|
Authored by: BJ on Sunday, March 10 2013 @ 11:35 PM EDT |
"[Mathematicians] they want to be confident that the algorithm produces the
correct answer for all inputs".
Correct as in terms of the program/software code and algorithms?
Or 'correct' as in terms of the output?
bjd
[ Reply to This | Parent | # ]
|
|
Authored by: BJ on Sunday, March 10 2013 @ 11:52 PM EDT |
Okay -- Since I'm submerged into Dan Dennett these days -- let me respond to my
own post with another question.
I said "That capability lies wholly outside of the computer, I'd
venture."
When does my assessment become fuzzy? When does said capability not become
exclusively 'human' anymore, and what implications does that have?
And does this very question (one 'computer' taking its input from another
'computer') bear in any way on what the patent(ability)-wars of the near future
will be like?
I recall from GEB that some programs came up with novel ways to prove theorems.
Did the algorithm 'foresee' this? Were we wrong to be astounded by it?
bjd
[ Reply to This | Parent | # ]
|
|
|
|
|