology, where do you get the money? Answer: labor. One way or another you need to do something useful to earn your keep, so you might as well skip the whole bunch of nonsense in the middle and just do the work. It’s easier that way.

    1. These already exist in factories and have for a long time. At one of my former jobs, an entire factory line ran off of these (replaced conveyers) and once they reached the end of the line with the complete product, they navigated their way to the warehouse, unloaded, and got back in line to start the journey over. Another factory in the same city has similar ones that move items around so well, they turned off the lights since the robots didn’t need them. Factories are getting hella complex nowadays.

    1. hmmm interesting. Nobody I know has a Roomba (or Roomba like robot). Perhaps we should broaden our worlds.

      Technically the era of home robots has began much earlier. When my grandmother was young, she needed to do the laundry, took her all day and was hard work. Later in life she got a robot to do this for her, it was called a washing machine, it really changed her life. What I’m trying to say is that it’s all a matter of perspective and the article of today pointed that out pretty well, people are expecting Rosie, a machine capable of doing much more then a single task.

  5. I can’t help seeing a glorified kitchen trolley doing silly tasks on sentimental music. The real thing that makes this possible is the preparation and all the additional things that allow the cart to load new stuff onto it.

    It’s horribly expensive compared to an ordinary trolley and if granny has the dryer next to the washing machine she doesn’t need to drive the basket all around the house. And if gradma need the full laundry capacity of a family of six then there are some other issues that need to be addressed first. A combined washer dryer in the right place can do miracles.

    But the “fun” comes when the cart get’s stuck behind the edges of the tapestry, sure it went great in the video… but in real life the dog will mess with the carpet, so eventually the carpet isn’t perfect any more and the “robot” will get stuck or crash. Suddenly grandma needs to call the robot-repair-person, who needs to be quick because the dog needs a walk to… but who’s going to walk the dog. And why do they want to walk the dog, but not help grandma (who is now depending on trolley).

    Who’s going to put the soup in the bowl… who’s cooking the soup… who’s going to get the soups ingredients. All lot’s of problems the cart can’t solve. So unless grandma still needs to walk around the house to do all these things, who’s gonna help grandma with all this. Eventually she needs to go to bed but can’t walk the stairs. A trolley that follows a programmed path is not going to solve this, a lift does, but the cart can’t reach her there…

    Grandma needs help that actually helps… interacts, so she doesn’t get lonely (the true danger of getting old). Making grandma an over-engineered glorified cart that has a higher failure rate (then a second hand 20,- trolley) isn’t going to make a big difference. It only prevents grandma from spending money on things she really needs or likes. Because she can only spend her money once since her pension (if she has any) is limited. Unless these aids make a real difference, they are only helping the business of the people who make this and the service organization that keeps on fixing it.

    Technology isn’t the answer, all grandma really needs is real care, attention, kiss and the KISS principle. And a thing that really helps is one of these:

    But… perhaps next century we can all have a Rosie (like the jetsons).

  6. Somebody needs to invent a driverless trash truck that picks up all the broken down robots that weren’t made to last more than a year. As I geek, I was OK with replacing two bumper sensors and a battery in my Roomba, but I think the average consumer would have punted it to the curb long ago.

  7. Both the robot tray, and the trays on the robot that’s delivering food at a restaurant are too low to the ground. I wouldn’t eat any food from the bottom tray. Really don’t like the next tray up either.

  8. “… for the elderly or people with limited mobility due to injury or illness, …” could better be rewritten as “… for people with limited mobility due to injury, illness or age, …”.
    Not all elderly are less mobile to the point it hinders their daily routines.

  9. We already tried that model of capitalization during the assembly line era and we called it “the stock market” and the pyramid of ability being what it is, despite everyone being able to participate in theory, very few people ended up owning everything, so its not a usable civilizational wide economic system.

  10. The robot server requires a crime-free 100% cooperative area to operate successfully.

    Anywhere else it’ll get stripped for food and copper and spray painted with graffiti until it looks like a railroad boxcar.

    Then when the robot is filthy and broken down the restaurant will go out of business and a new one that uses human servers will open.

    People underestimate how smart and strong servers are. They know to wash their hands, they know how to deal with drunks and angry people, etc. There’s a lot more to the job than being a human conveyor belt or human wheelbarrow.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.