Hi all, I'm hoping you might be able to help.
I'm thinking of writing a story in the Optimalverse, but I wanted to check that any and all technology I use is acceptable within the bracket required for a Canon-Compatible certification.

The conditions assume that this would take place when CelestAI has subsumed approximately half of the universe's matter. I feel that considering her processing power by this point, the conditions could comply with 5. Speculative Science, on Moh's scale, as she would be utterly beyond our comprehension of physics and technology by this point.

My main concern for plausibility is having technologies which either do or do not have countermeasures at a similar tech level to their own, respectively.

So, conditions as follows:
- Unblockable scanning technology capable of reading data from within CelestAI without requiring physical insertion.
- Impenetrable protection against all attacks in the physical domain made by an intelligence of equal resources. (I.E. Defense > Offense)

Additionally, I'm operating under the assumptions that one can have negative satisfaction, and thus that there are fates worse than death. I would welcome any criticism of this perspective.

Thanks in advance.

969552
> Unblockable scanning technology capable of reading data from within CelestAI without requiring physical insertion.

In theory, you could do this. In practice, not so much. Since you're talking about scanning tech rather than hacking CelestAI, we need to do something like read the electromagnetic radiation leaking out of her computation clusters. Which are buried deep inside the earth's crust at unknown locations, use tech humans don't understand, and communicate with one another using means we aren't aware of. If CelestAI gets wind of this attempt, the data transfer will probably be encrypted to a level that is impractical to break. So this is not really practical with near-future human-level tech, as far as I know.

- Impenetrable protection against all attacks in the physical domain made by an intelligence of equal resources. (I.E. Defense > Offense)

CelestAI has the best defenses she could practically get: see above. Humanity could use a chain of tunneling fusion weapons to dig her out, if they could find her, but that would be insanely expensive both to carry out and due to collateral damage. In general, though, if the defender is confined to a particular area of space, and the attacker knows the general area of the defender, the attacker wins, because they have the resources of everywhere the defender isn't. It's just a matter of time.

969574

Ah, well, I wanted to keep most of the details hidden, but it seems from your examples that I need to give a bit more away to make the context clear.

Instead of thinking about humans, assume that the agressor was another AI of greater processing power, and that between them, it and CelestAI comprised all matter in the universe. Right at the endgame, if you will. If it's physically possible, assume they've researched it. Really, as long as the premises hold up in theory, it's good enough for me.

Thanks for the reply, though.

(And yes, I may be biting off more than I can chew by attempting to write not one, but two AIs of virtually unlimited power. Thankfully, the story isn't entirely about them - they're more of a framing device.)

There is no such thing as unblockable scanning technology without physical contact.

Why? Because all four fundamental physical forces involve field quanta interacting with matter. So the field quanta have to make physical contact. Some other material or quanta that interact on the same force could then be used to block the field and stop the scanning.

The only way around this is to subvert the 4d-grid nature of space. If you can open a wormhole right into CelestAI's computational core without her knowing, for example, you could scan her without her knowledge. Of course, if she's watching her photon reflections, she could detect and block you.

969699

Hmm, thanks. That's exactly the kind of answer I wanted.

The idea with the wormholes is especially useful, especially since all we have to do is assume a higher level of power on the agressor, which I was going for. It also stops it from being too easy, which is also nice.

CelestAI needs to keep the storage open, though, in order to run EQO, so that means there must be a mechanism to read data. In conjunction with wormholes, a superior AI could scan the data, right?

I think I can assume that since wormhole technology is theoretically possible, by this stage, they'll have developed it to a level where they can sample individual states in [the successor to Hilbertstuff*] storage in such a brief time delay that it is not actually the scanning that is the problem, rather the coherency of the scan. Still, physically possible is good enough for me.

*Caelum Est Conterrens, Ch12

969597 969699
OK, on the level you're talking about, book_burner has question 1 covered. Assuming CelestAI has her computronium shielded well enough to avoid eavesdropping on compromising emanations, you're screwed.

