[geoalchimista]

Not sure I understand the reason this kind of model keeps getting spawned every few years. Idealized models with no relevance to empirical observations are nowhere near predicting civilization collapse. They can't even explain the observed past trajectories.

[cryptonector]

Right, because all these models are just that: models, and very simple ones at that. They encode a few functions, often linear, when the thing they model are enormously complex and dynamic. Humans generally fail to grasp the non-linear and the dynamic.

The Malthusian model is: more food supply -> more humans -> more food demand -> more food supply -> more humans -> ... A vicious circle. A positive feedback loop. But it's a positive feedback loop operating in machine with negative feedback loops too. It's been over a decade since it's been understood that global human population is going to plateau in just three more decades, then begin declining. Reasons for that are myriad, but if I had to summarize it it would look like this: low child mortality + high life expectancy + high standards of living + high taxes and costs + high retirement costs == low interest in reproduction -- i.e., price signalling works! Who would have thunk it? Not Malthusian modellers, for sure.

[rsj_hn]

I remember the Club of Rome was the first to do these models, assuming peak oil and the collapse of western civilization based on resource constraints. They would have these coupled differential equations, similar to predator-prey, calculating cost of extracting resources and the resulting declines in food production and thus population.

https://en.wikipedia.org/wiki/Club_of_Rome

But I thought all of that had fallen out of fashion by the early 1980s when oil prices fell and it was discovered that you don't have stable relationships or known parameters for most of this stuff.

[cryptonector]

Exactly. You can have very sophisticated, non-linear, dynamic models developed over many years by brilliant people, then still see the models fail miserably.

[dredmorbius]

World3, implemented in the Python module submitted here, is the Club of Rome model.

[mdp2021]

('thunk': «from an 1876 glossary of words in the mid-Yorkshire dialect in Britain», presumably jocular.

Used by Joyce: “I thunk I told you” (Finnegans Wake); “Have a good old thunk.” (Ulysses)

Does not come, as one may suppose, from a jocular integration from the area of the onomatopeic "thunk", which is 1952. Interesting. Considered in grammarphobia.com )

[cryptonector]

It's a purposeful misspelling of "thought". It's fun to use it once in a while.

[coolspot]

> high standards of living + high taxes and costs + high retirement costs

Add unsustainable mining/fishing/farming/pollution to feed/dress/entertain all those people with high standards of living.

Now you will see the plateu eventually, yes. But it’s not a natural plateu where we stop because we want chill life, but plateu where we stop because there are no resources (everything is expensive) to support our kids.

[nefitty]

You're not taking into account technological and industrial innovation. Presumably there will be new ideas iterated upon that get us closer to universal abundance.

My main concerns over the next few decades are cyberwar collapse (< 1% chance), runaway AI (10% chance), robotized world war (30% chance), and solar flare induced societal collapse (40% chance).

Yeah, cereal is 20% more expensive but DeepMind just beat GPT3.

[RealityVoid]

I wonder why solar flare societal collapse ranks so high on your list.

[kortex]

Not GP but it's a cascade failure mode with theoretically no upper bound in how much chaos it could cause. A toilet paper shortage caused millions to lose their goddamn minds, what do you think having no refrigeration on a world scale for even a week would do? A CME could potentially fry enough equipment that we would be left in in the dark and unable to repair for months, just due to availability of parts. I believe those estimates are at current production rates, too. If the supply chain and global comms are compromised, it's even harder.

It's definitely at the sweet spot of "terrifying" and "completely plausible."

[nefitty]

Yes, perfectly said.

Here's in-depth info on solar flares. Major ones are considered "Carrington-level events", due to a flare that occurred in 1859: https://www.history.com/news/a-perfect-solar-superstorm-the-...

Vox's 2014 breakdown of the systemic risks: https://www.vox.com/platform/amp/2014/7/30/5951263/a-catastr...

NASA article on the prediction that there was a 12% chance a Carrington-level event would occur by 2022: https://science.nasa.gov/science-news/science-at-nasa/2014/2...

My 40% is based on that information. I'm gut-extrapolating to the next few decades, with the understanding that large devastating flares occur every few hundred years.

