There might've been an over-update on what past-Keltham described as the difficulties of interstellar travel faced by dath ilan inside of non-magical physics, where you don't just have mortal-Aroden doing Interplanetary Teleports all over.  He'd guess that wasn't an industrial spaceship, just one magically hacked together; or a magically star-traveling lifeform from the Dark Tapestry, or some other draw from all-other-possibilities.

Seeding diamonds out of hot 1% methane 99% hydrogen in a zero-gravity space-looped demiplane is much much easier than interstellar travel the hard way, and nobody was selling synthetic diamonds to Efreet in the City of Brass before they got there.  (Unless the Efreet themselves, and not just Efreet trade goods, are partitioned by planet; and the Golarion-trading Efreet can't buy Wish diamonds for even their own use from Efreet who trade with hypothetical industrialized planets; it's possible, but improbable, counterevidence that adds to other counterevidence.)

How does Carissa feel about demands that divinities at least stop actively hindering mortals from making their lives better?  Does this feel like something she's terrified he can't get, will constitute asking for too much, and then Creation ends?

She thinks he shouldn't say 'or I'll destroy the universe' about it? 'or I and my allies won't do any positive-sum trades with you and will consider ourselves in a low-key state of war with you' is fine! That doesn't entail destroying the universe! Carissa will absolutely back him on withdrawing-cooperative-relations from gods that get in the way of mortals making their lives better. She'll back him on trying to assassinate Them!

Carissa agrees wholeheartedly that the gods could probably stop getting in mortals' way, and that getting the gods to stop getting in mortals' way should be the highest conceivable priority. But it is not acceptable to murder trillions of people because it was too hard to achieve, and so the means by which it is achieved cannot be threatening to murder trillions of people. They will simply have to pursue this incredibly important objective in some other fashion, once Hell is destroyed.

Mortals are creative, and prophecy hasn't been gone for long.  She doesn't think the hindering will succeed for long, let alone for forever.

And if, from a godly angle, it looks to him like the hindering will succeed?  If it looks like only Golarion alone will ever be free, unable to liberate anywhere else?  Or if the ancient gods demand Golarion be destroyed, as the price of other bargains?

Those entities are kind of assholes, Carissa.  These are eventualities to be considered.  They obviously don't want mortals getting uppity and, if it turns out They believe They've got a plan for keeping mortals firmly down, should he disagree with Them about that question-of-fact?  If they don't have a plan, like that, if Golarion survives, why wouldn't they demand some arrangement like the deliberate destruction of Golarion?

Why would they not demand every last scrap that she was willing to concede, if she persuaded him to concede it?

They haven't destroyed Golarion yet, despite prophecy being broken there, despite all the headache it has given them. Presumably because Rovagug is contained within.

Destroying Golarion and letting Rovagug out would be, she thinks, a great victory, for those gods who want to keep mortals down, over those gods that don't - and it's not all of them.

She doesn't think it's Abadar. Abadar kind of jumped at letting the mortals build Civilization, even in the face of what he knows is some risk they'll use to destroy themselves. They can ask Erecura, actually, a bunch of her questions for Erecura relate to god-alliances they can offer and join on ascension.

She just wants Keltham to, when he ascends, look around and see which gods he's allied with, and then be allied with them, instead of trying to unilaterally wreck everything.

Already part of the plan once he's legible to the Lawful ones and can do binding logical deals not to have info he provided be used against him.

He'll have unfortunately needed to do a lot of grim work before then, on his best present plan, like destroying Absalom and releasing Rovagug (both as a distraction, and to preemptively defang Asmodeus by placing Him in a situation where he'll fight on the anti-Rovagug side).  But if it were possible to just show the gods how it would go, and have that be credible, and move straight to the negotiated outcome, none of this would be happening inside reality at all.

She doesn't think she can get him to be sure that there's categorically nothing other than Hell worth destroying the world over.

But the system not being fair, fixing it other ways being extremely hard and uncertain - she thinks that actually, humanity would take that, over a high risk of being destroyed, and anyone trying to act on behalf of the mortals inside Pharasma's system would remember that, overwhelmingly, if not facing Hell, mortals don't want to be annihilated, that when slaves they don't want to be annihilated, when prisoners they don't want to be annihilated, that the route ahead being hard and uncertain and unfair is not a problem they solve by dying about it.

