Comment Common mistakes about Robotics & Economy (Score 1) 1457
Hi,
This summurize my answers to the most common mistakes, according to me, that I can read on the topic of robotics and its consequences on economy, especially the "end of work":
* "When the first robots will take the place of human workers, new jobs will naturally appear to build, repear, create those robots."
WRONG: AI-based robots can simply do, by definition, any kind of jobs, including the new jobs they will generate. You can build, repear and even design robots using AI-based (intelligent) robots. The only area they will probably not compete with humans is art and social relationships.
* "Using taxes, we can make the cost of robots use higher than the cost of human work force. This solves the problem using a political/economical trick."
WRONG: The quality and the complexity of the tasks robots can perform can put man out of the game in a decisive way. Many high tech products today could simply never be done by humans, even if they where very competitive no-cost slaves. They simply don't have the skills robots have.
* "We have seen this before many many times: workers in the coton fields where replaced by machines, workers in factories where also replaced by machines. The services sector offered to these people new jobs and it will still do it. Nothing to be worried about."
WRONG: It's a new situation. The mechanization mentionned here is related to automatization of physical work force. Here, we are talking about intellectual force, which includes the kind of jobs done in the services industry.
* "By 20xx the computational power of computer will be equal to the one of the brain. This will be the beginning of the AI era"
WRONG: Two mistakes. First, the computational power of computers (as they are designed now) is not to be compared to the one the brain. It's simply different hardwares and it doesn't make sense to compare them (neurons transistors). Second, it is very naive to assume that hardware is the only thing. There is also something called software! Even if we have the computational power necessary for AI, we still have to design the software for AI, which is a very challenging task. Moreover, there is no clue today on how to design this software. It is typically a matter on which there could be no prediction. We could find the idea(s) that leads to AI within 10, 100 or 1000 years. Nobody knows.
* "AI is making progresses, just look at the chess programs or expert systems".
WRONG: Theses domains (chess, reasoning, database management,...) are just logicaly driven tasks than can be automatized easily. The real challenge of AI is today into robotics & vision: how could we design a machine that can find his way in the streets, take a bus to go to a previously unknown place, summerize a book, etc... The most difficult tasks seem to be the one we, humans, do the most easily. On the contrary, the easy tasks (chess, maths) are the one we are the worst at. AI is in the domain of the difficult tasks. And we have made almost no progresses there so far.
* "Intelligent machines will soon fight us and kill us all" (the Terminator fear).
WRONG: It is most likely that all these aggressive tendencies are specific to our still animal-like specie, inherited from a time where life was all the time endangered by hostile predators. If the machines feel threatened by humans, it will probably be more economical for them to leave the planet and move to space, using solar energy as a power source and taking the control of other planets for the material needs. Why risking a fight? Anyway, they would probably leave the planet, even if they don't feel threatened by us. It safer for them to spread in the universe rather to stay concentrated on a single planet.
Well, that's it. They are other things I could react on, like the importance of emotions in intelligence, the nature of art, etc. But it would go to far from the subject of this post.
Regards,
Eric B.
This summurize my answers to the most common mistakes, according to me, that I can read on the topic of robotics and its consequences on economy, especially the "end of work":
* "When the first robots will take the place of human workers, new jobs will naturally appear to build, repear, create those robots."
WRONG: AI-based robots can simply do, by definition, any kind of jobs, including the new jobs they will generate. You can build, repear and even design robots using AI-based (intelligent) robots. The only area they will probably not compete with humans is art and social relationships.
* "Using taxes, we can make the cost of robots use higher than the cost of human work force. This solves the problem using a political/economical trick."
WRONG: The quality and the complexity of the tasks robots can perform can put man out of the game in a decisive way. Many high tech products today could simply never be done by humans, even if they where very competitive no-cost slaves. They simply don't have the skills robots have.
* "We have seen this before many many times: workers in the coton fields where replaced by machines, workers in factories where also replaced by machines. The services sector offered to these people new jobs and it will still do it. Nothing to be worried about."
WRONG: It's a new situation. The mechanization mentionned here is related to automatization of physical work force. Here, we are talking about intellectual force, which includes the kind of jobs done in the services industry.
* "By 20xx the computational power of computer will be equal to the one of the brain. This will be the beginning of the AI era"
WRONG: Two mistakes. First, the computational power of computers (as they are designed now) is not to be compared to the one the brain. It's simply different hardwares and it doesn't make sense to compare them (neurons transistors). Second, it is very naive to assume that hardware is the only thing. There is also something called software! Even if we have the computational power necessary for AI, we still have to design the software for AI, which is a very challenging task. Moreover, there is no clue today on how to design this software. It is typically a matter on which there could be no prediction. We could find the idea(s) that leads to AI within 10, 100 or 1000 years. Nobody knows.
* "AI is making progresses, just look at the chess programs or expert systems".
WRONG: Theses domains (chess, reasoning, database management,...) are just logicaly driven tasks than can be automatized easily. The real challenge of AI is today into robotics & vision: how could we design a machine that can find his way in the streets, take a bus to go to a previously unknown place, summerize a book, etc... The most difficult tasks seem to be the one we, humans, do the most easily. On the contrary, the easy tasks (chess, maths) are the one we are the worst at. AI is in the domain of the difficult tasks. And we have made almost no progresses there so far.
* "Intelligent machines will soon fight us and kill us all" (the Terminator fear).
WRONG: It is most likely that all these aggressive tendencies are specific to our still animal-like specie, inherited from a time where life was all the time endangered by hostile predators. If the machines feel threatened by humans, it will probably be more economical for them to leave the planet and move to space, using solar energy as a power source and taking the control of other planets for the material needs. Why risking a fight? Anyway, they would probably leave the planet, even if they don't feel threatened by us. It safer for them to spread in the universe rather to stay concentrated on a single planet.
Well, that's it. They are other things I could react on, like the importance of emotions in intelligence, the nature of art, etc. But it would go to far from the subject of this post.
Regards,
Eric B.
