[General Battuta]

I'm saying that the very framing of the "problem" as being "local vs global" is already problematic in itself, for it pressuposes that all the agents are behaving in a rational, optimal, strategic manner, constantly trying to get into a Nash equilibrium state. But all evidence shows that at least, we humans, actually do not behave like this. In recent studies it was shown that only two groups of people actually behaved like that, economists and psychopaths - which makes a lot of sense if you think about it, and it would explain a lot of our intellectual fascination with psychopaths in the first place but that's a different topic.

Brah I think you're forgetting what field I was in. People clearly aren't econs.

People do gather info and make decisions based on rules and heuristics. The irrationality you're talking about is the product of rules and heuristics that evolved (in a local, satisficing, good-enough way) across time. Once you abandon the idea of teleological evolution, you can understand that human behavior is governed by a bunch of locally optimized heuristics that try to get together to solve much bigger, newer problems.

You're charging in here to argue points that are in fact already central to the discussion and agreed upon.

The only way it works in BP is because it is written with a non-desperation feel-good ending in mind, so it creates this idea that we are somewhat special (unlike all those other 99% of species that were wiped out, ala Mass Effect style), and it is this grace of specialness that saves our asses (we can sugarcoat it with "moral" decisions all we like, but still), where in reality there's no indication of this being the case at all.

Why would you assume this? All these years of BP making the precise opposite argument and you think it'll abruptly veer back the way it came? Interesting.

[General Battuta]

The framing of 'local' vs 'global' in fact explains departures from 'rationality'. It's the definitions of 'rationality' and 'agent' that are misleading.

[General Battuta]

Waha I also remembered one more class of people with reduce departure from 'rationality' in a number of experimental setups: the clinically depressed!

A number of heuristic biases, such as positivity bias, are actually adaptive in a range of common situations.

[The E]

GB, could you please remind yourself that an edit function exists and that you don't need to triple-post?

[General Battuta]

Battuta phone posts
His think joy is too verbose
A vowel man weeps

[Phantom Hoover]

I think, Luis, that you're thinking too teleologically about the processes in question. For instance:

I'm saying that the very framing of the "problem" as being "local vs global" is already problematic in itself, for it pressuposes that all the agents are behaving in a rational, optimal, strategic manner, constantly trying to get into a Nash equilibrium state. But all evidence shows that at least, we humans, actually do not behave like this. In recent studies it was shown that only two groups of people actually behaved like that, economists and psychopaths - which makes a lot of sense if you think about it, and it would explain a lot of our intellectual fascination with psychopaths in the first place but that's a different topic.

Agents in competitive games converge to Nash equilibria because that's the stable strategic environment for the game, not because they aim to reach a Nash equilibrium and rationally try to bring it about. They don't need to know anything about game theory or strategy at all, or even have the ability to think.

[Snarks]

Now we can double down in the mental process and declare that, in some weird, obscure manner, our psyches are actually tuned to work in a much deeper Nash-like rational equilibrium that hasn't just been understood and modeled yet, but once such models were done, we would actually see how perfectly rational and strategic all of us really are. And in that sense, terms like "solution" and so on would make sense again. But all of this is entirely teleological. There's no "solution" here to be had, because all we have in our genes and culture is mostly the past and our drives, not any purpose or reasoned demanded output for our existence.

If we are to look at mankind's "solution" to the shivan problem, we might even run to several philosophical problems before even correctly defining what exactly the "problem" is, but if we handwavely and generally define it as "surviving" the shivans, then I'd venture that a lot of adaptation would need to occur, for all our past, both culturally and genetically, are just not sufficient tools to deal with this kind of stuff. Natural selection was great, but it deals terribly with new original problems of their own that have the capacity to instagib the entire genepool / civilization. Our famed intelligence is little better than that, most of our thoughts are incredibly stupid and incoherent, science, as the other writer said some time ago, is a human achievement, not a "human trait". Will all the social and scientific processes that the GTVA produced be enough to deal with a deep-time deep-space challenge? My guess is 100% no. The only way it works in BP is because it is written with a non-desperation feel-good ending in mind, so it creates this idea that we are somewhat special (unlike all those other 99% of species that were wiped out, ala Mass Effect style), and it is this grace of specialness that saves our asses (we can sugarcoat it with "moral" decisions all we like, but still), where in reality there's no indication of this being the case at all.

