7 ms·
This article has issues. XKCD explains it best here: http://xkcd.com/1425/ http://xkcd.com/1425/ Take for example... "Other low-wage and labor intensive jobs i
by dragontamer 12y ago
This article has issues. XKCD explains it best here: http://xkcd.com/1425/ http://xkcd.com/1425/
Take for example... "Other low-wage and labor intensive jobs in retail, such as stocking the shelves of supermarkets with food, will soon be replaced by machines that can do those jobs better and faster than humans could."
Here's a state-of-the-art robot folding some towels.
https://www.youtube.com/watch?v=gy5g33S0Gzo https://www.youtube.com/watch?v=gy5g33S0Gzo
The amount of image processing and calculation needed to fold towels or stock shelves is in fact well beyond the reach of current AI / technology.
Yes yes yes, there's that shirt folding machine prototype going around. But that only gets the job done with a good deal of human assistance. (Placing the shirt on the machine correctly, assuming shirts are a certain shape, etc. etc.). A "Shelf-stocking" robot sounds like a horribly complicated AI task, the likes this world has never seen before. A research team might be able to accomplish it, but such a thing is clearly beyond the reach of typical tech within the next 10 years.
- melling 12y agoThat's why it's referred to as "The Coming Manufacturing Revolution". Technology is advancing at a rapid pace. If the problem is already solved but it's works too slowly then it's probably close to production ready. We aren't too far off from robots replacing millions of workers who make our smartphones, for example. http://www.extremetech.com/electronics/195556-foxconns-robotic-workforce-isnt-precise-enough-to-assemble-iphones http://www.extremetech.com/electronics/195556-foxconns-robot...
- dragontamer 12y agoThe pace of AI advancements is slower than you'd expect. The majority of Siri / Cortana / Voice Controls were figured out in the 80s, but it took almost 30 years before AI advanced to the point where speech recognition was usable. As far as replacing workers who made smartphones... that was _already_ done in the 90s. The Sony Walkman was an entirely automated build IIRC. We went _backwards_ from that era of mass production, because Chinese workers were cheaper than machines. The key is design. Designing components that are easily grasped by robots. Taking advantage of chip shooters (https://www.youtube.com/watch?v=SRu02F6AOmg https://www.youtube.com/watch?v=SRu02F6AOmg), or other technologies. We have automatic pick-and-place machines with electronics. We have automated robots that can place screws into predetermined holes. We have assembly lines that can be designed with "press fit" plastics. But these are old innovations of decades past. When your chips are standardized sizes and are placed always at the same x / y coordinate on a board... and when reflow soldering allows a bit of error... things are pretty good for machines. In contrast, I have never seen a robot stock a shelf. You've got objects of differing sizes, non-uniform visual stimuli, changing conditions and little human assistance. Again, you're falling into the trap of not knowing what has been done before (mass production has been done since the 80s. Full lights-out factories have been made through the 90s and early 2000s. This is _not_ a new concept), vs AI problems that will take a specialized team decades to solve. (IE: Stocking a shelf).
- melling 12y agoI guess you didn't read the article that I posted because it clearly states that the robots can't do the work to the required precision. Yes, robots do a lot of work in factories but they still need further refinement before they can do some jobs that require a lot of dexterity. And yes, we've had automated factories for a long time. Steve Jobs had a highly automated factory in 1990 to build the NeXT. http://www.nytimes.com/1990/12/24/business/all-next-inc-s-plant-lacks-is-orders.html http://www.nytimes.com/1990/12/24/business/all-next-inc-s-pl...
- dragontamer 12y ago''' I guess you didn't read the article that I posted because it clearly states that the robots can't do the work to the required precision. Yes, robots do a lot of work in factories but they still need further refinement before they can do some jobs that require a lot of dexterity. ''' You miss my point. You can make complicated electronics entirely automated if the design were better.
- cryoshon 12y agoAh, but that video is four years old. Here's the shirt-folder: http://youtu.be/Mr7U9pQtwq8 http://youtu.be/Mr7U9pQtwq8 Sure, it sucks at folding shirts quickly, but it's most of the way there. Reducing the amount of human assistance needed is going to be a gradual process for the more complicated tasks, but you can bet it's being worked on. The core of the technology is already operational, it's just not quite consumer-ready yet.
- hiou 12y agoBut how much more expensive is that robot to purchase and maintain as opposed to a low wage human worker? Remember, you only lease human labor so when one breaks it doesn't cost you anything. Also, low skill human labor is much more flexible than an automated robot. You can have them folding shirts one minute, to answering the phone the next, to rounding up shopping carts or trying to harass some kids you think my be trying to shoplift. If you don't need them as often you can just cut their hours. With a robot you are still paying for it. And the robot won't transport itself for free off the premises when you want to suspend your lease. Decision making will be where AI is going to have it's largest immediate impact. Not physical labor. Because physical labor is much more expensive to replace than mental labor decision making.
- Houshalter 12y agoComputer vision is advancing exponentially. This year the imageNet machine vision challenge winner got 6.7% top-5 classification error. 2013 was 11%, 2012 was 15%. And since there is an upper limit of 0% and each percentage point is exponentially harder than the last, this is actually significantly better than it seems. That's about as good as humans are estimated to be at this task, and humans require much more time to search through reference images. The future of robotics is not manually programming robots with every move they should make, using handcrafted features to do vision,like your towel robot. It's letting the robots learn how to do tasks by themselves.
- dragontamer 12y ago''' The future of robotics is not manually programming robots with every move they should make, using handcrafted features to do vision,like your towel robot. It's letting the robots learn how to do tasks by themselves. ''' That's still programming. If you've ever touched "Logic Programming Languages" like Prologue, you learn that you leave an enormous amount of reasoning to machines today. For the more mainstream programmer: it is a brutal amount of AI that is used to calculate "auto" specifications and template parameters in C++. But in either case, the AI may be significantly faster or accurate today... but you still require a programmer that understands how to use the AI and give it direction. True, not every move needs to be programmed anymore. But neither are we programming in Assembly language. We've got huge advances in compiler technology (which really are stemmed from the automated reasoning groups in AI). At the end of the day however, the AI's purpose is defined by the programmer.
- Houshalter 12y agoYes of course the thing will have to be programmed to some degree. The point is that it's within the realm of possibility to use AI to do computer vision and robot control, rather than having some programmer trying to figure out every possible edge case of a messy real world problem. Instead the programmer designs the reward function and collects data and sets the hyper parameters, then the AI figures out how to solve the hard parts of the task itself.