] The next step was autonomous, humanoid robots. The
] mechanics of walking were not simple, but Honda had
] proven that those problems could be solved with the
] creation of its ASIMO robot at the turn of the century.
] Sony and other manufacturers followed Honda's lead. Over
] the course of two decades, engineers refined this
] hardware and the software controlling it to the point
] where they could create humanoid bodyforms with the grace
] and precision of a ballerina or the mass and sheer
] strength of the Incredible Hulk.

Iclough's memeings about robots made me go back and look at this essay.

This was making the rounds a few weeks ago. I ignored it because he started off talking about McDonald's Kiosks. We had those in Murfreesboro in the early 90's. Hasn't gone anywhere. Not exciting...

However, a fully automated McDonalds is an easy thing to imagine. We (finally) can buy robotic lawn mowers. The primary problem with these things is battery power. As fuel cells become a reality robotic lawn mowers and vacuum cleaners will become common place in American homes. People will have their weekends free.

What is a combine but a big lawn mower? Robotic combines are already running at Texas A&M. The food gets automatically harvested.

Transportation? They have a Humvee at Georgia Tech that drives itself. This can be aided with the right kind of infrastructure in the roads. When the time comes, we'll build it. The food will move from the farm to the plant automatically. Its already processed automatically. Then it will move automatically to the store, where a kisok will take your order. Cooking it? Easy I think. I'm actually somewhat amazed that people still cook fries. The mechanical motion is so simple. They really only need a person there to make sure that nothing goes wrong... And we're off...

I agree with the author that we will see a lot of robotics in our lives in the next few decades. It will be an important growth industry. I also agree that honda's humanoid robot is a tremendous achievement.

However, this article gets several things wrong.

A leap of logic is made between humanoid robots and robots doing human jobs. Many human jobs are a lot more complex then we think. Cleaning a room is incredibly complex, when you consider how varied the environments can be, and the need to ID so many different kinds of objects, and sort them properly. This is a tremendous knowledge problem. I will admit that more computing power could help you tacle it, but its just an example. The point is that human's easily handle situations that are very complex to program because humans aren't computers. More on that below.

This article assumes that Moore's law will continue unabated. This is a bad assumption. Moore's law has a shelf life that will end in a decade or two. You can't get smaller then atoms, and no engineering breakthrough is going to fix that. New science maybe, but you can't predict scientific progress in the same way. Science is about a better understanding of how things work, and how things work doesn't lend itself to linear progression of human capability, because its not driven by human ingenuity, its driven by cold, hard reality. Reality simply is what it is.

This article makes the famous AI deceit that computing power will someday equal that of the human mind. 1. The human mind is NOT a computer. In fact, humans are extremely BAD at computing. Thats why we all find calculators so helpful. 2. My $10 calculator has far superior calculating capabilities then my human mind, so if you are looking for dominance from computers, you need look no further. Your Pentium 4 is just a big calculator. Thats all it is. It adds, multiplies, and shifts bits. Thats it. Its not magic and its not the same as thinking. Neural Networks are more interesting, but frankly what we don't know about how the human mind really works could take several generations to explain, and almost certainly will.

This article presumes that the technical advances of this century are evidence that any imaginable technical advance will certainly happen in the next century. Hogwash. People have been predicting time travel and cyborgs for over 100 years, and yet we have neither. No one, on the other hand, really predicted the Internet. Not way way back. Closest was Bush's paper on microfiche.

If you know how to do X, but you don't have enough Y, and Y is getting cheaper at N rate, you can accurately predict that you will be able to do X at a certain time.

If you don't know how to do X, you can't predict when you'll be able to do it.

Robotics is somewhere in the middle of these examples. I think we're going to see a lot of exciting advances in this area, that it will eliminate a lot of jobs, and that this is all good. However, I don't think I'm going to be writing long winded essays at some point in the near future about the civil liberties of artificial life forms. If I do, it won't be because of something we are able to accurately predict right now.

Having said all of this, the point this guy makes about the economy is absolutely right... Jobs are going to go away. This has already started. It is going to continue. Just as in the industrial revolution the jobs moved from the farm to the factory, in the information revolution the jobs are moving from the factory to the mind, and those intellectual jobs are being outsourced to India.

Consider the impact that technology is having on the music industry... We know that the whole way that media is produced and distributed is going to completely change, and its going to look like linux; a place where no one really makes money. Not really. Not for the work that REALLY matters.

Its going to be terrible terrible. No one knows how to build an economy in an environment where there is diminishing scarcity. People are simply going to give their work product away for free, and they are going to compete and incent with reputation. This change will hit the whole of society like a ton of bricks, and a lot of people are going to fall through the cracks. Trying to figure out how to make this transition happen smoothley is something worth thinking about, but it is so far off at this point. I think we are really talking about problems that will primarily be born by another generation. We're just at the very start of this now.

Robotic Nation, by Marshall Brain