I'm glad you liked the intelligent outcomes without intelligence thing.wayfriend wrote:
If you copy the program and not the data, then you have another "being" which is exactly as intelligent but which is a different "person", having to learn things all over again, and capable of having a different "personality".
But if you copy the data ... then you have two copies of the same "person", where both believe that they are the original, possibly leading to disastrous consequences.
That being said,
It is possible that biological beings might have been designed so that only one parent is necessary.
Mix them up. Cull all but the improved. It would be artificial and arbitrary rather than random and uncontrolled. But it would be similar.
The first time that kind of idea occurred to me was because of people who can easily make change, recite multiplication tables, recall the sonnet they learned in 5th grade...but can't really add, subtract, multiply, divide, or talk about what the sonnet they know means, let alone read and understand a new one.
On the first couple...right. That's part of the experiments I was talking about.
And you can do all kinds of variations...from using a simple section of code on small/singular kinds of data, all the way up to millions of exact/complete copies working on different and duplicate kinds/sizes of data sorted in various ways/with varying emphasis on hierarchy [[if you've got a machine analyzing a batter, you might have a particular goal prioritized...say, how s/he swings, and what changes will lower strike-out rate, OR raise home-run rate. OR the machine might crunch the data, compared to other batters, to predict how s/he will perform in the future, and what can be changed to increase that performance, and/or which batter is/will be more valuable...I used batter, because I know those things are ALREADY being done with fair and rapidly improving success]]
On the next, some biological beings---even a few reasonably complex ones---do reproduce singularly. Some all the time, some occasionally depending on environmental conditions at the moment.
Last I knew, no one knows exactly how sexual reproduction got started...nor precisely why two-parent is vastly more common than three or more.
But, despite some seemingly serious disadvantages, sexual reproduction BOTH allows variation/adaptation/mutation AND error-correction/redundant systems/code-repair.
On the last, which overlaps in ways with the first couple...mixing/culling, those and competitive/cooperative interactions, modeled survival pressures... that's being done as well. It's an important part of how the bleeding-edge game-playing and other systems are made/evolve.
[[Which you almost certainly already know]]