I hold a lot of these same expectations. Very well written!

 I think it won't last forever, mostly because people won't be human-singleton-minds forever. But for a while, possibly a very long while, this is probably how things will play out.

I don't know what portion will be super persuaded into bad beliefs and what portion will be isolationist and what portion will be enlightened in our future. I'm okay with some limited extreme beliefs surviving. I hope it's similar or better than today... But it won't be, necessarily. 

Thanks for writing this.

(Lady of Mazes is a scifi book that describes how you might have to shift your beliefs/mind in order to communicate and move between communities and that this is a difficult thing to do.)

Hi, FelixG I really enjoyed this post!
I have a poor understanding of this topic. Most of all I wish I had a clear idea of what the biggest danger is from water mismanagement. I can't tell if it's pollution, aquifer drainage, mismatched timing, wrong location, or something else. 

I once tried to take a look at existential risks from lost ecosystem services and wrote this brief summary against it qualifying as an existential risk (I know you specifically are NOT claiming existential risk, but I think the commentary is still useful grounding): 

Environmental activists make it sound like we are running out of freshwater. But only 2% of watersheds use 75% of their freshwater. Data centers are not draining our water supply. To be clear, we are choosing to pump truly preposterous amounts of water out of aquifers for agriculture, and this is unsustainable and bad for the environment. But we could be far less wasteful, recycling and desalinating way more. Surface water is returned annually so we aren’t going to “run out.”

During my mini foray I was pretty shocked at how much water is being drained from aquifers. I wish I could spend some time thinking more deeply about the threats from water misuse. 

As a general rule, I have found resource depletion threats tends to sound ominous, but turn out to be avoidable upon closer inspection. Resource depletion is usually hundreds of years out. And there are usually moderately more expensive alternatives which we would switch to, once we started to experience scarcity. 

I would want to check on that for water bankruptcy: When would we run out. Where. Is this a new problem, or a fact of the environment. What viable solutions are there, if any. 

"...such system could collapse and lead to numerous cascading bad consequences. At least, to my understanding."

I would like to see some specific examples of a water-crisis-cascade, to get an idea of how it has manifested so far.^[1] Follow by some estimates of the maximum scale it could theoretically reach. 

I just did a research report on vultures as an ecosystem service and got some high numbers of irreversible human health damage. So its very vivid to me right now that environmental damage can be huge, and far cheaper to avoid than to try to reverse. I could definitely see aquifer drainage as being a case of eating the seed corn and heading for worsening disasters. But I could also see it as having simple solutions and being hundreds of years away from becoming an issue, in most places.
 

1. ^{^}

   (When I tried to do this for ecological collapse, I was shocked I couldn't find any real-world examples of cascading collapse. Which made me update that it's a much less likely source of unpredictable devastation.)

I'm reminded of this rather nice infographic:
https://forum.effectivealtruism.org/posts/pxEHC6uBzsXRdWqax/i-made-an-infographic-for-the-introduction-to-effective

Hi! From the title I thought you were offering advice. You might want to add something like "looking for/help requested" so people know to click and suggest ideas and recommend connections

Thank you! You are right of course that I didn't include long term resilience. I was just trying to start with the most basic first-approximation "baseline minimum" assuming perfect knowledge and handpicked extinctions. 

Resilience is difficult, because it is a noisy gradient, and ultimately a choice where to pin the "right amount" of stability. It co-depends on our fluctuating ability to cope with the 100-year extremes. I'm not accounting for stochastic events, tipping points, natural disasters, or 95% survival rates on 50-1000 year timelines. I'm also missing an unknown number of ecosystem services, and didn't include supporting ecosystem services(!).

Which is why the numbers I ran would realistically be higher. I'm not sure how much higher. 10%? 50%? 2x? 10x? Even at 10x though, its still around 10-20% of species.^[1] 

So, to answer your question, how much changes when we include long term resilience/insurance/response diversity?

From what I've seen of the biodiversity x ecosystem-service-benefits, it caps out surprisingly fast. I agree that different traits result in different performance under different shocks, and that it is important to system stability. But again, I'm guessing stability caps out surprisingly fast. 

Yeah. It looks like stability rises approximately linearly with log richness. You probably get most of the gains in the first handful of species. The gist is that if you halve the species, we keep 75% of the stability. 25% the species and we keep 50% of the stability.... Translating this to "required biodiversity for ecosystem services," we should never have one species providing an ecosystem service. Any service should have a minimum 3-5 providers. For the rest, I can't speculate what level of stability is needed. I vaguely feel losing most stability is probably acceptable, but this really demands calculating out the costs incurred and the acceptable rate of catastrophic failure. (intact ecosystems also have rates of catastrophic failure, so the increase in rate...)

Sorry to continue to be a nag, but managing ecosystems might be an easier way to increase insurance than species numbers. 

Maybe I'm overfixating on gloomy estimates to be "rational." I would love someone to show me some concrete numbers that challenge this picture.