From a godly standpoint there'll just be a probability estimate.  Not "the road ahead being hard and uncertain and unfair", just a probability of victory.  If you don't like your thoughts and feelings collapsed into probability estimates, then don't become a Keeper; and the same presumably applies to becoming a god.

Suppose that the probability of fixing the system looks like, say, 0.01%, because the ancient gods will it not so; and They demand of him that he personally scour clean Golarion with antimatter, in exchange of the end of Hell.  Does she bid him then concede?

Certainly it'd be absurd to destroy the universe rather than do it, what with how destroying the universe has all the downsides of doing it and then some. 

And...if that's the price of ending Hell it's a price worth paying. Even she thinks so. 

Depends on whether you think a god-corralled and herded subpopulation of the human race and the sapient kind, as that greater population exists across all the realities, is a positive or negative contribution to the whole.

And it doesn't have to be the price of destroying Hell.  It could be something the ancient gods demand only because they're assholes and Carissa is titanium-bent on conceding them everything they ask.  Why wouldn't they demand it, if he's predictably going to concede to every demand?

The proposal is that he not destroy any universes which are pretty excellent.

Carissa thinks that the universe, without Hell in it, with no other changes, would be pretty excellent, and constitute mortals - not Keltham, maybe, but normal mortals -- having gotten a great deal relative to their alternative of not getting this life at all. She thinks that if you take Hell out of the picture, mortals mostly live interesting and challenging lives, doing things that matter to them, around people who love them, and the ones who hate it could suicide without worrying about Hell; and Carissa, when she catches them, can do whatever they want done for them including sending them to Greater Reality if they want that. The difference between that and the universe not existing is much much much larger than the difference between that and a better universe.

...theory related to this. Dath ilani don't seem happier than Golarionites, and she suspects that some of that is them having bred themselves wrong; but the other obvious theory is that people are pretty much the same amount of happy under a really wide range of situations, and letting Golarion become like dath ilan in wealth levels would be only a small favor to the people in it.

They could run a lottery, if they wanted to test this: if you win, you get 10,000gp, enough to never work again and live in luxury, and if you lose you get fed to daemons. Carissa predicts that basically no one will take this lottery (and the ones who do will be doing it to save sick children or resurrect damned loved ones, not for selfish reasons) and it's the lottery Keltham proposes entering for them as a collective.

He would want to explain about anthropic survival and Greater Reality; and he's not sure he can, even leaving aside the unshareable evidence-from-his-own-perspective of starting in dath ilan and ending up in Golarion.

Carissa spoke of slaves choosing not to die.  The vast majority of those slaves don't have the guarantee of Axis or Elysium, if they died.  Especially if they suicided, as Pharasma defines to be the 'Evil' of murder.  (Another Evil that isn't remotely bad!  Another way to end up in an Evil afterlife without being a bad person!)

But perhaps Carissa is right as a matter of human-variant psychology, that even a slave with the promise of Elysium would still not want to die.  Would this be coherent reasoning?  And not, for example, the result of natural selection against people who took afterlives seriously, who really believed in Hell and didn't have kids, or who took Heaven seriously and sacrificed themselves altruistically against demons?

He remains skeptical about the proposition that clinging to an awful life is something that Golarionites would be selected to really coherently prefer, rather than them just being selected to not-really-anticipate the promised afterlife.  Assuming, of course, that a slave with the promise of Elysium would not just want to go to Elysium like a sane person.  (Further assuming that Elysium doesn't turn you into a cheerful bright glowing ball, which would be another excellent reason to want to stay alive though miserable.)

Mostly, that should be set aside for later experiment; it's the sort of disagreement they might be able to resolve by using Detect Thoughts on a volunteer, at this intelligence level.

But if the Golarion variant of humanity is bred not to really believe in Hell or Elysium, it's probably also impossible for them to really emotionally believe in anything he says about Greater Reality.  Even if they use 15 wishes and a +6 intelligence headband and an Owl's Wisdom to make the underlying arguments understandable.  He supposes that they could find that somebody with that boost would suddenly say, "Oh, wait, what the ass was I thinking, of course I'd rather go to Elysium than be a slave," and then they might be able to coherently process arguments about Greater Reality; but that experimentaloutcome would still undermine Carissa's point.

