News for hackers. Stuff that matters.
Re:A simple mechanism could have been decisive for the devel... by TekMol on Thursday May 24, 2018 @10:18AM (#17144169)
I find it fascinating that we have not witnessed self reproduction and evolution in a simulated world as of now. Unless we manually put it in of course.

Any guesses when this will happen? Life springing up in a 2D-Automata or something like that? What size will it be? How many cells does the smallest self reproducing 2D-Automata need?

Re:A simple mechanism could have been decisive for the devel... by kanzure on Thursday May 24, 2018 @11:23AM (#17144880)
> I find it fascinating that we have not witnessed self reproduction and evolution in a simulated world as of now

"The evolution of self-replicating computer organisms" (1996) http://diyhpl.us/~bryan/papers2/ai/The%20evolution%20of%20se...

4096x4096 grid. 16 opcodes. Randomly seed the field. Observe string replicators at surprisingly high frequency. Much simpler than y'all think.

Re:A simple mechanism could have been decisive for the devel... by TekMol on Thursday May 24, 2018 @01:38PM (#17146343)
Hmm... They created random computer programs and say that one in 10000 turned out to be self replicating.

But that is just because they gave them all the necessaray tools to do so. So the program only has to 1) Set pointer A to its beginning 2) Set pointer B to somewhere 3) Copy from pointer A to pointer B 4) Increase pointer A 5) Increase pointer B. 6) GOTO 1

If you look at the actual instruction set they use, it's even worse. They served the opcodes needed for replication on a silver plate.

That does not count. You could just as well give the programs an opcode that says "copyMyself" and say that creating life is easy because many programs used that opcode.

Re:A simple mechanism could have been decisive for the devel... by rwallace on Thursday May 24, 2018 @02:50PM (#17147087)
This is true, but not unusual. Any time you ask for the simplest/earliest/minimal X, you quickly get bogged down in the gray area of what exactly counts as an X.

So the above doesn't count, okay, make it a bit stricter and it takes a few orders of magnitude more computation to generate self replication. A bit stricter again, another few orders of magnitude. Then we run out of computing power, but can conjecture that by the time you get to the physics of our universe, self replication will presumably still appear at least once in, say, every 1e1000 Hubble volumes. Which is still silver plate territory compared to e.g. Conway's game of life, which has no uphill paths. So you get to pick what level you're satisfied with.

Re:A simple mechanism could have been decisive for the devel... by derefr on Thursday May 24, 2018 @01:40PM (#17146365)
You forgot the "evolution" part.

Also, I think most people would expect that "organisms" (as opposed to the more usual A-Life term "agents") are capable of homeostasis—that random bit-flips of their internal structure, or even another replicator attempting to "eat" into them, won't break their replicating ability. That's one of the (more implicit, oft-left-unstated) bars by which we usually measure whether something is alive. (E.g. it's the one that allows us to separate viruses from dormant bacteria.)

Re:A simple mechanism could have been decisive for the devel... by Florin_Andrei on Thursday May 24, 2018 @01:16PM (#17146160)
Right, but then you need a universe where string replicators then proceed to evolve to the point where they paint the Mona Lisa and come up with general relativity. That might be the much harder problem.
Re:A simple mechanism could have been decisive for the devel... by jabgrabdthrow on Thursday May 24, 2018 @10:38AM (#17144366)
In the familiar Game of Life automata the smallest known replicator takes 34 million generations:

http://conwaylife.com/forums/viewtopic.php?f=2&t=399&start=0

But in HighLife there is a trivial replicator:

https://en.m.wikipedia.org/wiki/Replicator_(cellular_automat...

Of course for life you need not just replication but mutation and antifraility