This is a very good critique. I'll respond to this first since it'll make more sense. First, let's get some definition stuff out of the way. A "game" is the abstract interaction between the actors and a set of states or "win" conditions. Solutions are states in which the system is in equilibrium, so for the most part, we can say they're the same thing. I will actually refrain from using the term equilibrium except in the specific case of a Nash equilibrium. Actors are the distinct entities that are making decisions within the game. All of this stuff is conceptually from the field of game theory although notions such as stability or solutions are concepts applicable in the broader fields of mathematics as well.

Defining the "problem" or "game" is certainly an important part of the question. What I think has been intrinsically proposed is that the problem is the one of existence. The "win" state is very simple; the actor in question has to continue to exist. To "lose" is simply to cease having a form. Note, in this game, everyone can win or lose; there does not need to be a loser for there to be a winner. In this case, both Terrans and Shivans have won the game. But here's where the difference is important. Terrans won by employing an unstable solution whereas Shivans seem to have won by employing a stable solution. The Terran solution, the post-FS2 GTVA, is probably an unstable solution because the GTVA X number of years later will be a different GTVA (especially if we assume things like BP). The FS2 Shivans (and this is based on how I have interpreted the Shivans) have not changed in billions of years (perhaps even more). They fly around supernovaing stars and killing anything that gets in their way. They (and this is my interpretation) have not evolved their technologies or doctrine. And by this static set of choices/behaviors, they have managed to continued existing.

I'm saying that the very framing of the "problem" as being "local vs global" is already problematic in itself, for it pressuposes that all the agents are behaving in a rational, optimal, strategic manner, constantly trying to get into a Nash equilibrium state. But all evidence shows that at least, we humans, actually do not behave like this. In recent studies it was shown that only two groups of people actually behaved like that, economists and psychopaths - which makes a lot of sense if you think about it, and it would explain a lot of our intellectual fascination with psychopaths in the first place but that's a different topic.

Those assumptions aren't necessarily prerequisite to the idea of local/global solutions or equilibrium.

What needs to be noted is that the framing of the question or actors is essential to the interpretation. In economics, Nash equilibria tend be solutions where neither actor (usually a person or a firm entity) will want to deviate from their current behavior. But this doesn't give us enough information to analyze the overall system if we deviate slightly (say someone makes a mistake or an irrational actor comes in and makes a random decision). If we refer back to the idea of stability, an unstable solution is one where some random action (perhaps from irrationality) can result in the players moving away from the Nash equilibrium. If we were assuming an unstable solution, then your critique of irrationality is entirely valid.

However, solutions can also be stable. In a stable solution, some deviation will be corrected for by whatever forces exist in the system over time. Most economists usually concern themselves with only stable solutions because unstable solutions are subject to the critique that mistakes/irrationality does exist.

Two bits of empirical counterpoints: 1) Economists do have models where actors may be assumed to be irrational (i.e. making random choices, making choices on an entirely algorithmic pattern, or even purposefully picking choices that are harmful to them). 2) You can abstract away from the idea of perfect rational people. An actor can be anything so long it is involved in the defined game. Refer to Battuta's post on rocks (and why they are also engaging in the game of existing.)

Now it's fair to also allow a bit of leeway for when we say something is in a solution. If something hovers very closely to a solution, then usually we can say it's a solution. Imagine if the post FS2 GTVA does not change at all. People's genetics remain the same, they have the same system of governance, and so on, with the exception that every few years people start wearing different hats. We can probably say this GTVA is a solution for the problem of existing.

