January 11, 2005

• Interesting stuff. But this chap overlooks the ingenious ways which have been found to eliminate human labour from all sorts of processes without needing humanoid robots. He says that a store in 1950 would have been essentially the same as a store now, but that's quite untrue. In a 1950 grocer, you would have had to describe what you wanted to a man or woman behind a counter, who would have taken it off shelves, weighed it out, sliced it, and packaged it for you; and then it might well have been delivered by a boy on a bike. Now you just walk round a kind of warehouse helping yourself and all they have to do is scan the products and take the money - in some stores you even do that yourself. There's no role left for any robots. Similar changes have been made elsewhere in the economy; there's always an easier way to get rid of a given job than putting a robot in. I think he has an unconscious prejudice in favour of the assumption that robots will inevitably do jobs that humans regard as low-status, ie manual labour: but in fact robots are rubbish at real-world physical tasks - far better at providing sophisticated entertainment. Here, Mr Threepio, let me get that suitcase for you...
• One thing that strikes me in all of his comments is that if the work force is sitting around in some sort of welfare home, w/o money, then why would we need all these robots? Who's going to be buying stuff when it appears that all of us will be out of work? What good would they do?
• I highly, highly recommend you monkeys take a bit of time and read his story based on his ideas: Manna damnItKage: the story will answer your questions. With two possible outcomes. IM*very*HO, this is very good.
• Toyota to employ robots
• killThisKid, thanks for recommending the story... it's a great illustration of the different possibilities one can have using the same basic tools.
• Marshall Brain is a personal friend of mine. We are in the same fraternity and went to school together in the 80s. (I admit I think I've seen him only once since then, but we've corresponded a bit by email since howstuffworks.com hit it big.) And yes, Marshall Brain is his real name.
• Oh... and the article. In a way, the robotic age has been here for a while. Go to any manufacturing plant and you see dozens or even hundreds of jobs being done by machines that used to be done by people. Thus, maybe 20 people can monitor the machines doing the work of a hundred or more people. The high-tech bubble postponed noticing this phenomenon, but it's been increasing steadily since the 80s.
• Um...Doohickie, perhaps you are mixing the concept o of automation with that of robotic labour? You don't see humanoid robots standing in the same place where factory workers used to, but tools that are attached to moveable parts that do the same job. It's like saying self-checkout machines in libraries are robots. It is automating the processing of book borrowing, but I would hardly say it's a sign that robotic labour is taking over human labour.
• I think what I'm trying to say is Machines ‡ Robots and Automation ‡ Robots.
• It's fascinating stuff. If you accept Alnedra's point that by "robots" we're referring to computing machines that are not merely automatic, they're autonomous, then I think he's a little off in his predictions of where the replacement of humans will occur. Pilots, for example. Human pilots may be expensive, and may be increasingly reliant on automated mechanisms, and may be the source of many errors that cause accidents... but would you get on a plane flown by a robot? More importantly, would lots of other people? I'm not sure he's taken into account how human reactions to automation and autonomisation may yet have a profound economic effect. Although technically they can be replaced earliest, I reckon pilots would be amongst the last to go, for purely instinctive and emotive reasons The one thought that says no to that? You can't hold a gun to the head of a robot pilot. Which may be important. I also wonder if the assumption that robots will necessarily undercut the costs of minimum-wage menial duties is accurate? It seems to perhaps be an extrapolation too far of Moore's law. I suspect that it's equally likely that autonomous, mobile robots will never, ever be cheaper to create, supervise and run than humans. Humans are really cheap, you know. Rather, I think we'll see more of what Pleggers said about automisation not needing anything robotic - simple systems will suffice, and human behaviour will alter to accomodate the new system. I think autonomy in machines is expensive; in humans, it's second nature.
• All well and good, until...