Carissa expects that for the most part, for most people, there has never been any serious selection at all around ability to think or not think about afterlives, because the overwhelming majority of people are farmers who know only vague legends about the gods that may or may not cohere to anything in particular. Most people who have ever lived probably have no idea if Hell is real, or if it's really very bad, or if Elysium turns them into a strange alien, or if they'd prefer either to nonexistence. (This is testable: they can go ask people in remote villages someplace). The thing Keltham is postulating, selection, where people who thought clearly about the world and realized they didn't want to live in it had fewer offspring, could only possibly have happened among a tiny segment of the population which had any confident knowledge about the afterlives (high-level spellcasters and people who get trustworthy information from them) and the ability to based on that control whether they had children, which most people can only do through lifelong celibacy.

(She does suspect that humans are selected against finding lifelong celibacy easy.)

Furthermore, coherent reasoning is useful for tons of other things that affect whether your children live to adulthood: planning which crops to plant, how much seed corn to hold in reserve, how to interpret signs of a coming storm, how to decide when to flee and when to stand and fight. There is extremely strong selection in favor of coherent reasoning; her best guess as to why everyone isn't ridiculously good at it is that there are some deep physical tradeoffs.

Regardless, she doesn't think this matters. The process by which people arrived at their current reasoning abilities and attitudes about the afterlives aren't relevant to her. The reasoning abilities are relevant and the attitudes, but not the process.

And she doesn't care if they 'emotionally believe' anything Keltham has to say. If they, in fact, told about Greater Reality and about the existing afterlives, want to go to the existing afterlives, then even if this is not what Keltham thinks they should want, or not what he thinks a different person with different abilities would want, it doesn't matter, because they are not Keltham's slave to dispose of as suits him; they should get the thing they actually want, not the thing that Keltham believes they would want if they were bred for or otherwise predisposed to greater coherence.

It really and truly feels to Carissa like it is profound Evil Keltham is describing here, to destroy someone's soul over your own conviction that, while they prefer Heaven to the distribution of outcomes in Greater Reality you described to them (or reject the premise that experiential-thread is what matters), if they were better at reasoning they'd most likely want something else. She just doesn't think it's all right to do that to people, ever; she thinks you'd kind of have to lack some essential human sensibility that dath ilan perhaps bred out of its own population in order to want to.

There may have been effects on her moral-cognitive reflexes for that she grew up in Cheliax - where every authority and every person who wants to overrule you, for what they claim is your own good, in fact doesn't really care for you much.  Or, insofar as they do care about you, is actively steering you against where you'd want to go if fully informed, i.e., trying to steer you into Lawful Evil.

Someone who grew up that way, might come out of that with a different moral outlook on parentalism, than someone who grew up in dath ilan.

Civilization's prediction markets are basically well-calibrated; Governance has a functioning delegacy that puts nonassholes in charge and could immediately kick out any who turned asshole later.  Dath ilan is nonetheless very conservative about applying parentalism, because it's understood that parentalism incurs long-term costs for short-term gains.  When you prevent people from making their own mistakes, you also prevent people from learning from them... well, of course that's also more of a long-term benefit in dath ilan, where people and societies actually do learn from mistakes.

But if you don't guard the heavy machinery well enough, a kid can wander in and get killed, and then they can't learn from that.  Or rather, the lesson comes in the Future, but that's a little late.

Dath ilan has always staged its degree of parentalism - as measured by the difficulty of the competence tests you need to pass before doing something supposedly harmful - with the goal of giving the smaller minds (like kids) the chance to make their own mistakes; provided that those mistakes aren't going to cause severe irrevocable damage.  That is, Civilization is heavily but not massively conservative in the direction of more letting people make their own mistakes.  Adults with learning disabilities who'll never be able to pass the relevant competence test get stronger guardrails set around them, because they can't learn from their mistakes, and so there's not as much long-term benefit to letting them hurt themselves.

The fact that people in Golarion have probably been implicitly selected to instinctively distrust, hate, and resent authorities, in a world where authorities weren't in fact looking out for their best interests... well, mostly it strikes him as possibly reflecting a trope about unreasonably-difficult+thoughtexperiments:  what if an organism couldn't steer to avoid hurting themselves, couldn't think about the question intelligently, couldn't learn from experience either because the problem was oneshot-else-die or because they were too dumb to generalize the right lessons, and they'd been evolutionarily selected to disbelieve in the possibility of anyone else helping them steer, and to feel angry and awful if you tried to help them anyways?