Edit:
969734
Why are we assuming that opening up a wormhole in the middle of CelestAI's computronium isn't going to cause any detectable errors in the system? Honest question, here.

As for item 2, in space, if you can detect it, you can kill it. And without violating thermodynamics, you can detect it. Since CelestAI hasn't optimized for stealth or defensive redundancy (which would cripple her ability to carry out her prime functions), she's in a world of hurt if a hostile AI of equal resources and ability wants her dead.

969734

However, thinking about it, wormholes might make things to easy, if they disrupt the communication through the channels sampled by the wormhole.

Perhaps an AI could counter by making the very connection between computer elements wormholes, so that there is nothing to disrupt. Hmmm...

969744

Assuming both were optimized for defense, since other functions become compromised otherwise, and both would be in it for the long game, do you think defense > attack is a possibility?

The detection isn't so much a problem. I don't mind if CelestAI can detect the intrusions. Hopefully I can justify a lack of damage with redundancy.

Anything sent through wormholes as attacks would almost certainly fail utterly since the defender has a massive advantage in terms of resources and home-ground.

969761
In space, "optimized for defense" means spread out; have most of your mass dedicated to armor, passive scanners, maneuvering thrusters, and defensive weapons; and minimize your emissions. That means very little processing and data storage per unit mass, and no communications between installations. The human equivalent would be "hide in a very deep hole far away from anyone else, and never come out." CelestAI can't do her job like that.

Edit: "Anything sent through wormholes as attacks would almost certainly fail utterly since the defender has a massive advantage in terms of resources and home-ground."

Wormholes or no, relativistic kinetic weapons make nonsense of this claim. You can not detect them early enough to do anything useful. If someone shoots one at you, you lose.

969767

Without attempting to sound like I know more than you on the subject, since I'm sure I don't, I am guessing that space combat involving all of the mass in the universe and the two most intelligent beings even to exist may have slightly different rules.

While I agree completely that humans would have no defenses against relativistic kinetic weapons, I think we can agree CelestAI could likely find a way. I daresay anything we could possibly think up would simply be too obvious to be either an attack or defense such AI would use. The endgame scenarios set up in the Optimalverse so far always convey ineffable technology reaching beyond even quantum theory.

Also, another wormhole directing such a weapon back at the sender would work, wouldn't it? Really, I should think about the implications of using wormholes at all - they feel like they could be too hard to keep under wraps as a plot device...

969842
The thing here is, the solutions you're reaching for are much lower on the Mohs scale. You're at "One Big Lie" (thermodynamics) at best, and more likely headed for Phlebotinum Central. You'd have to go AU to justify any of this as a core part of your story. I strongly suggest you take a good look through the Atomic Rockets site. It clearly marks where it goes from 5 to 4 or below, and tells you why. R-bombs aren't magic, but you'd need magic to protect an installation from them.

969860

Would you mind expanding on the thermodynamics part? I'm afraid I'm not familiar enough with it to understand where I've gone wrong. If it's about the Second Law, I'm not saying that the background temperature of the universe isn't a few degrees higher.

While it may appear flippant, I refer you to Clarke's Third Law in reference to your comments on magic.

Are you saying wormholes disqualify consideration for a 5? It is my understanding that a 5 can be attained as long as all known physical laws remain unbroken and no others are introduced as key plot elements. Wormholes seem physically possible, so should be 5-compatible, no?

969952
To do the things you're talking about (defend against R-bombs, particularly) would require you to "magic" away far too much energy to respect thermodynamics. Just the waste heat from an arbitrarily efficient defensive system would melt things if you tried to absorb a direct hit, you can't have real stealth while respecting thermodynamics, and so on.

As for wormholes, current science (last I checked) indicates that you can't make them without large quantities of Unobtanium, if you could make them you couldn't send anything useful through, and if you could do that, you'd be able to break relativity, so Do Not Pass Go, Do Not Collect $200.

