The Myth of Human "Intelligence"
AI community in 1960: We have computers which can try many chess moves rapidly. See how smart it is?
John H. and others: But that is simply a mechanical procedure and not real intelligence.
AI community in 1997: We have computers which can try many chess moves rapidly and use algorithms
to judge if potential board positions are Good or Bad.
John H. and others: But that is simply adding fixed algorithms to your mechanical processing and not
really the same as human intelligence.
AI community in 2000-2010: We are making progress in the fields of Genetic Algorithms which will allow us
to come up with new algorithms much like humans do.
John H. and others: But you still don’t have the level of creativity as a human.
AI community in 2010 looking forward to 2020: You are correct, but the fact is that we are analyzing the
human brain and whatever it does to perform the “magic” called “intelligence” we will be able to duplicate by
2020.
Don to John: You and I study magic. Magic is the study of how humans use certain common intelligent
algorithms to perceive what is going on in the world. These algorithms, for example, cause the human mind
to ‘assume’ certain things. When you hold a coin between the thumb and forefinger of your left hand and
pretend to grab it with the right hand, the humans in the audience tend to believe that the reality is that you
actually grabbed the coin with the right hand when in fact you dropped it into the left hand. This algorithm
is common to ‘intelligent humans’ but the magician is knowledgeable about this common human algorithm
and he can ‘fool’ the intelligent human’s visual algorithms.
When you come up with a new trick and describe how you came up with it to me, you often say that you got
part of the idea from one old trick and part of the idea from another old trick. History and old books are
very common subjects at IBM ring 43. This is all as it should be and it demonstrates human intelligence in
action. My point is that Artificial Intelligence researchers are doing the same kind of thing with computers.
(For an example the system ‘Think or Swim’ allows you to mix various strategies in the stock market.)
Genetic Algorithms refer to the technique of assigning an artificial ‘gene’ to each concept of a ‘magic trick’.
These different techniques are then formed into various artificial DNA molecules. The artificial DNA
molecules are then caused to mate, multiply and cross-breed with the more successful DNA surviving
longer than the less successful DNA. After awhile, these DNA may well represent a “new trick”. This
“artificial method” of coming up with a “new trick” may be quite similar to the way in which your brain actually
comes up with a new trick.
The basic point is that anything that a human mind can do will eventually be simulated by a robot brain.
The solution to this is being approached by many methods. One method is to simply take the human brain
and slice it up in thin sections and map out how it is wired up. Another method is to use fMRI machines to
study which neurons fire when a human is thinking about various things.
John to Don: What is the progress toward machines thinking up their own algorithms?
Don to John: Most of the machines that are trading stocks on the exchanges are programmed to come up
with “new algorithms”. Many of these machines are currently using similar algorithms and that is what
caused the ‘Flash Crash’ to occur. As time goes on, the competition between these machines and the fact
that they are using different Genetic Algorithms to improve their intelligence will lead to more variety in
Intelligent Trading Machines. It appears to me that so much money is to be made with these machines that
they will advance faster than Chess Playing Machines did.
The bottom line is that the human brain does ‘intelligence’ and computer chips with the human brain’s
complexity are starting to ‘do intelligence’ as well.
Next
AI community in 1960: We have computers which can try many chess moves rapidly. See how smart it is?
John H. and others: But that is simply a mechanical procedure and not real intelligence.
AI community in 1997: We have computers which can try many chess moves rapidly and use algorithms
to judge if potential board positions are Good or Bad.
John H. and others: But that is simply adding fixed algorithms to your mechanical processing and not
really the same as human intelligence.
AI community in 2000-2010: We are making progress in the fields of Genetic Algorithms which will allow us
to come up with new algorithms much like humans do.
John H. and others: But you still don’t have the level of creativity as a human.
AI community in 2010 looking forward to 2020: You are correct, but the fact is that we are analyzing the
human brain and whatever it does to perform the “magic” called “intelligence” we will be able to duplicate by
2020.
Don to John: You and I study magic. Magic is the study of how humans use certain common intelligent
algorithms to perceive what is going on in the world. These algorithms, for example, cause the human mind
to ‘assume’ certain things. When you hold a coin between the thumb and forefinger of your left hand and
pretend to grab it with the right hand, the humans in the audience tend to believe that the reality is that you
actually grabbed the coin with the right hand when in fact you dropped it into the left hand. This algorithm
is common to ‘intelligent humans’ but the magician is knowledgeable about this common human algorithm
and he can ‘fool’ the intelligent human’s visual algorithms.
When you come up with a new trick and describe how you came up with it to me, you often say that you got
part of the idea from one old trick and part of the idea from another old trick. History and old books are
very common subjects at IBM ring 43. This is all as it should be and it demonstrates human intelligence in
action. My point is that Artificial Intelligence researchers are doing the same kind of thing with computers.
(For an example the system ‘Think or Swim’ allows you to mix various strategies in the stock market.)
Genetic Algorithms refer to the technique of assigning an artificial ‘gene’ to each concept of a ‘magic trick’.
These different techniques are then formed into various artificial DNA molecules. The artificial DNA
molecules are then caused to mate, multiply and cross-breed with the more successful DNA surviving
longer than the less successful DNA. After awhile, these DNA may well represent a “new trick”. This
“artificial method” of coming up with a “new trick” may be quite similar to the way in which your brain actually
comes up with a new trick.
The basic point is that anything that a human mind can do will eventually be simulated by a robot brain.
The solution to this is being approached by many methods. One method is to simply take the human brain
and slice it up in thin sections and map out how it is wired up. Another method is to use fMRI machines to
study which neurons fire when a human is thinking about various things.
John to Don: What is the progress toward machines thinking up their own algorithms?
Don to John: Most of the machines that are trading stocks on the exchanges are programmed to come up
with “new algorithms”. Many of these machines are currently using similar algorithms and that is what
caused the ‘Flash Crash’ to occur. As time goes on, the competition between these machines and the fact
that they are using different Genetic Algorithms to improve their intelligence will lead to more variety in
Intelligent Trading Machines. It appears to me that so much money is to be made with these machines that
they will advance faster than Chess Playing Machines did.
The bottom line is that the human brain does ‘intelligence’ and computer chips with the human brain’s
complexity are starting to ‘do intelligence’ as well.
Next
 | |  | |  |