And there's an old fallback in overly-difficult thoughtexperiments like that one:  Go back to what's actually true.  People can tie themselves up in tangled knots of belief and trust and resentment of parentalism, but afterlives are just real, cryonics is just real, the Future is just a place and so is Greater Reality.  The core idea cutting through all dath ilan's arrangements is this: that when the test shows that you see reality clearly, you are said to be ready to navigate it yourself, even if society doesn't like your choices.

The reason why kids are prevented from killing themselves generally, and truekilling themselves especially, before they pass the respective competence tests to choose either, is not that Civilization thinks it proper for parents to have different interests in children than children have in themselves.  It is that Civilization is reasonably certain that it does in fact know something that the kids don't know, and to be more meta-rational than the kids about which of them is more likely to know it: namely, the value-of-information from staying alive longer to see what unfolds.  When prediction markets become sufficiently certain that somebody would say years later that they should've been allowed to suicide years earlier, that's an automatic license to do so no matter how young you are at the time.

He's sympathetic to the idea that INT 10 people in Golarion should be allowed to hurt themselves, maybe deal themselves lifelong crippling injuries that they can't afford to get healed, if they say to the likes of gods and INT 29 mortals:  "Stay out of my business."  Past-Keltham fell on the Individualistic side of Civilization's moral balances; he wanted Civilization to move more in the direction of people looking out for themselves, rather than looking out for others.

You could make a strong case for letting eight-year-olds get themselves killed that way, once they're old enough not to make the wrong choices in the Boneyard; in the long run that provides the benefit of calibrating future generations to accept the right amount of advice.  Conversely, if you prevent kids from killing themselves or force them to take advice, standard dysgenic mutational pressures may produce increasingly suicidal or intransigent kids.

But he starts to worry when it comes to the question of letting people go to Hell; or for that matter, having children in Cheliax if Cheliax stays the way it is.  At that point it starts to become proper for the Government to step in and do something about it, at least if it's dath ilani Governance rather than Golarion royalty.  That's a kind of mistake that smaller agents really can't recover from, and it's the proper job of larger agencies to guard them parentally until the smaller agents pass the relevant competence tests for having understood what they're getting into.

Past-Keltham thought in dath ilan that people ought to be more selfish, and wished that Civilization were set up to work with more selfish people.  But even he considered there to be an obvious caveat for having systems in place such that children could successfully grow up, and not just, like, get eaten by the first adult who talked them into signing a legal release allowing them to be eaten.

Carissa agrees that letting people go to Hell is taking this principle too far. She thinks that the way you can tell it's taking the principle too far is that generally, once they're in Hell, they don't want to be there anymore. But she doesn't think that applies to people choosing Heaven over Greater Reality, assuming Heaven doesn't mind-edit them to like it once they're there; she doesn't think they'll later regret it, and if they do they can in fact go try out Greater Reality at that point. And Keltham was just hypothesizing people who they told all of the facts known to them about Greater Reality. She thinks that at that point they know the facts and they choose Heaven it's a choice there's no justification for overriding them in.

She agrees that the parentalism of dath ilan is much less bad than the same thing in Cheliax. But she observes all the same that if there were people in dath ilan who were miserable not to have more freedom, not to be allowed to go start their own Civilization, to have had the crucial decisions about their childhood made for them, well, those people wouldn't have kids, and so dath ilan probably does have selection in the direction of being all right with lots and lots of parentalism.

In an absolute sense, yes, there's lots of parentalism in Civilization; and decreasing it is an ongoing problem that lots of Very Serious People worry about, not least because they're explicitly worried about selection pressures for being okay with whatever level of paternalism is allowed to prevail.

In Golarion he observes a notable tendency for city governments and kingdoms to ban books that might lead people into heresy; and not offer those books for sale in Ill-Advised Consumer Goods stores to anyone who's old enough, or any child who passes a requisite competence test about it.  In Golarion, once the city-level or national-level government or maybe just a powerful local church decides you're not allowed to buy a book, or a drug, or in some places a night with a sex worker, that's just The Rule and there's no exception for anyone.

