Something I wrote elsewhere
[About an article in LinkedIn by Shelly Palmer: "I consider this to be a superb article."] My answer:
That puts me in one of the two classes of human beings. The members of this class will agree to the statements of Shelly Palmer (aspects of Artificial Intelligence).
The members of the other class will disagree. It is just as simple as that. Two classes. In the last 50 years I have had discussions with people about artificial intelligence and I am surprised that the argumentation of the members of the second class has never changed. Now I am deliberately offending this people: they behave like criminals during an interrogation. Don't admit to something that has not been proven yet!
Yes, it is true. Definitely a program can not excel in some disciplines that have not been tried yet. I have an ongoing discussion with a quite intelligent guy who claims that a computer could not have a feeling for art. Personally, I am still working on that argument although I don't consider it a "killing threat".
A hundred years ago, a very famous Austrian writer and journalist wrote about the general attitude of the people that could not accept the steam engine and its consequence, the train. If the train is going faster than 30 km/h the devil will come and fetch you. This is just one example of many, where people could not cope with the imagination of other foreseeing humans. I admit that I also have my limits of imagining certain contents of science fiction. Actually, there are just two of them; a) time travel backwards in time and b) beaming (if you know Startrek). Both hesitations of mine are based on the problem of very, very big numbers. Even if we have Terabyte hard disks nowadays, the actual storage must be much bigger for this disks in order to take care of statistical errors. But with beaming every single particle even on a subatomic stage would have to be reproduced without any error.
I had my own resentments in the past: why would be need color monitors? I could not see any advantage it it. I was wrong - just like Ken Olson from Digital Equipment Corp. when whe underestimated the importance of the personal computer.
I thought Go would be still a winner for the next ten years. I was wrong. However, I was right in one specific assumption. You have to work with several strategies and have them operate in conjunction. I have claimed this in a different context of quality control. It is not possible to test software by just using one strategy. (e.g. test automation) I could prove this by a geometry example. My assumption still holds true. It is one of the main features that AlphaGo combines two Neural Networks and Monte Carlo methods apart from the nitty-gritty part of handling the game.
If people argue about what the computer will never be able to achieve they will be surprised about some details which are not so favourable for the human comfort in the future.
That puts me in one of the two classes of human beings. The members of this class will agree to the statements of Shelly Palmer (aspects of Artificial Intelligence).
The members of the other class will disagree. It is just as simple as that. Two classes. In the last 50 years I have had discussions with people about artificial intelligence and I am surprised that the argumentation of the members of the second class has never changed. Now I am deliberately offending this people: they behave like criminals during an interrogation. Don't admit to something that has not been proven yet!
Yes, it is true. Definitely a program can not excel in some disciplines that have not been tried yet. I have an ongoing discussion with a quite intelligent guy who claims that a computer could not have a feeling for art. Personally, I am still working on that argument although I don't consider it a "killing threat".
A hundred years ago, a very famous Austrian writer and journalist wrote about the general attitude of the people that could not accept the steam engine and its consequence, the train. If the train is going faster than 30 km/h the devil will come and fetch you. This is just one example of many, where people could not cope with the imagination of other foreseeing humans. I admit that I also have my limits of imagining certain contents of science fiction. Actually, there are just two of them; a) time travel backwards in time and b) beaming (if you know Startrek). Both hesitations of mine are based on the problem of very, very big numbers. Even if we have Terabyte hard disks nowadays, the actual storage must be much bigger for this disks in order to take care of statistical errors. But with beaming every single particle even on a subatomic stage would have to be reproduced without any error.
I had my own resentments in the past: why would be need color monitors? I could not see any advantage it it. I was wrong - just like Ken Olson from Digital Equipment Corp. when whe underestimated the importance of the personal computer.
I thought Go would be still a winner for the next ten years. I was wrong. However, I was right in one specific assumption. You have to work with several strategies and have them operate in conjunction. I have claimed this in a different context of quality control. It is not possible to test software by just using one strategy. (e.g. test automation) I could prove this by a geometry example. My assumption still holds true. It is one of the main features that AlphaGo combines two Neural Networks and Monte Carlo methods apart from the nitty-gritty part of handling the game.
If people argue about what the computer will never be able to achieve they will be surprised about some details which are not so favourable for the human comfort in the future.
steppenhund - 14. Mär, 18:08
read 982 times