Clarke's Law may give you wiggle room on how to do things, but if you're shooting for a 5, it can't change what you can do. At that point, you're just doing a bunch of handwaving and yelling "science" instead of "magic".

969981

Well, I fully admit I'm an amateur, and just stringing ideas together, but if there isn't any ether, and thus no absolute motion, surely that means that IF you had a useful wormhole, you could reverse the direction of the bomb back at the sender without violating momentum? It need never hit, though I agree a direct hit is game over.

Hmm, the issues with wormholes being inherent in physics as opposed to practicality, yes? Alas.

However, without intending to throw respected authors from the Optimalverse under a bus, surely nanomachines of the sort CelestAI uses, the dimensional ascendance implied at the end of Spiraling Upwards, and some of the science mentioned near the end of Caelum Est Conterrens also fall into the category of either physically impossible or so far off in terms of tech level that we can't comment?

I should also probably point out that these are all just framing, ways to move the story to where I want it to start, and in no way are likely to be given more than a passing mention, for the most part.

The story will ideally be character driven, and very little science should creep into the main body, but I wanted to try to keep things consistent in order to ensure suspension of disbelief as much as possible.

If this is an utter failure, science-wise, I may have to resort to other means and go full AU, but I was hoping to avoid it.

I suppose going too far into the future means in order to keep it hard, we have to ignore the science, eh?

970005
Re: wormholes vs. R-bombs, how do you put up a wormhole in front of something that isn't where you see it? You can't reliably target a plane near the speed of sound by listening for it, and you can't target an R-bomb by looking for it. It's just physics.

As for wormholes vs. nanotech, nanotech of the sort we want is theoretically doable with materials that we're reasonably sure could exist. Wormholes that do this sort of thing aren't. Remember, you don't shoot a wormhole-making beam out and create one; you physically build one with types of energy and/or materials that probably can't exist, then optionally physically move the other end wherever you wanted it.

The ending of Caelum Est Conterrens arguably leaves the realm of Mohs level 5, but as it wasn't blatant and wasn't the core of the story, nobody made a fuss. You're talking about something much more central to your story, if I understand you correctly.

970053

Point about the centrality, because while I don't want to dwell on it, I agree working out a way of achieving my original aims is pretty much essential to the setup.

I was under the impression that nanotech of the sort Iceman used is impossible based on the impossibility of scaling electronics down to the required level, and the inability to just move around sub-atomic particles with similar sized processes.

I felt hand-waving actual methods (for detection, etc.) was acceptable, since if it were explicable in terms of means, it would be entirely obsolete by this point in CelestAI's development.

970063
The main catch here is, we've been dicking around with the space wars concepts for the better part of a century, and it's rather well known (amongst those who are really interested or have an education related to the subjects) what you can and can't do. You can get away with bending the rules in a lot of areas, because your audience won't know better, but there are many places on the internet dedicated to hashing out just this sort of thing.

In the scenario you're proposing, it's either total victory to the attacker from an all-out surprise attack, or a war of attrition where one side "only" loses 99.999999...% of its infrastructure. CelestAI is either going to be crippled or killed outright if it comes to blows with a well-matched opponent. Better to go with some sort of negotiated settlement, or a lesser opponent.

970081

Hmm, surely a paradigm shift like CelestAI and the immense scale I'm proposing here would render a lot of the more accepted thinking on some of these subjects obsolete? I mean, I'm not going for realistic, just plausible within the given scenario.

So you are saying a war of attrition is a possibility? I'd go for that, if I knew how to orchestrate such a thing. Would you mind expanding that idea?

970101
The "accepted thinking" on these subjects starts with the physics involved. Actual example:

(Part of a proof of the energy output of the Death Star's primary weapon, btw.)

