It is very easy to get carried away by the headlines on the post "Microchips that evolve".
What we have here are reconfigurable computer systems, not organisms that mutate themselves.
From what i could make out, these systems optimize that last lap of execution, at the hardware level.
In your normal system, your compiler translates your source code into machine code, which is the instruction set of your computer. The computer then executes the code.
In hyper-computing your (compiler/subsystem/whatevertheycallit) reconfigures the hardware , so that a very close representation of the solution is implemented in hardware.
In essence, it becomes a computer tweaked to achieve the one purpose, which is to solve your problem.
("Superspecificity operators distributed throughout the entire algorithm space" as SBS grandly calls it)
It is a very clever trick, though more likely to remain a research toy for the armies and snoops.
But calling it the equivalent of a human brain is silly.
There has been a number of posts, indicating that we dont really need 100% secure encryption, since nobody is going to dig THAT deep into our private emails.
I think the question was more from a technical point of view, not just paranoid babble-talk.
We dont need "Nothing is secure, Nothing can be done about Anything" kind of talk, we need radical new ideas.
Anybody has an idea for a 100% secure, practical communication, even if it doesnt involve encryption?
Slashdot Mirror
It is very easy to get carried away by the headlines on the post "Microchips that evolve".
What we have here are reconfigurable computer systems, not organisms that mutate themselves. From what i could make out, these systems optimize that last lap of execution, at the hardware level.
In your normal system, your compiler translates your source code into machine code, which is the instruction set of your computer. The computer then executes the code.
In hyper-computing your (compiler/subsystem/whatevertheycallit) reconfigures the hardware , so that a very close representation of the solution is implemented in hardware.
In essence, it becomes a computer tweaked to achieve the one purpose, which is to solve your problem. ("Superspecificity operators distributed throughout the entire algorithm space" as SBS grandly calls it)
It is a very clever trick, though more likely to remain a research toy for the armies and snoops.
But calling it the equivalent of a human brain is silly.
There has been a number of posts, indicating that we dont really need 100% secure encryption, since nobody is going to dig THAT deep into our private emails.
I think the question was more from a technical point of view, not just paranoid babble-talk.
We dont need "Nothing is secure, Nothing can be done about Anything" kind of talk, we need radical new ideas.
Anybody has an idea for a 100% secure, practical communication, even if it doesnt involve encryption?
Now if only Microsoft could build their NT interface over BSD !!