(One could conceivably argue that Golarion's policy is really the same as dath ilan's policy; it's just that nobody in Golarion can pass those competence tests, so nobody in Golarion is allowed to go to an Ill-Advised Consumer Goods store, so they might as well not exist.  However, it seems to him that the competence test of the individual ought to be calibrated to the competence of the government that wants to overrule them, and governments in Golarion couldn't pass those dath ilani competence tests either.)

People in Golarion are being selected to be okay with much greater levels of obnoxious interference in their lives, so far as he can tell.  It's just that the claim of those governments to be acting on behalf of the individual, with all their meddling rules and prohibitions, is not even that believable.  So it's not, at that point, selection pressure about being okay with real actual parentalism from an entity that plausibly does have your best interests at heart.  It's about submitting to a Golarion government that will otherwise torture or execute you...

...but this is getting off-track from the key point, and the rule he knows for cases like that is to review the local stack trace.

A couple of stack layers up, Carissa was putting forth the predicted-but-not-confirmed-observation evidence of "an average slave prefers to stay alive, and is predicted by me to want that even if they were guaranteed to go to Axis or Elysium" (is his paraphrase okay, he wants to check?) as moral input into the question of how bad it is to prevent people from existing in a life that's hard or unfair.  Carissa had also proposed an experiment about explaining to someone about Greater Reality, and then offering them 10,000gp in exchange for a fifty percent chance of being eaten by a daemon under controlled conditions (paraphrase okay?), as moral input into how bad of an individual-choice violation would be involved in destroying Creation.

The problem from his perspective is that if an average slave doesn't choose Elysium, he's not clear on to what extent that implies a general problem with truly believing that Elysium is real, versus a strong coherent preference to stay in Golarion.  Carissa then put forth a claim that it was morally imperative on him not to test whether an individual's preferences were coherent or their beliefs accurate, just obey their stated decisions (paraphrase okay?).  But in the moral system he knows, there's no obvious systematic way to put forth an absolute deontological-rule like that, rather than a defeasible deontological-pressure, without implying that eight-year-olds should be absolutely allowed to take truicidal actions that they don't know are truicidal but that they say they want to do.  Carissa replied that letting people go to Hell is indeed too far, implying it's a deontological-pressure rather than a deontological-rule, and said that Heaven vs Greater Reality should be up to the individual.

He tentatively agrees with leaving Heaven vs Greater Reality to the individual (if Heaven isn't actually gotchaing people into lantern archons, but yeah sure that seems unlikely).  What next step of argument from there?

If he agrees with leaving Heaven versus Greater Reality to the individual then they might just...agree? Carissa doesn't think she is reasoning in deontological-rules, she's not sure how those fit into Carissa-values and they may not at all, she's mostly just thinking in terms of civilizations-she'd-want-to-preemptively-cryopreserve-or-not.

She thinks that as much as she hates it, Keltham is at least maybe-right about Hell, and she's just scared that because Keltham doesn't think anyone else is reasoning clearly he will make tradeoffs between non-Hell places and Greater Reality which aren't, in fact, those they'd make.

Yay for apparent-possible-progress-pending-further-discussion+toward-agreement, then!  But not to overpromise progress, he's worried for trope-based reasons that, after they have a few more days to mentally pursue this debate, they won't end up actually agreeing about the expected value of Greater Reality to people vs. "any living standards higher than literally Hell".

In particular, he's worried for trope-based reasons that the real heart of the disagreement between himself and Carissa will turn out to be that past-Keltham experienced his True Death and continuation elsewhere, and now remembers apparently direct confirmation of isekai-immortality being what truedying people actually experience.  That evidence centers on his self-observation and his update off finding himself to be this self rather than somebody else: noticing himself to be somebody who remembers beginning in a simpler more real world and truedying there and continuing in a less real one.  From the standpoint of somebody else watching past-Keltham get copied out of dath ilan, they haven't made that same observation, and it's fundamentally unshareable with them.

The famous central difficulty of writing a romance novel is finding a conflict that can't be resolved immediately through clear communication and emotional maturity; nobody wants to read a romance novel about silly people ignoring obvious solutions.  Anthropically unshareable evidence is one way of providing a romantic obstacle like that.