From your first link: Within-species portfolios accounts for ~2× stability?! Huh, Des Roches et al. 2018, Raffard et al. 2019 compare intraspecific-variation effects against species-replacement effects on ecosystem functions, and within-species effects show similar magnitudes! This seems to indicate that stabilization comes from decorrelated units. Of any kind. Populations, age classes, genotypes, species, anything that delivers varying responses... Which should move some of our concern to the ongoing loss of species genetic diversity (and maybe habitat heterogeneity, shared susceptibility to natural disasters, disease, etc?)

I appreciate the pushback!

2a,b,c) Agree, people discount the long-term benefits and positive externalities of ecosystem services quite a lot!

2d) Totally, agree with this and I wasn't adequately considering it. Its usually costly and uncharted how to reverse the loss of ecosystems. Infrastructure is a sunk cost and requires a lot to remove, and we acclimate to how things currently operate, so we wouldn't just move back even when its strictly better. 

2f) Option value: I agree. But I think it can only account for maybe 4x the species we currently value?

3.) Ah, sorry, I don't expect to "harvest everything there is to know" about a species anytime soon. This is more about a hypothetical distant future to acknowledge that, technically, knowledge is finite and not identical to the species. This is not relevant for hundreds of years. After hundreds of years I expect we will be able to simulate/understand adaptations to climate before they happen. Like how we can semi-predict how the flu will have mutations each year, and are trying to accurately estimate the climate phenotypic plasticity of species today. We will eventually map out all the possible mutations, their likelihood, their impact on survival, and all that will remain to discover is which course history actually takes. But this is just about the far future, where we are able to easily create entire species from scratch, and things are very different. In between then and now, wild populations have lots to teach us about adaptation, behavior, climate, biochemistry, nature, ecosystems, umwelt, and imo more than we can name.

1. ^{^}

   I came into this expecting to show it would be around 30% of species, and ended up calculating 1%. :( I'm not confident about any of this. I'm trying to get my bearings. It's lower than I expected. It seems clearly less than 50% of all species. It seems clearly higher than 0.2% Probably around 1-10% That seems really low. I'm more comfortable assuming 20-30% Augh.

An analogy from climate/environmentalists: development and carbon emissions are "bad" for the environment. But I don't worry about the "bad stuff" that companies are doing. They are powering cities, inventing things, distributing goods, and generally pushing progress and making the world a better place on the dimensions they are most paying attention to. That's their job. It's my job to care about the bit of the world that I know about that needs fixing. So I am fine supporting the good they are doing and then spending my time combatting the bad side effects they aren't aware of/the world needs more attention on. Roughly same principle with AI. Their job to do good AI things, your job is to fix bad AI things they aren't doing enough for. 

Things that are different for AI: AI companies might be net bad instead of net good. AI is in a different tier of urgency and xrisk.

So how does that change the calculus from the previous picture? (It might also help to think about a middle tier case like weapons manufacturing.)

Morality doesn't exist. Welfare isn't an objective weight. These things are fictional.

Responding to your meaning, I am bothered you are right: experiences are the only things that should have moral weight. I want to agree. But I hit some problems I can't resolve easily:

Pursuing the world that is nothing but tiny single purpose happiness circuits, or pond scum only experiencing blissful happiness and nothing else, is an end state I cannot accept being the best outcome. The problem is things like consciousness, variety, truth, perception, concepts, are at best extraneous and get in the way of pumping out maximum utils/welfare. Delete everything -> only welfare.

The way out, that I'm aware of, is either to add "quality of happiness" and suppose complex happiness is stronger than pure simplistic happiness. Or suppose some kind of saturation of similar kinds of happiness. Or that variety is some additional value. 

How do you grapple with it?

Excellent job describing this clearly and accessibly. (And teaching me about Saturationism) You portray this well from where I stand on the opposite shore!

I think that truth and variety matter and would prefer to live in a universe with more truth and more variety even at the cost of "loss of perfection" or suffering^[1]. Even if I could not tell the difference (unaware of deceptions) or never experienced any positive effects (from variety). I personally value these things, despite them not simplifying to additional positive valence. 

Here are some values I have that don't pan out to positive welfare:
1) I think if someone dedicates their life to some endeavor, and they are feeling the rush of emotions as their life's project comes to an end, it would be "bad" if those exact same emotions/experience were via beaming that feeling into their mind. They cared about the project itself, they wanted it to be due to their own effort, they are not doing it for the pleasurable experience portion. 
2) I think it is bad if we, with well meaning, erase^[2] everyone's memory to replay their perfect happiest possible moment over and over again for the rest of their lifetimes. I think, even if nothing could be a better world of experience for everyone, and no one will ever know it's a lie, that this is a flawed^[3] life and a flawed world.
3) ^[4]Someone who changes the future but doesn't live to see it has, in some sense, a better life than someone who tried and failed, even though their felt experience is identical.
4) ^[4]I think it would be a shame if many things are never understood or discovered before the universe ends, regardless of how this translates to total happiness experienced. 
5) I think an empty universe is bad in comparison to what it could have been, even if no one exists to prefer something else.