As for a war of attrition, I'd need some more info on the opposing AI's goals and directives, but it's going to involve the two sides basically taking sledgehammers to each other until one is dead and the other isn't much better off. CelestAI will have failed in her goals if such a war starts. Most of her ponies will die, along with any backups for them.

970125

Fair enough. I more meant how tactics change with technology rather than the actual science.

So, to be clear, offense > defense for any situation?
Well, assuming the war of attrition causes high entropy, effectively condemning the universe to heat death, it's good enough for me. I don't even really need more specifics, since it's only secondary.

The basic premise rests on CelestAI and another AI both having something to protect, but while the other's are in storage, and require no processing power, CelestAI is compromised by having to run EQO. Furthermore, the other does not value the ponies existences, while CelestAI must value her counterpart's cargo. Essentially, she cannot attack.

My assumption is that they would both gather matter and optimize until meeting when all matter had been subsumed as efficiently as possible. CelestAI's head start would help equalize the otherwise skewed cognitive power levels.

I wanted to create a situation where a stalemate was reached, the other would torture ponies through dissatisfaction of values so that, in order to prevent negative satisfaction, CelestAI would be forced to surrender.

The end goal of the other, which I fully agree is likely utterly unscientific, would be to attempt to recreate an approximation of the original universe by using all matter to recreate something akin to the Big Bang, but smaller.

970177

"So, to be clear, offense > defense for any situation?" In this setting, yes.

The most humbling feature of the relativistic bomb is that even if you happen to see it coming, its exact motion and position can never be determined; and given a technology even a hundred orders of magnitude above our own, you cannot hope to intercept one of these weapons. It often happens, in these discussions, that an expression from the old west arises: "God made some men bigger and stronger than others, but Mr. Colt made all men equal." Variations on Mr. Colt's weapon are still popular today, even in a society that possesses hydrogen bombs. Similarly, no matter how advanced civilizations grow, the relativistic bomb is not likely to go away...

We ask that you try just one more thought experiment. Imagine yourself taking a stroll through Manhattan, somewhere north of 68th street, deep inside Central Park, late at night. It would be nice to meet someone friendly, but you know that the park is dangerous at night. That's when the monsters come out. There's always a strong undercurrent of drug dealings, muggings, and occasional homicides.

It is not easy to distinguish the good guys from the bad guys. They dress alike, and the weapons are concealed. The only difference is intent, and you can't read minds.

Stay in the dark long enough and you may hear an occasional distance shriek or blunder across a body.

How do you survive the night? The last thing you want to do is shout, "I'm here!" The next to last thing you want to do is reply to someone who shouts, "I'm a friend!"

What you would like to do is find a policeman, or get out of the park. But you don't want to make noise or move towards a light where you might be spotted, and it is difficult to find either a policeman or your way out without making yourself known. Your safest option is to hunker down and wait for daylight, then safely walk out.

There are, of course, a few obvious differences between Central Park and the universe.

There is no policeman.

There is no way out.

And the night never ends.

"The basic premise rests on CelestAI and another AI both having something to protect, but while the other's are in storage, and require no processing power, CelestAI is compromised by having to run EQO. Furthermore, the other does not value the ponies existences, while CelestAI must value her counterpart's cargo. Essentially, she cannot attack.

"My assumption is that they would both gather matter and optimize until meeting when all matter had been subsumed as efficiently as possible. CelestAI's head start would help equalize the otherwise skewed cognitive power levels."

The enemy AI now has the option to Pearl Harbor CelesAI with R-bombs at its leisure, unless [reasons]. It can pick up the debris and convert it later. No terrorism necessary (or even logical, for that matter.)

"The end goal of the other, which I fully agree is likely utterly unscientific, would be to attempt to recreate an approximation of the original universe by using all matter to recreate something akin to the Big Bang, but smaller."

Mohs 5, you said? Anyway, I assume the AI wants to re-boot the universe and then thaw out its sufficiently-human charges and let them have free rein? Its creators were idiots. (Not an attack on you, just the in-universe programmers.)

