The slowdown of the productity in the society can be dated back to the 1970s. In that decade electronic machines were invented which improved the productivity. Most people in the 1970s believed, that in the coming decades many more innovations will take place. The vision of accelerated innovation was described under the term Fifth computer generation. It was a vision about modern computers which are better than what was available in the 1970s.
In the 1970s, electronic devices like a barcode reader and the first computer software for example the C compiler was available. That was a nice sort of technology, but it wasn't enough to automate factories. The imagined technology was a robot which is capable of replacing humans. The plot of robotics automation and thinking machines was described under the term Fifth computer generation and the first real projects were introduced in the mid 1980s. That was the upraising of the LISP workstation and the first biped robots were developed.
It's important to know, that the Fifth computer generation never was realized. It remains only a vision but was not available in reality. Somebody might argue, that in the year 2019 factory robots are available, expert systems are used by the doctor, and neural networks are able to detect the content of images. Are they are really doing so? No they don't. According to the vision of the fifth computer revolution, robots are introduced in the economy which increases the productivity, and expert systems are replacing human doctors. The problem is, that such kind of technology was tested but the engineers struggled. A robot like “baxter robot” was produced by Rethink robotics in the 2010 and it was used in some factories, but the robot didn't increased the productivity. The same is obvious for medical expert systems. Some of these advanced software was published as a prototype, but in a project there is no obvious advantage over not using such software.
The only technology which can used in reality was invented in the 1970s. Normal electronics innovations like the C programming language, a barcode reader or a pocket calculator provide a clear improvement over not using such technology. The only problem is, that no further innovation is available which goes beyond such simple devices.
The next logical step after a barcode reader would be a robotic check out desk in the supermarket, and the next step after an electric car would be a self-driving car. Unfurtunately, such technology wasn't invented. The existing prototypes which can be seen in youtube videos are only working in the laboratory but not in the reality. Some attempt were made to sell robotics and advanced artificial Intelligence to the mainstream audience, but the customer doesn't like the products. They have tested out the robot and they come to the conclusion, that the costs are too high and the advantage is zero.
What we can say for sure is, that the promise of the fifth computer generation wasn't fulfilled. Advanced technologies like humanoid robots, neural networks and fifth generation programming language are available from a technical perspective. That means, anybody can download an expert system from github or program such software from scratch. But in contrast to the outlook of fifth generation computing, such technology can't be utilized for solving real problems. It's a useless piece of technology which fulfills it's own needs.
To understand the problem we have to go back to the 1970s. Technology from that time provided two aspects at the same time. It can be used in reality, and the technology itself can be teached. The inner working of a pocket calculator is explained at the university, and the product can be sold to the customer. This is no longer the case for robot technology. THe inner working of robots is explained in university courses, for example with the help of Lego Mindstorms. But it's impossible to use this technology in reality. That means, Lego Mindstorms is an educational tool and it won't increase the productivity in a factory. This problem is described in the literature as a productivity paradoxon. It means, that on the one hand, the engineers have developed advanced supercomputers, and powerful robotics algorithm and at the same time, it's not possible to do something useful with these machines.
For the reason of simplify the overall explaination, it make sense to assume that the fifth computer generation never took place. It is simply not available and the vision can't be realized. The only thing what is available is the vision itself. It's a description of a future, in which robots are working in the factories, in which the productivity has become better and innovation is used for improve daily life.
This kind of optimism was only realistic since the end of the 1970s. With the advent of the next revolution, namely the fifth computer revolution the productivity has stopped. The new innovations are there, but they can't be utilized in reality.
From a naive perspective the task is easy. All what the engineers have to do is to realize the vision of fifth generation computing in reality. They have to build robots and expert systems which can increase the productive further. So that innovation never comes to an end. And exactly in this point, the engineers have struggled. They are not able or not motivated to realize the vision of fifth computer generation. They have build some prototypes, but they aren't working. Other engineers have tried to fix it and they failed too. I'd like to give a small example.
In the 1980s, Joseph Engelberger invented the Helpmate robot. Suppose, that Engelberger and his team wasn't the best technican. Are today's AI experts are able to invent a robot which is more powerful and can be used in reality? No they don't. Even if they are using the latest hardware and software, they will struggle again to build a simple hospital transport vehicle. They can build a prototype, but the consumer won't like the device very much. It won't increase the productivity and after a while, the project will be canceled. This kind of failed robotics project is visible everywhere. No matter, which engineer in which decade is trying to realize the vision of the fifth computer generation he will fail. An extra problem is, that most computer students are not aware of the problems. They believe, that their knowledge is high enough to build a robot, because the Lego Mindstorms prototype is able to follow a line on the ground. But there are two different kind of robots available, the first category are university robot projects, in which the challenge was optimized so that a robot can fulfill it. The second kind of requirement is the reality, in which a robot should increase the productivity and help existing employees.
On the first look, the technology in the year 2019 is much better than what was available 40 years ago. For example, modern youtube is much more powerful than an old black white television screen. A closer look into the technology will shown, that apart from entertainment industry, the computer didn't have improved the society so much. The amount of human work until a car is produced or until a book was written is the same like 40 years ago. The measurable productivity in the economy is the same or it is a bit smaller than in the past. The reason for the productivity paradoxon is the missing fifth computer generation. Robots and Artificial Intelligence struggles in solving tasks from the reality. Instead the domain has build it's own domain which has nothing to do with the needs of society.
Fifth generation computers
The first books about fifth generation computers were written in the early 1980s. The hope was that a new computer generation will increase the productivity in the service sector. That means, at the end of the 1970s it was clear, that society is working with a low productivity and there is need for advanced technology which helps to overcome the issue. As key technology it was imagined, that Artificial Intelligence and robotics will help to deal with the problem. It was assumed, that neural networks, expert systems, graphical workstations, Prolog like languages and fast computer networks will be able to increase the productivity of the society.
The funny thing is, that from a technical perspective all the wishes become true. An advanced graphical workstation which can execute LISP programs in realtime and has access to the Internet is sold to a mainstream audience. And the latest generation of Intel processors is much faster than it was expected for fifth generation computers. On the other hand the requirement of increasing the productivity at the workplace wasn't reached. The problem is, that the productivity is exactly the same like in the mid 1970's, it didn't even increased by a simple one percent.
No comments:
Post a Comment