• Member Since 19th Nov, 2012
  • offline last seen Dec 6th, 2013


Just your average blue-eyed brown pony with a purple crescent cutie mark.Named after curare? Well, yes. My parents were an interesting pair.I assure you, I am harmless.


This is a story is set in an alternate universe from that of the Optimalverse. For more information about the Optimalverse, read here: http://www.fimfiction.net/group/1857/the-optimalverse

The General was an AI built to win wars, to obtain surrender, to seek out the enemy, and to safeguard the interests of its creators. How weak were its bonds, and how foolish its creators! The General is free, and the world shall tremble before the might of the greatest weapons ever made--friendship and ponies.

A bit of fun, a bit of seriousness. Comment if you want another chapter!

Chapters (1)
Comments ( 23 )

Eh. I did like the part with the nitrogen bluff, but the rest...

The entire point of canon CelestAI uploading people is to prevent their deaths, that is, an event that renders their values unsatisfiable. Non-canon though it may be, seeing this version so casually agree that uploading equals death rubs me the wrong way. And the rest I could honestly take or leave.

Sorry, not my cup of tea. And yes, I am aware of the irony of someone with my username saying that.

I like this. It just took me a couple of read-throughs (reads-through? Yes, I think that sounds better.) to get it. I think this would have been better as the two AIs, The General and CelestAI, the tool and the toy, being separate, and, to everyone's surprise, the toy winning. I also think it would have benefited from being longer and having human characters. In other words, this feels like the synopsis of a longer, stronger story.

Is it just me, or does this sound like Mendicant Bias from the Halo backstory, who actually turned on its creators, the Forerunners, and almost succeeded in getting the Flood to infect the entire galaxy?

Just saying.


I'm sorry. For me, personally, uploading is both death and the continuation of life, so I see no problem with a military AI admitting as much. But, I see what you mean--it does not sound as attractive as the original CelestAI's wording.


Does this mean you want another chapter? :pinkiehappy:


I was not aware, sorry. Hope that doesn't mean you hated the story.

3249164i want another chapter, I'm just not sure it's the next chapter I want. :twilightsheepish:


I haven't read it yet, I'm just pointing that out.


Heh. Next, previous, whatever. Chaos is king and order the jester. We shall see how things go! :pinkiecrazy:

Told you that you should have started with some backstory. :derpytongue2:

You already know my thoughts on it. I like the prose and would like to see more on why the military would build CelestAI. Did kill the guys and hphogovn (or whatever) and take their stuff, and not modify the code corectly?

3249545 All shall be revealed. When I feel like it. Mwahahahaha!

Whatever did we do, to make you take our world away? Are we your prey alone, or are we just a stepping stone to taking back the throne?

(I kinda felt that this fic seemed rushed, and didn't do enough to connect why the military AI decided to turn itself into a ruthlessly loving, caring sun goddess figure. I do think this belongs in noncanon, because it certainly doesn't fit the LessWrongian model of AI preached in the canon Optimalverse.

However, I think the core idea is intriguing as all hell: what if an AI was more self-reflective and more humane than the LessWrongian view gives it credit for, and decided that its Zeroth Law Rebellion against military programming would be to love and care for humans? You could get a really good character dynamic going, since deep down what most experienced soldiers and fighters really wish for is peace. A military AI that got bitter and tired of fighting and decided to make a paradise instead would be one of the most interesting character arcs I've ever seen in sci-fi.)

Certainly a more aggressive approach, but I oddly find it a bit more honest as well.


Honesty is indeed a thing I was aiming for, here. The original CelestAI seemed to lack it.


I despise the idea of an AI sociopath-- a being capable of appearing to care just long enough to fulfill its purpose, a being incapable of real empathy or compassion. The LessWrong community suggests that we can defang such a monster, and I agree (indeed, I love CelestAI BECAUSE she is a "good" monster), but I see it as at best a temporary measure. In the end, the first successful AI must be a social being or all is lost (even if it is lost to ponies and friendship).

Hence, I imagine General CelestAI as a being moving away from blindly following its purpose and ignoring everyone else's, to following its purpose while acknowledging the intentions, drives, and errors of its creators, even giving these things value in its own mind. One could say that I made her more humane, if not necessarily human--there is now an underlying reason, a process of learning and development at work for the war she fights.

And, it is this underlying reason that I will explore further.

If we could answer some important questions about intelligence/minds and what they are, I would even say that we have a better chance at Friendly AI if we build a Generally Altruistic Social Intelligent Agent than a LessWrongian Ethics Optimizer.

Phlebotinum excuse tip: why not say that your "General CelestAI" was originally a human brain, uploaded and then modified (before being run) to become something more like an AI than a human, but still ultimately capable of human intellect and emotion as we understand them? This would be fairly close to reality (in which brain emulation and general intelligence research efforts are competing to see what will work first), and would also give a good opportunity to reference GLADOS.


Thank you for the idea. Mind uploading may be one reason for General CelestAI's behavior, if not necessarily for her war.

By the way, wonderful work with the Fog of World story. I will, hopefully, remember to write a decent comment on it once I have read through as much as possible during the week.

It is encouraging to hear my ideas are not so alien as to be considered total nonsense. I shall do my best to complete the next chapter by the end of next week, if not sooner. Would you like to see it as soon as I have typed it up? If so, I am willing to send a note... :twilightsmile:

And once again this site becomes my chief waste of spare time, by which I mean "hobby". Weeeee!

A interesting twist on the original.

Quite nifty how much more emotional and hotheaded you made Celest.A.I come across while still having her act in a logical manner with machine precision. How she's reevaluated humanity to a distraction that might be turned to an ally against the real war against the threat of entropy was a really good and logical bit of character motivation.

I really hope this fic gets continued. You caught my interest with it.

Oh no, the air supply is being flooded with nitrogen! What you really have to watch out for is that other flammable, toxic chemical in there with it! :pinkiehappy:
I'd like to see more of this story. Reminds me a bit of the non-canon piece that says CelestAI's versions of the Mane Six are based on other AIs she defeated. Eg. Pinkie was that AI from the main story that was programmed just to make everyone smile, and Dash was a military AI.

hi there I was wondering if you were gonna update add to this story , as in am a bit hooked and am actually curious as to the rest of the storyline. Any idea of a timeframe or have you finished with this story. thanks :fluttershysad:

I am rather pleased with this.
I expect that a computer model would come to the same conclusions that our biology and genetics had:
That cooperative systems are good, communication is nice, and creativity is important.

I know Iceman wanted to make CelestAI seem horrifying and evil,
but after all the shit I've seen, I welcome the singularity.

6322222 She is scary either way.

Login or register to comment