You cannot imagine an unwatched world without mentally watching it
Yes, but I think that if you delete the outside view then lots of things are incoherent and worse. 
The badness of death disappears. We know that being dead is bad because of the outside view  where they could be happily living. Same for unborn people. Accounting for past and future experiences is using an outside view, no one is actively having those experiences. Similarly, if we forcibly rewrite everyone's desires to be something terrible, (like to desire everyone dies horribly, and then we all happily satisfy this!) this only is bad from outside. 

The summing view makes arbitrary outside view choices too. Why sum? Why not only care about the single happiest life ever lived? No one experiences the sum, the lives that didn't happen, or the future. You are asking the accountant to look at total welfare achieved out of potential welfare possible, not just total welfare and without judgement on worlds which could have existed instead. There is nothing to evaluate, without an outside view to compare against. Choosing to "add them up" instead of "average them" or "prioritize the worst-off" is a world-shape preference the accountant imposes, not something anyone experiences. You want to say the Totalist's account is fundamental while the Saturationist's account is doing something superficial. But the Totalist is just as arbitrarily choosing how and what to aggregate when he's adding up experiences, and using an outside view to do it. Yes, variety across experiences is an outside view parameter, but that isn't unusual.

Its also a little weird to assume positive welfare "exists" and "can be summed." Maybe when you "sum" it, the redundant identical feelings overlap, occupying the space of one single unique experience, and summing it results in the Saturationist's account. Maybe that's just how adding up felt experiences works.

Edit: The next sections I think I'm being uncharitable/shuttered in some way, but I'm not sure how.

No one proposes an ugly impersonal value.
No, variety and truth are very much an ugly impersonal values. These could have very high costs and often result in a lot of suffering. All kinds of stories celebrate people who wrecked their lives forcing truth or justice into the world, or chased perfection/uniqueness (even when it doesn't work!), which is strange behavior if these values are just about netting more pleasurable experiences. Their value did NOT evaporate when it had an net-negative-welfare cost. Sometimes we admire it as a life more well-lived, in some non-welfare sense. They had purpose instead of happiness, or something. We often hold these values dear albeit with conflicted uncertainty about what is best and right and matters most when it had such a terrible downside. 
We don't celebrate the people that are happiest and heroically help others bury their head in the sand to protect their happiness. Which would indicate that happiness is not the single final value...

Ask A whether she wants to switch worlds: she is indifferent, and she must be, because if she preferred one, the welfare would not be equal. Ask B, same answer, because nothing changes for him.
Turning around and sating that "this must mean the welfare wasn't equal" is hard to respond to. Maybe its high quality welfare without being higher quantity? People do things for for results that they will never personally experience all the time. Its weird to say those values are fake.

I think there are additional fundamental values beyond welfare, like variety and truth. Unfortunately I'm not sure if they can be explained further than "I value them..." so I may have to just describe them and then stop.

I think I might be motte and bailey-ing. I think I'd probably retreat to variety greatly increasing welfare/valence, and be willing to confine this to only cases when it resulted in actual felt experiences at some point down the line. I may just think experiences are much greater relative to your position, when they are complex and varied experiences. We do have to come back to actual beings with actual values, (probably also being actually experienced). 

I'm not so willing to accept that identical copied lives add just as much value as varied equally valenced lives.

Anyway, all this results in this belief that I have:
If we reach the world where everyone has the most perfect life, and it's the same life, I think this world is worse than many non-identical, not-quite-perfect lives. I wouldn't force worse lives on anyone, but I'd volunteer to live a worse life so that the universe contains something different. Normally sharing my different life would benefit others; but even if no one benefited at all and I wasted my perfection, I think I might still want to do it.

I'm open to the possibility that I'm a limited idiot who can't see what the god-minds of the future will want. In my defense, I'm not hopelessly unimaginative. I fully expect that we will want to do currently-crazy-sounding things likes erase ourselves, limit/reshape our minds, replay experiences on loop, merge into amorphous monads, do tedious activities in deep utopia, and tons of other unimaginable things that might seem undesirable, even hideous, to me today...) The god-minds might all agree that tiling the universe with hedonium is obviously correct. Maybe I've just intrinsified some stuff that is a mere side-effect of promoting progress or weeding out traitors or something that will no longer apply in the grand scheme of things. But from my best attempts at imagining the god-minds at the end of the universe, I think they would value variety and truth even at the expense of some lost hedonium.

1. ^{^}

   I would also not maximize truth and variety at all costs.

2. ^{^}

   assuming they didn't agree to it. I expect we will choose to do this to ourselves quite a lot actually.

3. ^{^}

   Maybe not a bad life compared to today's good lives, etc. It's an alright future. But I think its far worse than many other life-shapes and potential worlds that include variety.

^{^}

I'm not entirely sure I hold this actually.

Show all footnotes

What is, "Rutger Bregman"? disagree votes made me google...

From this summary I don't know what the book's arguments are, nor what original pro-colonization arguments it is responding to. (x-risk?)

Interesting topic though, and sounds like a good book from what I can tell. Thanks for the alert and invitation to read it.

 Is there a "right size" for space colonization, (one solar system, a few nearby stars, one galaxy)? Or is it basically inevitable that any space colonization is going to lead to infinite expansion?