What Undercover Boss and The Jetsons Tell Us About the Future of Jobs
slate.com
slate.com
"Genuine Food - touched by humans". "Authentic Underwear - touched by humans".
Robots don't spit in coffee. Robots don't under-cook your burger when they have an ambition crisis. Robots sew neat seams.
All you need is software that determines what needs to be done, and a minimum wage work force wearing headsets to receive task commands to carry out. Short circuits the need for ANY advancements in AI to enslave entire social classes.
I would love this to be wrong.
Doesn't sound like it's a huge problem to me.
Having a 3 segment robot arm pick up a salt shaker requires balancing 7 codependent feedback loops, while constantly checking if you're hitting anything and requires constant evaluation of forward and inverse kinematics. As if that problem is not bad enough, you have to do it not just fast, but in real-time (as in, you have to calculate all this at ~50hz at least. If a single decision takes 100 msec, you've just damaged or lost the robot). State of the art in robotics is picking up and unboiled egg 10 times without breaking it, and most robots can't do it. If you start to use those robots to hand people burgers, expect broken bones at least, casualties at worst.
Replacing a burger flipper requires solving not one, but actually most unsolved AI problems. You need voice recognition + understanding. Okay that's one. Then you need to understand the meaning in what was actually said. You know, the problem that google's 35000 employees from amongst the "world's best" don't quite know how to solve. I mean, they're far ahead of the pack, but nowhere near open-ended conversations with a human, never mind doing so with social grace. If your robot can walk around you need to manage an unsolved 30-dimensional feedback loop problem. State of the art for that problem is pathetic as well[1][2]. If it can't walk around you have other problems, and it's still at least a 14-dimensional feedback problem (two six-dimensional arms + 2 grippers). There's the mechanical problem of creating a robot that stays robust when working with "gooey" stuff like food (it will build up inside the robot, then fail or cause overheating). Then you need to solve all the problems that the "manna" system solves, or there's no point having a robot in the first place.
[1] http://www.youtube.com/watch?v=vjgkV77gVcs (note the obviously reduced dimensionality that wouldn't fly in the real world, and that the robot sees no problem moving through itself. It's stability under disturbances is also not at the point that it would keep operating for long periods. Do that in real life and the thing will topple over) [2] And before you say Bigdog, note that 4 feet locomotion is a much simpler problem. Bigdog is solving for 8 dimensions, and that's state of the art (and it is not sufficiently reliable to operate a single day without assistance). 2 feet locomotion in a human is ~140-dimensional, the minimum you can reduce it to while still having a nice mobile robot is 12-dimensional.
But the thing that I found interesting, both in the story and in real life, is that the story's author has his characters as 'worker bees' with not thought to living their lives, and then switching over to living full lives. If you subscribe to that philosophy, then nobody would ever be 'useless' for more than about 8 years since you could pretty much learn everything from high school through a masters degree in some subject in that time if you dedicated your time to it, and assuming like the story author did, that these people didn't have to do anything to maintain their lives (food, shelter, clothing).
The difference between people who are naturally curious, and the people who aren't, is where I am most intrigued. When do people switch over to being victims? And what keeps them in that mindset?
Thanks for the link, it was an interesting read.
The more stringent accounting standards that were created over the last 20 years would have led to the demand for thousands upon thousands of white collar workers, if we still kept books and records the way we did in 1910.
On the other hand, companies would not have been able to grow large enough to need new accounting standards in the first place, if not for the productivity increases since the introduction of the computer.
I think it's more like computers allows us to do the same stuff faster and at a broader scale, but it has had little impact so far in which stuff do we choose to do.
I'll actually wait slightly longer for the automated checkout at the market, mainly because I want to see it in more places.
However, I'm not quite sure that will be enough for shops to automate. The problem for shop is that an automated system most likely cause of failure is to miss items. That added to the likely need to remove some security features to simplify automation and the already existing shoplifting problem means shop owner will be more interested in more clever (read more obnoxious) self checkout system than automated ones.
The reason given is to be more "customer friendly", which seems to correspond to the upscale neighborhoods where the lanes are being removed and, not surprisingly, more upscale grocers are starting to move in (Roundy's/Mariano's, Whole Foods, Fresh Market).
Turns out it takes a lot of staff intervention to keep the self-checkout lines moving.
I remember seeing a commercial for something one time where somebody just pushed a cart full of groceries through a portal-thing, and it automatically found all of the items and their prices, by RFID or something, found his credit card in his wallet the same way, and auto-charged it for all of the groceries. If they had that, I'd do it.
I think that we will see a period in the coming years that could be regarded as a "growing pain" for the human species. Jobs and professions could shift to be more inline with rapidly advancing technologies. During this time there will probably be people who find themselves either out of work, or unqualified. After this time I think that jobs and professions will again be available and that many people will be qualified for them. Perhaps even more jobs.
Who knows, maybe I'm an optimist.
Like what?
The problem is that we're creeping up the capability ladder, and it will only be a few decades until the robots can do the math, science, design, engineering, and assembly of new robots, which can in turn produce any other objects.
What's there for humans to do?
The real problem is that robotics makes humans unnecessary for economies to function, and we'll need to change our social model to account for that.
Or just let a huge swathe of the people starve.
Actually, I would see it the other way round, on one condition. If people had a guaranteed income (or housing, healthcare & food) then paying humans to do work that robots can do would seem like an unnecessary luxury to me.
