I’m talking a robot in my home that can do my ironing, fold my clothes, get me a drink from the fridge, etc.
I look at the current state of technologies like robot mobility & dexterity, object recognition, machine learning, voice recognition etc… and it seems to me like personal robots must be very, very close.
How soon until I can buy one? Which requisite technologies, as of 2018, aren’t sufficiently matured to allow it?
Is there a reason to think they’re any closer than self-driving cars? We’ve got an ongoing thread in this forum about their prospects.
A good question to start off with here is, what are the reasons to think personal robots would become a reality before/after self-driving cars do?
I kook at those same things and I think 20 years or more.
I think the real question is when will personal robots be cheaper than hiring a maid to come in once a week. I’d have to have a lot of disposable income to buy a robot to fetch me my beer every night.
By robot, I assume you mean an autonomous, mobile device capable of handling a variety of tasks? Building a robot that could do any single household task is a tall order. Building one that can do all of them is well outside our current level of technology.
Have we even developed a power source that is quiet enough, portable enough, and has enough juice to run a hypothetical household robot?
I think we’re a long way off. If there’s enough demand, we’re more likely to see a washer/dryer that can fold your clothes for you, a kitchen that can assemble a meal for you, etc. Those things can plug into the power grid and don’t have to be compact or mobile. Their guts can live inside the house.
This. See, the way you should look at it is you should evaluate the engineering effort vs the value added.
SDCs : Large engineering effort, but the problem of driving at it’s core is fairly smooth. Solutions probably exist that will allow any common type of car (as in, same number of wheels, similar steering capabilities) on any road anywhere and at least not crash. And, most of the time, make it to the destination.
Household tasks are much more varied and complex.
The next thing is the model. Most people won’t own SDCs, they will come pick you up. This lets them cost 60-100k each, full of expensive sensors and electronics, but you don’t have to pay upfront.
Is the household robot going to stay at your house? How much are you going to pay for it’s services?
We will see household robots but they are farther down the reward vs difficulty curve for robots. It’ll be years after SDCs.
An electric powered home robot could easily plug itself in to a wall power socket to recharge its own battery.
Actually, this one problem isn’t too bad. The trick is to stop using hydraulics and use direct motor drive, with regen. This slashes the power required down severalfold. This is why the newer Boston Dynamics bots are so compact, quiet and don’t need engines.
But even then, yes, the robot would need to take recharge breaks every few hours. But we’ve managed to speed that up a bunch - both electric cars and your smartphone can recharge about 60% of capacity in 20 minutes, so obviously household robots would do the same trick.
If the robot’s labor were very valuable, it would swap batteries.
There is already a $1000 prototype that doesn’t work and a $16,000 prototype that doesn’t work.
The biggest impediment to self-driving cars is that they have to be able to deal with an uncontrolled range of situations and if they screw up people die.
Household robots could be ‘educated’ about the layout of your house, but would still need to be able to deal with -or step around- any number of unexpected things. However it seems relatively unlikely that a maidbot mishap will kill anybody, so from a bug-fixing and legal standpoint getting these things to market should be smoother.
Also it’s worth asking, are we talking about having one general-purpose robot, or a bunch of separate specialized robots (like washing machines that fold). The latter is much easier - the individual robots only need to know their own job, many won’t have to move around much, and generally won’t have to know about or understand things outside their wheelhouse.
In fact, I’ll go out on a limb and say we might see Roombas hit the market within the next five years.
These are some of the most difficult tasks for a robot. Which is why textile manufacturing (mostly sewing) is still done in third-world sweat shops rather than fully automated factories. And handling your dirty laundry is even more difficult than handling brand new rolls of fabric.
Even when we come up with robots that can iron and fold clothes reliably, it probably won’t be a general-purpose “personal robot” that walks over to the hamper, takes it to the washing machine, etc. I think it’ll be a fully automated laundry machine - you throw your clothes into the chute, and an hour later, clean, ironed, folded clothes come out. And instead of a robot that can get stuff from your fridge, you may have a robotic mini-fridge that comes to you when called.
I think expecting a single robot to perform a wide variety of household tasks is a pipe dream. However, automation that decreases the amount of manual labour around the house? Already here, and increasing all the time. And I’m not just talking about Roombas (though those obviously count). Just the other day I was speaking with my mother on the phone, and made some comment about having to do a couple loads of laundry but that wasn’t really much work, and she responded agreeing that these days laundry isn’t a lot of work - implying that this certainly wasn’t always the case. And thinking back to the wringer/washer she used in my childhood, she’s certainly right.
Don’t expect personal robots. Do expect specific household tasks to continue to become more and more automated, usually by single-purpose devices.
