I was taken with the language Fortran which means "Formula Translation" and reminds me a lot of when I first learned Algebra in Junior high or High School.
a useful thought about all this would be in relation to computer languages of A + B = C
The A and the B and the C are all unknowns. However, if you plug into A
A might equal 1, 2, 3,4, 5 6, 7, 8,9 or 10
B might equal 1, 2,3, 4, 5, 6,7, 8, 9 or 10
And C would equal whatever A and B added together would be.
For example if A is 2
and B is 3 then C equals 5
However, once you have the capacity to do something like this you also could represent each number symbolizing something else too which creates a decision tree.
For example, if you say that 7 is a decision tree meaning yes and 10 is a decision tree meaning no. then you can flow chart something to give it two choices about anything "Yes or NO" Yes meaning 7 and no meaning 10.
OF course all the numbers are completely arbitrary that you might chose for these decision trees but even in Artificial intelligence it's still using these simplistic decision trees (only billions of them) or more to give you the answers that AI is giving you.
Once you assign numbers meaning other things then they represent those other things whatever they are.
This then is the elementary basis for designing all computer languages every since, the A+B = C decision tree.
Of course I'm being very simplistic here so it makes more sense to you. Because all this gets more and more complicated the more decision trees you have going on.
What is a human being?
A human being makes thousands of decisions every day otherwise they don't survive. The same could be said for every computer and every computer program. Computers survive by executing decision trees one by one.
So, the danger is that some decision tree might be fatal for one or more human beings or it could just as easily save a life or 10 lives or 100 lives or 1000 lives or 1 million lives or even a billion or more lives. But, it could go the other way too because at core it is just mathematics and algorithms functioning really as a basis of how it makes decisions.
The biggest initial problems with AI likely will be hidden from the public as much as possible because they wouldn't want this known (the developers) of wars between different AIs programmed by different people and using different values (each one) that the others.
Just like everyone as a person is different also every programmer is different and has slightly different values that they automatically program into whatever they are writing. Then you are going to be setting loose these programs that design other programs whether they are a simply as what is programming what you are reading right now at this blog site.
I'm not having to program this at all and "compose" is literally programming into HTML everything I'm typing here. All I have to do is to type in all of this and then hit either "Update" or "Save" or Confirm to save or publish what I'm writing. But, the problem becomes when I quote from another source online programmed either by a person or another Auto-Coder like the one here at blogger.com and they don't get along for one reason or another and then I have to find ways to modify the output here at my site or give the whole thing up and not publish what I'm quoting from another source.
However, it is also true that this happens less and less now than it used to starting in 2007. Then what I wanted to quote wouldn't work up to 50% of the time and slowly but surely this has reduced down to around 10% to 20% of what I'm trying to do for you doesn't work. So, obviously programs have been written to reduce this problem between auto-encoders who I guess could be considered a type of early AI too just like "Autopilot" in early Teslas likely would be considered an early form of AI too.
No comments:
Post a Comment