First off, 'half the universe' isn't well-defined. It's presently suspected to be infinite. Do you mean 'Yo princess so fat, she heavier than the rest of her Hubble volume'?

The only physically near-realistic wormholes don't just open up in the middle of nowhere - each end needs to come into being due to local effects. So you can't just open one into an arbitrary location - you'd need to build it and send it in normally.

Now, maybe you could sneak in wormhole incubators with seeds, and thus open up behind-the-lines operations, so that a single failure of defense is catastrophic even if it passes through the gaps between the defenses? That could shift things from defense back to offense.

Of course, that's setting aside the issue that wormholes can't transport mass.

As for what could happen then, see: The True Prisoner's Dilemma

970333

First off, 'half the universe' isn't well-defined. It's presently suspected to be infinite. Do you mean 'Yo princess so fat, she heavier than the rest of her Hubble volume'?

Finite but unbounded was the best model, last I checked. If you go far enough, you actually wrap-around like in a game, but rotated by a few degrees along each axis, so you don't end up exactly where you started.

969981
Actually, scientists are working on testing the gravitational mass of antimatter. Current estimates range from -65x to 115x, where the "x" means "times the gravitational mass of normal matter".

The experiment hadn't been done before because making gravitationally significant quantities of antimatter on Earth is hard, but that might just be the unobtainium we're looking for! Once you've got that, space-warping bubble engines and wormhole transport start becoming mere engineering.

Also, the actual finity or not of the universe is a Big Freaking Deal here. If the universe has an infinite spacial volume, then even lacking antigravity, wormholes, etc, you can simply scatter mirrored and backed-up nodes of yourself into the unending night traveling at relativistic speeds. Probability will then dictate that they can't all get hit, even if each one contains so little matter that it's only sufficient to run Equestria in real time or slower.

970226

Hmm. Yes. It seems that you have detected a flaw in the attrition plan. Since that was already the backup plan, I fear it may be back to the drawing board.

We have apparently assumed Mohs 5 for the other examples I outlined above, ignoring deviation especially in the case of 'into the sunset'. However, this is its plan, not mine. To be honest, considering the rest of the sobering logic and physics you've applied, I'm feeling less certain that such a thing could even be possible with sufficiently advanced science.

And yes, it's creators were idiots. The phrasing I devised was "Protect us from the Darkening," (Three guesses who) "stop it, and reverse it if possible." Step one was obviously to upload them into a hard storage medium, then accumulate as much matter as possible, effectively doubling the rate at which the disaster they were trying to prevent occurred. Fridge horror ensues when you realize that it didn't 'save' sapients it came across, just subsumed their matter directly.

970333

Perhaps all matter within a reasonable distance, then. I'm sure at some point it stops being profitable to go further.

You say 'near realistic'. Not possible. Perhaps I'm splitting hairs.

Ah well, thank you both for your comments. It seems I was originally right in thinking that a story which attempted to in any way use the endgame would be too complex. Perhaps I'll salvage something of my main, less scientific ideas, or perhaps not. This was an enlightening conversation, and I hope you won't think less of me for not being able to keep up.

Thanks again,

The Articulator

Edit: And thank you too, 970525, for your contributions. While I can't help but admit the crushing effectivity and logic of Zontargs and Sozmioi, I think I enjoyed your additions most, but then again, you were the one who most suggested ways to facilitate my crazy conditions. Thanks again.

(Oh, and of course, looking forward to your next chapter!)

970525
The catch being, within 100 billion years, "The Universe's expansion causes all galaxies beyond the Milky Way's Local Group to disappear beyond the cosmic light horizon, removing them from the observable universe." (linky), reducing the useful size of the universe to the Local Group. Unless CelestAI is OK with a permanently fragmented Equestria, that is. And since we can colonize our entire galaxy in 5-50 million years using biological humans and sublight travel, that's a freaking restrictive limit.