Handling textiles is particularly difficult, but in general, a robot to perform a wide variety of household tasks as instructed by an untrained human ‘operator’ would have to be able to understand complex, multivaried commands in English and other languages, which is itself a very difficult problem. And it isn’t as if there is either a great financial or social benefit, or a particular broad demand, for robotic household servants. I think assuming this could be available to the general public in two decades is optimistic unless you restrict the desired functionality to some very limited subset of tasks and capabilities.
If there were robot technology capable of this degree of flexible operation at a comodity price point, the primary application for it would be in health and elder care where we will have a growing need and insufficient labor force unless we relax immigration laws in the developed world.
Stranger
I’d guess it’ll be the 2040s or 2050s before bipedal robots that are cheap and effective enough for domestic work are common.
I base that on nothing.
What single use robots will be coming though? We already have washers/dryers, we’ve got dishwashers, vacuum cleaners, etc.
I’m not sure what single use robots could come about. Major household tasks that still haven’t been automated include:
Grocery shopping
Putting items away when you get home
Cooking
Vacuuming
Scrubbing (floors, tubs, toilets, kitchen countertops, etc)
Taking out the trash
Folding clothes and putting them away
Some of this stuff isn’t so much being automated as streamlined. Grocery shopping can be done online, then you pick it up. Or you have it delivered.
Robot vacuums, for now, aren’t that great. the time it takes to move everything out of their way is longer than the time they save.
A cooking robot prototype exists, but it is still expensive. However I’m guessing it’ll be under a few grand in a decade or so.
A single robot which does all of those things would be idiotically complicated and expensive to build. Do you expect your washing machine to fix you a martini? Do you expect your microwave to iron your shirts?
We haven’t reached the limit of home automation. Nowhere near. But we’re also not moving in the direction of Rosie from The Jetsons, for multiple reasons, the big one being that it’s a lot more bang-for-your-buck to make a Roomba, a relatively simple and stupid machine which does one well-defined task reasonably competently. That’s the direction practical robotics has seen the greatest success, and that’s the direction it’s going to continue to pursue.
Smart cabinets and refrigerators which use RFID chips on food packaging to keep inventory and get your phone to alert you when you’re out of peanut butter make a lot more sense from every possible perspective than trying to solve computer vision so we can make a robot which fold shirts and dusts bookshelves.
For a pither take, see this comic. Be sure to hover your mouse on the image and read the text.
[quote=“Wesley_Clark, post:15, topic:812349”]
What single use robots will be coming though? We already have washers/dryers, we’ve got dishwashers, vacuum cleaners, etc.
I’m not sure what single use robots could come about. Major household tasks that still haven’t been automated include:
Grocery shopping
Putting items away when you get home
Cooking
Vacuuming
Scrubbing (floors, tubs, toilets, kitchen countertops, etc)
Taking out the trash
Folding clothes and putting them away
Some of this stuff isn’t so much being automated as streamlined. Grocery shopping can be done online, then you pick it up. Or you have it delivered.
Robot vacuums, for now, aren’t that great. the time it takes to move everything out of their way is longer than the time they save.
A cooking robot prototype exists, but it is still expensive. However I’m guessing it’ll be under a few grand in a decade or so.
[/QUOTE]Regarding grocery shopping and cooking, we’re going to see human-facilitated services (like takeout) slowing down the pressure to have robots doing the same thing.
I would guess it’s at least 20 years out for a general purpose robot able to do a variety of tasks. For uni-taskers there are already a few out there, but for what you are asking it’s going to be a while. The most advanced robots I’ve seen demonstrated still cost a lot and I don’t see any way to bring their price point down…and most of them are still not capable of putting it all together into one package, so to speak. IOW, they are good demonstration platforms for a vertical set of tasks, but they don’t do everything. I’ve seen robots that are getting pretty good at, say, walking or picking something up or movement over rough terrain, robots that are fairly good at speech recognition or interaction, and robots that incorporate machine learning or at least weak AI…but not one that does it all and certainly not at a price point that the average person could afford (I’d go with the $1000 dollar magic number). And I don’t see anything indicating that someone is going to put all of it together into one package at any price point any time soon, let alone getting it down to something the average consumer could afford and would want to shell out the money for.
I think it will be a very long time. Yes you could make specialized machines to somewhat clumsily carry out mundane tasks as long as nothing significant changes in their work area. But put a red shirt in a load of white laundry or substitute salt for sugar and they will carry on like nothing is wrong.
Computers do exactly what they are programmed to do and nothing more. They are just a bunch of algorithms. In the past things like printers or tape drives could make really spectacular messes when something went wrong and they were not programmed to handle it.
We really won’t have machines that can adapt and learn effectively until we have AI. And once we have functional AI that brings in another set of issues, like once a machine can think will it become self-aware and decide it doesn’t want to clean your toilet?
its been theorized that once we do get AI its going to run along the planet of the apes route … first thell be pets companions (for a literary version of this read jeph Jacques questionable content web comic) ect then become more advanced and then be programmed/trained to do more and more and eventually become part of society or overthrow us …