So the idea of solutions is a fine framework for analyzing the Shivans. But as you've pointed out, it is also very important to identify what we mean by solutions. Here, we have a mostly clear game and solutions that address it appropriately.

If we start critiquing the idea of what it means to be an actor though or what it means to maintain identity as the same actor, we'll get into really funky discussions about what it means to exist. If the Shivans vaporized all of humanity and we became dust, does that mean we have also "won" the game of existing by picking the strategy of being space dust?

Edit: I love this forum. I can't think of anywhere else I get to discuss game theory, philosophy, and fictional space aliens from a video game.

[General Battuta]

I think for me the vital insight about winning the existence game is this: a very strong tactic is to gain the ability to reshape existence to suit your tactics and deny those of others.

A simple case is farming. We can use land and crops to create food surplus. Later we may develop fertilizer and artificial climates and specially bred crops. By this means we reshape the ordinary terrain of food scarcity in our favor.

As we get better at this, we begin making models that predict the tactics of others. We predict how they will act and we counter play. This is Steele in BP. He uses understanding of his opponents to precisely disrupt their tactics. In the farming analogy, we might choose crops that flourish in the presence of pests our opponents are vulnerable to.

This tactic is so good that I would expect galactic civilizations to build hegemonies from it. 'We have a model of everything you may do and we have already solved the counter plays.'

Yet we see none of this in the Shivans. They do not show any interest in building models of us. They don't target their force at gaps in our capabilities.

I agree with you, wholeheartedly and with great delight, that the Shivans represent an optimum in the existence game. But I don't think that points to stasis. I think it points to tactics that prey on the tactics I described above.

[Luis Dias]

I'm saying that the very framing of the "problem" as being "local vs global" is already problematic in itself, for it pressuposes that all the agents are behaving in a rational, optimal, strategic manner, constantly trying to get into a Nash equilibrium state. But all evidence shows that at least, we humans, actually do not behave like this. In recent studies it was shown that only two groups of people actually behaved like that, economists and psychopaths - which makes a lot of sense if you think about it, and it would explain a lot of our intellectual fascination with psychopaths in the first place but that's a different topic.

Brah I think you're forgetting what field I was in. People clearly aren't econs.

People do gather info and make decisions based on rules and heuristics. The irrationality you're talking about is the product of rules and heuristics that evolved (in a local, satisficing, good-enough way) across time. Once you abandon the idea of teleological evolution, you can understand that human behavior is governed by a bunch of locally optimized heuristics that try to get together to solve much bigger, newer problems.

You're charging in here to argue points that are in fact already central to the discussion and agreed upon.

My problem with this answer of yours is that I had foreseen it and already pre-answered to it right above it:

....But all of this is entirely teleological. There's no "solution" here to be had, because all we have in our genes and culture is mostly the past and our drives, not any purpose or reasoned demanded output for our existence.

Why would you assume this? All these years of BP making the precise opposite argument and you think it'll abruptly veer back the way it came? Interesting.

You're a damned good writer, and so I fear assuming anything with you, but my feeble mind cannot but assume that this is the unavoidable weakness of the plot, that no matter how you dice it, it will either come from a "specialness" of terrudankind or sheer luck. You might hide it so well that it will be invisible but for those who are going to nitpick everything till they find the crack, but it's like a perpetual machine for me: no matter how clever you made your maths, there's a crack somewhere down there that just doesn't cut it.

[Snarks]

So laying out the strategies:

UEF - be peaceful and exist on what's given in the environment
GTVA - eat the peaceful by manipulating the environment
Shivans - be the environment?

The Shivans are such a massive entity that their mere existence shapes how everyone else reacts.

On a past note about local/global strategies: if the Shivan strategy truly is the optimal global strategy for existence, then does that imply the most effective strategy at winning the existence game is to exist on an unprecedentedly large scale? It almost feels like we're saying to be the biggest winner, you have to be the biggest winner.