It would not have occurred to Carissa to complain that among the problems with their relationship were that they didn't have enough genuine non-manufactured conflict.

In fact she can't think of a single romance she's ever encountered in which that was the problem! Any two people naturally have irreconciliable differences as a product of the one of them selfishly optimizing for their own interests and the other selfishly optimizing for their own - okay now that she completes that sentence she can see how in dath ilan that would not describe the plot of many romance novels.

But the unshared evidence thing does seem like it might arise. Carissa does not really think that in an important sense the thing that happens, if you truedie is you waking up somewhere else; someone might make a copy of you somewhere else, which is good of them, and if you truedie enough places then on that planet, the copy might have the experience of truedying and finding itself on a planet that likes making copies of dead people from elsewhere. And she grants that she ought to care about exact copies of herself as much as about herself. But the intuitions from instead-you-wake-in-a-nearby-universe-where-you-didn't-truedie and instead-your-experiential-thread-usually-dwindles-slowly-enough-there's-nothing-to-continue suggest to her that she doesn't think 'you wake up in Greater Reality when you die' is the right way to think about it, even if it or something like it did happen to Keltham this time from his perspective.

Still viewing the conversational stack from a step back, they shouldn't be trying to fully-resolve this now unless it looks like it can be fully-resolved quickly.  The important thing in this conversation is to get an overview of the things that might form an obstacle to cooperation and kept oaths between them, going forwards.

The critical question here is: what happens if the plot successfully places the two of them in an epistemic Cooperation-Defection Dilemma, where rather than the two of them just having different goals, Carissa believes that he is mistaken about what happens when people greatly diminish in reality / get deleted from the universes that contained most of their instantiatedness; and conversely he thinks Carissa is mistaken about him being mistaken.  (This is the thing that's hard to do to a romance between INT 29 people who know about agreement theorems (as she soon will), unless the plot throws anthropics into it or something equally overclever; which is why he worries about overclever tropes.)

In this case, Carissa could end up believing that to play 'Defect' against him would be to serve even his own goals, better than her Cooperating would serve them.  Betraying him might seem like a friendly act, an act of aid.

(The classic presentation of this situation to dath ilani children supposes two 12-year-olds on an island otherwise full of 9-year-olds, and they're all facing a contagious disease:

The two 12-year-olds disagree about which disease it probably is.

The 9-year-olds collectively have enough resource that they can take thorough precautions against one disease, or 80%-effective sloppy precautions against both diseases.

Either 12-year-old could sabotage the other by waiting for them both to present their case, and then telling the 9-year-olds all the reasons why they're sure the other 12-year-old is wrong, in which case the 9-year-olds will take 99%-effective precautions against that disease only.

If both 12-year-olds Defect in that way, the 9-year-olds will be dispirited and only take 20%-effortful 50%-effective precautions against both diseases.

These two agents both have the same altruistic goal - to save as many of the people on the island as possible - but their believed Cooperation-Defection payoff matrix has the classic ordering of the Dilemma:  DC > CC > DD > CD.)

Carissa is, actually, pretty sure at this point that she is going to be a Lawful god; she wishes she had retained the option to betray Keltham to Otolmens, but she does not actually intend to do so, having not retained that option. That logic does not, to her, feel specific to whether she would be betraying Keltham for his own good or for the good of every other person in the world.

Yeah, he figured, but wanted to check explicitly.

Okay to shift discussion to continue a previous topic up the stack?

Yes. And she noticed he asked and will try to do the same if she's changing topics, though on the set of norms more familiar to her you just change topics and the other person can object if they object.

That sure is an impressively low amount of meta!  He's not sure he wants to go that low on meta.  Their conversation has been sorta uncontrolled, not really by his past-self's standards for arguing something complicated with a friend over a meal, but definitely by the standards of Very Serious People in dath ilan trying to use two people's intelligence in a coordinated way.  Though his guess made at the start of the mental meeting, and still held by him now, is that their time is still mostly better spent on mostly not-meta, due to their lack of mutually established protocols for meta, which is why his current meta-meta strategy is to do only very brief metas like that one he just did.  But you can still, like, briefly meta-think about whether it's okay to change topics or if the other person was still in the middle of thinking.