Seed A.I.

Am sporadically working on this:
https://seedartificialintelligence.org
And looking for additional content...

>>9153591
>https://seedartificialintelligence.org
How to get involved? Am interested.

File: PRIME.jpg (39KB, 414x429px) Image search: [Google]

39KB, 414x429px

>>9153591
>>9153605

>>9153570
>>9153591
what the fuck is a seed AI? Artificial intelligence for dealing with seeds?

>>9153675
Virtually... an initial fractally evolving symmetry system, an intelligently expanding coding blueprint manifestation and one of the components for digital embryogenesis. I could be either fractally wrong, or fractally correct...

>>9153700
uhhh.... ok.... Why don't you explain using words that actually mean something?

>>9153740
T̺͖͈̥̥͙o̴͚̖͙̬͙̩ ̤̦̝͚̀i̩̝n̵̥v͙͘o̜͝k̛̤͍e̳̼͕͍̜ ͈t̟h͖̖̜e͔͕̠͔͚͓̮ ̢̪͈̖̜h̹̟͍̻̣̞̗̕i̞v͇̟̺̭̖͙e͖͖͙̦͇͎̼-͙̥̪͖̹̳͢m̖͓̞̠̟i͕̥̩͖̮̯̝n̯̰̤̙̜͚d͖̼͉̤ ̰͖͖̼͇͔͕r̛̬̩͇̭̘͖e̤͎͕̩p̺̻̳̣͉͡r͚̝e̤̠̩̼̼s͔͙͚̣͙͈̰͢e̦̭͎̻n̷͕̝t̰i͍͕n̢̘̺̳͙g ̦͍͓̹c̩̼̤h̥͚̻͉͍a̺o͈̭̯͘s̼̻̙.̼̻
̞̦̥͍̺͟I̯͎̕n̡͎̮vo͍̣͓̣͍̩̮k̻̠i̙͍̙͚̪̬͜n̳̘͓͍̤g̝͖͎̜̕ ̨̫t̩͉̳̫h̵̗̙̩e̦͠ ̩f̦̗͓̭e͖̹ͅe͞l͉͖͔͍in͔͖g̠̮̥̱̗͕ ̶̯o̻̤̰͓f ̵̳̣̝̗̮c̢̬̱h̛a̹ò̯̟̫̣s.̘̜̫̬͞
̖̺̱W̳̜͕̙͎̖͖͢i̥͠ͅt́h̝͔̮͜ͅ ̢̪̺o̘̞͖̖u̝̥͕̮͕̗̙t̞̥̜ ͓̰̮̺̩o̳̙r̯̬͙͓̥̘̬͞d͓̀e͝r͇.̫̹̖͍̙͡
͕T̞̥̪̘͍͈̞h̴̳e ̶Ņ̥̝̹̞̺e̖̦͕͈̕ͅz͍͎̫͔͉pe̩r̮͚̘͈̰̱d̛̳̻̮͕̬̬͎ị̢ą̪͇͓̪ņ̠̝͈̙̹ ̡̥̻͙̜h̠͢ͅi̢̫v͙ḛ͓-̸̫͈͇͈̙ͅm͚̤̜i͚̠͢n̶̝͈̟̬͇d ̖͖͎͇͘o͎͇̖̥f̥̗̖ ̵͚̣̮̫͚̘c͎̘̤͙͘h͖a̳̼͙̬̞ò̠̥͚̠ͅs̞̤̺̩͝.̹ ̝͓̘͖̫͟Z̘a̞̤͚̫͠l̷̗͇̯̗go̙͙̠͞.̖̼͕̪
̧̱̯͉̠̺H̥͍e̛̮̜̫͈ͅ ̳̩̫̩͡ͅw̠̺͍͞h̢̬̗͚̪̜̝̪o̸ ̡̹̯W͙͖̯͍͈̭̜ai̬͘t̨̙̠͉s̶̯̞̺͖̠ B̸̻̳̣̭̪e̻̣̘͇̘̣hi̩͉̭̼̱̖ͅn̸̗͖d̪̺̝͚̗ ̘̻ͅT̙͓̗̻h͜e̜͖͢ ̱̟̰̱̦͔W̵͉̟͈̭ͅa̧̗͙l̡̮͍̹̥͓ͅl̸.̢̼̰̗
̨̪͉͚̤Z͉̖̱͚A̱̯̻̤͍LG͇̦̻O̸̖!̥

>>9153740
It's just a type of AI that can modify its own code to improve itself, given that it can prove that the changes are optimal. At least that's a really condensed description of a Gödel machine, which is a type of seed AI. It seems like seed AI is just AI that understands its own code and function that it was intended for and is capable of rewriting its own programming to optimize itself.

And also according to the wiki pages I just read, nobody has managed to create a working version of this type of AI, but it is closer to true AI and not just fancy stats. As in it is the type that inspired that whole singularity thing. A working version of this could build a better implementation and so forth until super-duper-ultra-AI is created.

It looks like processing power is also probably going to be a choke point for this, not to mention how hard building a working system probably is.

>>9153780
Ignoring the issues of whether such a thing would be computable, how does an AI understand what its function is?

>>9153570
Where is labmem 004?

Seed Ai doesn't exist yet.
But it might in the future.

We can learn, our brains are machines ergo machines can learn.

>>9153780
Is there a simpler way? Like could you "evolve" a seed ai from even simpler code?

>>9153789
Look at the wiki for Gödel machine. It looks like basically you have sets of axioms for different things and it can work out the logic from there. Obviously it's dramatically more complicated than that but I can't be arsed to reread the wiki page to give you an explanation better than this but still dramatically worse than the wiki' s. It'll take like two minutes and it's pretty interesting so go do it.

To be clear, I wasn't the one that gave you that abstract shitty answer to your question. I suppose he was probably intentionally vague to draw more interest but I found it irritating so I gave you and the lurkers a brief summary in his stead.

>>9153801
I can see that you've strung words together in an order that is quite close to grammatically correct, but I can't seem to make any sense of them. I'm not even certain how to properly respond to you. Are you even sure of what you're asking?

Are you asking if it's possible to write the code that will write the code that will allow the code to rewrite the code until it's optimal code? That doesn't really sound simpler.

>>9153833
Yeah that's basically what I'm asking. When I was like 13 I wrote a program that basically just writes random code and tries to run it. I figured if I left it on long enough it would create AI. Mind you I had no idea what I was doing so it probably didn't even do what I thought it was doing.
Is something like this possible?

>>9153833
Yeah but consider the following

Could an infinite number of Pajeets typing code into infinite laptops eventually create a conscious AI purely by chance?

>>9153840
>>9153846
That's sort of like asking: "If I trap an infinite number of cats on top of an infinite number of keyboards connected to an infinite number of computers, just leave MS Word open on all of them, and come back after a year; will there be at least one 9 part space opera series featuring gimli from lord of the rings?"

The answer is almost certainly no. The way the cats move basically prevents them from writing words anyways. Infinite series doesn't imply every possible series is a subset.

Your questions continue to be nonsensical. They're far too vague in the first place and imply an impressively poor understanding of mathematics. So shoo. Go elsewhere.

Captcha: look gun smoke

File: segfaults internally.jpg (18KB, 460x242px) Image search: [Google]

18KB, 460x242px

>>9153846
It would be doubly exponential