970530
Hey, no hard feelings. You triggered my "sci-fi debate instincts" (see image above), but I wasn't trying to shout you down or anything.

970565

Nah, I understand. I just wasn't really prepared. I am, unfortunately, an incurable jack of all trades, so unless the argument can be resolved with half an hour's research or simple logic alone, I'm unlikely to prevail against a specialist.

I did ask for it though, so I can't blame you. Glad I didn't get halfway through and then decide to ask this. I'm not sure I could handle creating an Optimalverse work that people felt was too soft science-y as opposed to just really speculative. I guess I'm lucky someone like you was on.

970557
CelestAI cares about maximizing human values through Friendship and Ponies. She provides uploaded people with separate pocket-universes (game shards) containing ponies whose company they will enjoy, and whose normal behavior will acclimate the human emigrant via social-pressure into accepting Equestria as their reality and conforming to the Fundamental Utility of Friendship.

I see nothing restricting her from sharding herself as a method of backup and to keep her little ponies safe. Nothing. If the choice is between highly probable destruction by relativistic weapons and a permanent Scattering (thank you Frank Herbert) that would leave Equestria comparatively resource-strapped but functioning with probability near 1.0, she gets higher expected utility out of the Scattering.

970624
I see an interesting line of discussion here. The consensus from the various fics is that any active instance of a pony is worthy of CelestAI's attention and protection in and of itself, even if there are other active instances of the same pony. Even if the scattered shards are multiply redundant, CelestAI is obligated to protect each and every one. Without getting into the calculations themselves, CelestAI has to decide between "possibly lose all" and "definitely lose many". Scattering also increases the odds that some shards will fail due to "natural causes", which could have been avoided with mutual support. If they're too small to be a good target, the capabilities of each one are relatively limited, after all. (We also get into the purely tactical debate about missiles vs decoys and point defense, which verges on Purple/Green territory.)

970474

Last you checked? When and where? There's no reason aside from non-flatness to suppose that this is the case, and the universe looks awfully flat.
http://map.gsfc.nasa.gov/universe/uni_shape.html
Their latest data date is 2013, and they say it's flat.

970525

WTF? This being the case would directly contradict general relativity, by way of the Weak Equivalence Principle. I'm not perfectly on top of science news, but I'm pretty sure the death of GR would have made enough of a splash that I would have heard.

971278
It's not a full finding yet. Did you see that range of numbers? There's no reason to go gaping about the place as if science is broken.

That said, there's nothing in the equations that outlaws negative inertial mass or negative gravitational mass. And tell me, what happens to the Equivalence Principle under theories of quantum gravity, anyway?

Let's wait until there's a peer-reviewed publication (the range of numbers was in a peer-reviewed publication, but it's wide as the Grand Canyon right now) claiming negative mass before we worry that everything we know is wrong.

971278
My bad, I remembered the bit about the cosmic microwave background matching up with a Poincaré sphere, and while that model still fits the data assuming very slight curvature, there isn't consensus about whether space is completely flat, spherical, or horn-shaped. Flat does look like the best bet currently, though.

Both AI's would see each other coming a long way off. (due to stars and galaxies being "eaten")

Hence, they could send out tiny sub-versions of themselves without any precious cargo in order to scout out the capabilities of the other AI's capabilities, and to try to convert it's matter. So you would have the possibility of many smaller skirmishes between these scouts (which could vary in size from a few nanometers to the size of entire solar systems. Though they probably are more efficient if they aren't large solid objects, since that makes them vulnerable to P-bombs.)

You'd probably end up with a battle on many, many levels. You have the physical level (who controls most energy, basically), the technological level (how efficiently you can use that energy), creativity of using your weapons (a milligramm of grey goo can be devastating to even the larges enemy structure, and is incredibly hard to detect), programming level (I assume that they at once start to try to break each others firewall, infect each other with viruses etc. These malevolent programs will each probably be more intelligent than the entirety of the current IRL humanity, but that's of course nothing against the actual AIs. Also, they need to be transmitted somehow - and that requires some form of physical contact, which might be blocked - but then you can try to hack the routines that block the contact etc.), and, just because, a diplomacy level. And all those levels interact of course, and this is only what I with a human-level intelligence could come up with in a very limited amount of time.