[Luis Dias]

Sorry for the doublepost, but doing renders and commenting here is stretching my PC resources thin .

I want to answer to Snarks (and partially to Phantom Hoover), because I see I was defining terms incorrectly, thus following terribly the reasonings.

Sure, I totally see the point in analysing these things in game theory in that manner, and alas, it's somewhat of a sketch of what the GTVA mega-computers were actually doing in making those social predictions that urged them to go to war in Sol. I surged in the thread because I kinda didn't like Battuta's simplistic vision of future prediction by free agents as "local" vs "global", when all I can see is just a difference of scale in analysing future possibility states (the tiger hunting does have lots of problems to solve in order to get its meal, the cowboy has his own shares of other problems to solve as well), with apparent qualified differences being just the emergent result of more neurons working at the problems at hand.

What I sense about the "difference" between those two above mentioned is that they come from two separate parts of our brain, one we just call our primal drives or needs, the other we traditionally call "intelligence", and we as human beings feel the inner tension between the two, so why not call them apart? But again, it's like chess. Once it was thought that "long term thought" was so different than "local thought" that AIs beating super-GMs was a sheer impossibility. This was proven false, all that was required was more neurons thrown at the problems (discussions about Kasparov's brain neuron nets being hugely more efficient at playing chess than Deep Blue's rooms of chips notwithstanding).

I agree with you, wholeheartedly and with great delight, that the Shivans represent an optimum in the existence game. But I don't think that points to stasis. I think it points to tactics that prey on the tactics I described above.

Tactics of beating tactics of beating tactics. There could be countless levels of this in the shivan design. Purely the result of extreme intelligence (if we define it as capability to predict future states). Now I know, it could be just random chance, or natural selection of some billions of years ago, that these agents just happened to come across these tactics, but since we also know the computing power of a single shivan corpse, I'm betting that intelligence was the main driver here.

On a past note about local/global strategies: if the Shivan strategy truly is the optimal global strategy for existence, then does that imply the most effective strategy at winning the existence game is to exist on an unprecedentedly large scale? It almost feels like we're saying to be the biggest winner, you have to be the biggest winner.

Size does give the shivans an advantage. I wonder how they behaved when they were smaller in scope, their strategies were obviously completely different.

[General Battuta]

Man it's not foreseeing if you're just saying what we've all been arguing! The heuristics we use to navigate the world arose from an ateleological existence game, and it's hard to get those blindly evolved local heuristics to work for our new, 'rational', teleological mode of global reasoning. You're trying to argue against something by restating it. All we can do is keep saying 'yes, exactly!'

So laying out the strategies:

UEF - be peaceful and exist on what's given in the environment
GTVA - eat the peaceful by manipulating the environment
Shivans - be the environment?

The Shivans are such a massive entity that their mere existence shapes how everyone else reacts.

On a past note about local/global strategies: if the Shivan strategy truly is the optimal global strategy for existence, then does that imply the most effective strategy at winning the existence game is to exist on an unprecedentedly large scale? It almost feels like we're saying to be the biggest winner, you have to be the biggest winner.

I think of the Shivans as an environment, yeah! But not just a system of systems interacting, like weather or tectonics. Something much more dangerous.

To borrow more of your language, I think of the Shivans as a function. A transform. It accepts input and produces, eventually, a response. What input?

Exactly the kind of stable existence game equilibrium tactics you talked about earlier. And the output is the devastation of that tactic. The Shivans are whatever network occupies the space between.

Imagine a point in the game space so unstable that when it encounters a stable equilibrium it immediately diverges wildly - and does so in a way that eventually selects for a counterstrategy.

It's not the kind of super smart chess mind I was talking about earlier. Rather, it's a function intrinsically hostile to that mind, one that eats the enemy's local victories and transforms them into global defeats. It can fight anything because it has no internal structure to predict and destroy: only engines for transforming input into radically divergent behaviors that eventually produce a winning tactic.