Your so called "efficient farming" does increase the yield of calories per acre at the cost of killing of the microscopic ecosystems in the topsoil and bankrupting farmers. The only thing those practices are good at is to increase the margins of fertilizer and pesticide producers.
What bankrupt the farmers is that industrial farming does increase the yield of a given field at the cost of relying in an ever increasing and more advanced industrial inputs. This has the double effect of increasing both the fixed (machinery) and variable (GMO seeds, pesticides, fertilizers) costs, while at the same time increasing the offer and saturating demand. The farmers find themselves in the predicament to ever increase production just to survive on an ever decreasing margins. And when they can no longer afford the fancy inputs, the soil is already dead so they end up the proud owner of a wasteland and go belly up.
It would be hard to come up with sources for all of these assertions, specially in an informal debate, but for starters look at: http://www.joshuabradley.org/green_revolution/index.html
I am not saying that organic food is pointless - heck, I buy organic onions because they taste better and I buy organic milk and chicken meat because I vaguely care about how these animals are kept. But I do not subscribe to an idea that everyone absolutely has to eat organic food. It is not only non-essential, but I think it might actually be impossible.
Edit: Found the research: http://www.bbc.co.uk/news/health-19465692
30% less likely to contain pesticides is a nice positive benefit.
The other way to look at this, of course, is:
47 percent of our labor force will be freed from their menial tasks.
Our societies will collectively be able to produce and accomplish almost double what we currently can.
You'll never convince me that Luddism is a progressive policy.
Europe and the US have been historically bad at that. Europe hides those workers in "pre-pension" (loads of steel/mine worker get their state pension around 45), "long term unemployment"; US hides them in various disability assistance program. (source: there was an article here I think a few months back)
Luddism want basically to stop progress in order to keep people employed, which was stupid 100 years ago, and is stupid now. GP on the other hand suggest to embrace the changes but accept some/lot people will need assistance.
... or the "handmade" of the future?
The only way to have a completely automated restaurant is have a big box that spits out meals. The alternative is no automation or a lot of little boxes that are used for pieces.
The thing is that a restaurant would still needs counter-help and clean-up help so further. Let's assume the box reduces the cost of prep-labor by 50%. That reduces the labor past of your costs by 25% and your total costs by 12.5% assuming labor is 50% of costs. So it's a good savings but not an astronomical savings. But the counter-weight you have sell people on buying the food that comes out of the box. A few places compete on price alone but that's hard. Just consider that restaurants are already competing with microwave dinners in the frozen section that quite possibly are already spat out of a better, cleverer box at a central factory. It would seem restaurants compete on ambiance and human relations and a 12.5% prices savings might not compensate.
Another thing about "food out of a box" restaurant is that the box might not have the flexibility of a human preparer. If you need to change your menu, you might find you have to do expensive retooling on your box rather than a quick retraining of your employees. If you have to close a given restaurant, you may find no one wants your box. You may find yourself more dependent on your small number of box-technicians than you were previously on a large number of easily replaced unskilled cooks and so-forth.
The thing is that humans are a fabulous deal once you remove the price differential. Human comes with amazing skills, often train themselves with more, supervise and trouble-shoot each other, require little capital investment and manufacture themselves. This kind of thing is a good deal for nearly everything except production at a large scale (and the most important items, cars and iphones, are going to that large scale). The only unfortunate thing from the buyers perspective is that humans have a nasty habit of expect a share of the output instead of just a bare wage. But the solution to that problem is visible all around us ... humans.
That's true today, but don't you think artificial intelligence will change that?
AI has made a lot of strides doing what it's good at and basically stagnated outside that realm (it's been forty years since the first seemingly intelligent system appeared and none of these systems has become good at dealing with an uncontrolled environment). AI hasn't gotten noticeably better at dealing with complex and confusing environments.
Why do you assume that those tasks won't be automated? To some extent they already are in some places (kiosk ordering at some fast food places, cleaning devices à la Roomba).
Roomba's are really slow. There's no robot that can quickly find and clean a spill while keep people around the area safe. That kind of progress hasn't been made with robots and AI.
You can have something like a system where the restaurant is a big box that cleaned automatically, yes, but then it also becomes unpleasant. Kiosk ordering might work but then cashier is also the "public face" of restaurant.
It all comes down to "why do people go to restaurants instead of microwaving Trader Joe's stuff", right?
http://www.popsci.com/gadgets/article/2010-04/students-inven...
Notice that it is created to manufacture Chinese food. A change in taste to Korean might obsolete it.
Prototypes can do lots of things. They tend to have durability and reliability problems.
Plus, many chefs I know also have durability and reliabilty problems.
Try baked Italian dishes. Clearly there's a limit to the thing's flexibility. Not all Chinese dishes was conventional wok cooking for example.
Also, most humans aren't reliable but generally, you "rent" them rather than buying them - your cook get truly unreliable, you can find another cook. You would almost certainly have to buy a specialize Chinese food (or Wok-cooking) machine and so you'd be providing not just "health insurance" but "health care" for it if it broke down.
[1] http://www.marketresearchworld.net/index.php?option=com_cont...
The point isn't that automated food service is impossible. It's certainly possible. Indeed, it's been around for a long time [1]. The point is that it isn't necessarily a better investment than human-run food service and thus human-run food service may stay around a long time.