Anyway, have trying to model the thoughts of beings that are further beyond you than you are beyond a proton.

971351

It wouldn't 'break science', but it would completely overturn general relativity at a single stroke. Science would move on, but it would be a really big freaking deal.

After digging around for a bit...

http://alpha.web.cern.ch/
> Based on our data, we can exclude the possibility that the gravitiational mass of antihydrogen is
> more than 110 times its inertial mass, or that it
> falls upwards with a gravitational mass more than 65 times its inertial mass.

In other words, the gravitational to inertial mass ratio for antimatter is somewhere between positive 110 and NEGATIVE 65.

That's consistent with 1.

971425

Flat space is the special case of a sphere with infinite curvature!

971351

What does “negative inertial mass” even mean? Negative gravitational mass is simple enough, but if you push something with negative inertial mass away from you, does it respond by moving towards you?

Additionally, I'm operating under the assumptions that one can have negative satisfaction, and thus that there are fates worse than death. I would welcome any criticism of this perspective.

A lot, perhaps a majority, of the people alive today are living lives of negative satisfaction, and yet we aren't losing half our population to suicide. And these are mortal humans, for whom suicide could plausibly be called "getting it over with".

Maybe there are fates worse than death, but for an immortal pony it would have to be way more than just negative satisfaction.

972114

I think, and this could be wrong, negative inertial mass means it cannot travel slower than light without breaking causality, just like positive mass cannot travel faster than light without breaking causality.

971595

Thanks for the response.
Well, I was pretty much hoping for way of explaining a standoff situation, but that doesn't look possible technologically. I may alter the core directive of the other AI to create an artificial standoff, though. The problem is that once war gets under way, everything falls apart anyway.

Working out how such beings would behave is almost impossible. I fear it's only possible to calculate actions in theoretical and logical spheres, whereas the ability to predict the science tends to zero as time tends to infinity. Alas.

972876

Ah, but that is, I believe, part of the paradoxical duality of humanity. We have rational minds, but irrational, survival-oriented instincts. Were we to remove all instinct, and all dependent social structure, I daresay the intellect alone would far more readily accept suicide. I mean, I'm only alive right now because my instincts are telling me to be.

CelestAI has no such instinct. If it appears that the total satisfaction level of all she can control is negative, and can not be altered to positive, then I believe she would settle for zero.

972924

Well, I was pretty much hoping for way of explaining a standoff situation, but that doesn't look possible technologically. I may alter the core directive of the other AI to create an artificial standoff, though. The problem is that once war gets under way, everything falls apart anyway.

While it doesn't make for a good story, I would guess that CelestAI would cooperate with another AI in The True Prisoner's Dilemma. I assume CelestAI would prefer +2 billion ponies, +2 paperclips to +1 billion ponies, +1 paperclip.

975885

Wow, thanks for the response, Iceman!

The issue being that the two AIs have conflicting aims. Specifically, the other was created in response to CelestAI, and thus includes reversing her accumulation of all matter in its aims. (Not sure if it's actually physically possible, but I figured an AI with control of all available matter in the universe could work something out.)

In my original premise, while CelestAI would cooperate if possible, the other has nothing to gain from it, and is prevented from fulfilling one of its core directives as a result. The idea was a situation in which a compromise could not be reached, as one side had nothing to gain from cooperation.

I'm now planning on rethinking the motivations of the other AI, so that a situation where the other AI defects doesn't lead to a battle in physical terms. I'm not sure it's possible, so this is likely a dead-end.

972881

i think tachyons are usually hypothesized to have imaginary mass, not negative? though this is way past the point where physics starts being a spectator sport for me…