Note that this crazy anti-chessmaster function may of course produce chessmasters from its sea of possibilities. They'll be effective at optimizing force locally, but may create fragility by over optimizing and becoming predictable. In BP these are called anima. They recruit elements from the totipotent base state to satisfy their local objectives, and they last until something takes them out.

And yes - you have to be huge to do this. Ridiculously so. It's such a locally wasteful tactic that you have to be able to absorb galaxies and millennia of losses in the name of ultimate victory.

The Shivans have given up on the Princess Bride problem that haunts chessmasters: if you do this, I do that! But you know I know, so you'll do THAT, so I must do THIS, but you know I know you know...

They've stepped back and said, ok, how do we take the limit of this problem and become the victor at infinity? How do we become a function that transforms strategies into counterstrategies?

In a sense the Shivans are very lazy. They ask you to work very hard to figure them out, or to throw all your strength at them, or whatever, man, it's all cool with us. They outsource the job of figuring out your weaknesses to you! They're just exceptionally good, in the long run, at listening. They'll grow the organs they need to kill you once you've made it clear what those should be.

There's a reason BP said the Shivans were calculated, not made. Though that means a few things.

[Luis Dias]

That sounds incredibly smart, and I notice the calculated bit. Has been a while since I last dwelled my thoughts into this so I missed a couple things from canon here, but I do wonder. I wonder at how magical it just seems to be. I wonder how it is possible to have all these diagnosing abilities without any hint of intelligence. Kinda reminds me the weakest point from The Dark Knight IMHO, the one where the Joker just laughs at the schemes of everyone else, while he just goes and "chases dogs", he has no clue what the hell he's doing and so on. And yet, he pulled several batmans in the movie and predicted almost everything about those people's schemes.

I do understand the point of using intelligence locally as a tool, and I am indeed fascinated by this idea of intelligence being used by a simpler function to do its bidding ( a phrase I never got to use in my Aken's Diary BP fanfic ran like this: a panicked crew member of the Iceni anima started shouting to his crewmates "Don't you see? They turned us into mere middleware!!"), and the designed point is indeed important. It also smells like a poignant statement over the human mind itself, as if saying, "You think you're in control of your body, don't you? Well, let me write you a scenario that might hint otherwise!".

[Kestrellius]

Welp, I'm still feeling out of my depth here, but I did a bit of reading on game theory and I get the jist of it.

Also, I just had an idea. Suppose someone created a computer to run simulations of various scenarios. Specifically, let's say the GTVA created one to predict the actions or motivations of the Shivans.

And then, suppose everything. Every possible situation, run through the computer. Victories, defeats, both, neither. But the computer is programmed to prioritize success, so when running a scenario, the computer doesn't progress to the next event until an acceptable outcome has been achieved and documented. Then, if that scenario arises, it is dealt with precisely in the manner laid out by the optimal run.

Do you see where I'm going with this?

[Luis Dias]

That looks like a simple chess AI player. It's a weeed more complex than that. See "local" achievements vs "global" achievements - the difference between short-term outcome and very very very long term outcome, the difference between a toddler playing chess and Kasparov.

[Kestrellius]

Oh. Yeah, my point was something completely different.

Alpha 1's superpower isn't immense flying skill, it's the ability to retry his or her missions an infinite number of times until an optimal outcome is achieved. What I'm proposing is that this, rather than simply being an out-of-universe phenomenon, is the result of Alpha 1 being connected to a computer simulation device in some way. Maybe Alpha 1 doesn't even exist, and it's the computer flying the fighter once it runs the ideal simulation.

Furthermore, this concept allows all campaigns to coexist, despite the fact that even if you take alternate-universe shenanigans into account, it's hard to justify the Shivans having entirely different origins and motives. However, if most modded campaigns are simply speculative simulation scenarios...

Since the Shivans are pretty much an unknown variable -- the GTVA has no idea what they want -- it would make a certain degree of sense to just go through every single motivation they could think of and prepare for anything.

[karajorma]

It's not particularly hard to believe that someone with an incomplete understanding of what the Shivans were could go off in completely the wrong direction. It would also explain why the GTVA kept helping Bosch with his plans. If you take the simpler "Bosch was right all along" approach you basically make them out to be idiots who simply didn't think things through.

Why?

If you take the Bosch was right all along point of view you make a set of assumptions about the GTVA which I'm not particularly fond of.

1) You assume that humans definitely can understand Shivans well enough to be able to communicate with them. You further assume that an alliance with them is possible.

2) You assume that despite the number of people studying them only one middle ranking officer (Bosch probably wasn't an admiral at the time) understood what was going on.

3) Bosch must either decide to keep what he has figured out to himself for some unknown reason he doesn't bother to explain, even when monlogueing or he does tell the GTA and they ignore him because he's a poopyhead.

4) The GTVA bury the information Bosch used for no good reason. Yeah, I know military intelligence can be stupid but we are dealing with a genocidal species here. You would have thought someone would want the research to continue.

5) The GTVA aid Bosch several times because of his work on a secret project. This is despite having such a vague knowledge of what the secret project is that for all they knew it could have been a super recipe for chilli. One so good that the Shivans would just have to know how to make it.

5b) Or the GTVA knew that it was a communication device and they let Bosch continue to develop it because leaving it to a rogue admiral who has already caused one war is better than asking the janitor for the keys to that room you've labelled "Shivan stuff we can't be bothered to look at - TOP SECRET" so that you can restart the research you should have been doing for the last 20 odd years anyway.

6) Having continually ****ed up when they finally find the Iceni the GTVA spend the minimum resources they possibly can in order to retrieve the information on board. Cause there's no real issue if the Shivans blow it up and we never figure out what message Bosch sent.

Basic conclusion - The GTVA are a bunch of idiots.


On the other hand, if we assume that a select few within the GTVA know a bit more about the Shivans than Bosch things start getting interesting.

Suppose that the truth about the Shivans is as horrible as we've been talking about on this thread. There's no way you'd want something like that to become public knowledge. Further suppose that for some reason the best strategy for the GTVA is to communicate something with the Shivans (It really doesn't matter what).

The GTVA have a problem now. They can't just open the portal and knowingly start off a war without explaining why they did it.  And looking like they blundered into the war looks bad on them.

Enter Bosch. Bosch had some crazy, incorrect views about the Shivans. The GTVA realise they can use him as a cat's paw to achieve their goals. It wouldn't be hard to make sure that the scientists who end up working for him are their people and since Bosch probably can't read quantum pulses they can be sure that it will be their message that gets sent.

With this scenario most of the seemingly idiotic decisions the GTVA made suddenly make sense. Bosch gains an incomplete picture of Shivan motivations precisely because he isn't a scientist studying Shivans. He's ignored because he is wrong. The information on Shivans is buried not out of stupidity but because it needs to studied in much greater secrecy. And the GTVA aid Bosch cause they want him to succeed, not because they think allowing a rogue officer to develop WMDs is a good idea.

In the end, we don't even need Shivan motivations to be understood or even understandable, we only need the GTVA to believe that the message they were sending would do something useful. Whether it was an offer of an alliance, surrender, peace treaty or even an Independence Day style computer virus is rather unimportant. All that matters is that the GTVA thought it would be worth it.

[General Battuta]

I don't think any of your assumptions are necessary but reading the rest I think it's just because we have different interpretations of 'Bosh was right' on the way to similar objectives.

This convo is spoken to a lot in the COLDMORN leaks.

[General Battuta]

The Bosch monologues are too lovingly written and too central to be thrown out as pure delusion, though

[karajorma]

Not pure delusion. The work of someone very intelligent working with incomplete data.

Large parts of what he says is correct. The rest is based on sound data but interpreted incorrectly.