[xsarin]

Society is so incredibly dependent on technology that is not EMP resistant and if a solar flare knocks out most of our electronics the bottom is going to fall out. No phones or internet to figure out what's going on, no cars since the onboard electronics are needed- same with trucks, trains etc. Can't go to the grocery store and buy food with your debit card, can't get cash from the bank, grocery store isn't going to get shipments anymore etc.

Our dependence on it all puts us on really fragile ground given that we know there have been sun events in the last 150 years that would obliterate everything we have today

[RealityVoid]

I have read that the dangers of EMP's are greatly overstated and they could damage stuff, but mostly the power grid, since you need long lines to induce sufficient power for it to matter.

[baq]

resources are finite and technological innovation is finite, too. thermodynamics gives us very precise upper limits to efficiency increases and to the surprise of no one, low hanging fruit are all picked.

[jstanley]

You could have said the same thing 100 years ago.

[CuriouslyC]

You're making the same mistake all the star eyed futurists make, which is to not realize that in reality, all exponential curves are just early/mid stage sigmoidal curves. You can paradigm shift all you want, that's not going to make infinite progression a thing.

[baq]

a single example: a 100 years ago the efficiency of ICEs was less than 50% of today's. we're left with maybe 2x until physics doesn't allow us any further improvements.

another example - transistors - we've got maybe 10, maybe 20, maybe 30 years of improvements ahead of us, after that there are fundamental limits that forbid progress.

the rate of change of total technology improvements will slow down and at some point will start to approach zero; maybe even go negative as we as a civilization start to forget how to do things faster than invent new ways of doing things.

[cryptonector]

And people did. Malthus was over 200 years ago.

[cryptonector]

Yes, but you're assuming indefinite growth. Again, in the real world we're already on a path to decreasing population.

[cnlevy]

Sure, African agriculture is already at max. efficency. Giant farms fully mechanized, of course /s

[snapcaster]

Kind of unrelated but if you're interested in African agriculture you should look into Seeing Like a State. The book argues mechanized farms aren't necessarily the highest form of farming for maximum sustainable yield in many places (like many regions in Africa) and has many many hard to notice downsides. A lot of the pressure to consolidate and mechanize has to do with legibility for taxation purposes compared to true yield maximization

[nefitty]

Great point. I'm starting to look at Africa for that low hanging fruit.

[spodek]

Because the data keeps matching their Business as Usual model. I think there's more recent research, but here's a paper from 2014, showing 40 years of reasonable fit: http://sustainable.unimelb.edu.au/sites/default/files/docs/M.... If we keep matching their model, we'll soon see dramatic population declines.

Also, the patterns revealed in their models are useful for understanding patterns in nature and human interaction with it.

[geoalchimista]

Just looking at Figure 1 on page 8 of the paper you mentioned, it can go either way because we haven't seen a reversal in any of the trends. That is hardly a prediction. For example, if one looks at the "Services per capita", is there any sign that it will meet an inflection point in the next few years?

I know it is always appealing to tell a story with a grand unifying narrative. But sound research must prop it up with empirical evidence. Could there be limits to growth? Probably. But a highly simplistic model not informed by appropriate data or economic understanding is not the way to tell such limits.

[baq]

> Could there be limits to growth? Probably.

that's what an economist would answer. ask a physicist.

for best results, get an economist and a physicist in the same room and ask them both.

[api]

There have to be limits to growth, but we won’t know where they are except in retrospect. We also may never hit them due to fertility decline which occurs in all wealthy societies. Fertility decline plus efficiency increase could cause consumption of some resources to fall.

Space migration doesn’t change the equation much on Earth unless you start mining and manufacturing off world and importing product, and that is pretty far off.

[teabee89]

"Could there be limits to growth?" Yes 100% (you cannot extract indefinitely more and more every year from finite resources), the debate is simply when we'll peak for a given ressource (20 years, 100 years, 1 million years) and the size of such an impact on the economy.

[AlotOfReading]

It's worth noting that world3 model is fairly sensitive to initial conditions and small perturbations can yield population predictions exceeding 20 billion or under 1 billion [0]. It's pretty unclear whether the patterns are actually useful or actually "natural" rather than artificial. For instance, virtually all scenarios in world3 end up in a peak->collapse, but it strains credulity to imagine that similarly applies in the real world.

[0] https://doi.org/10.1287/mnsc.44.6.820

[actually_a_dog]

> For instance, virtually all scenarios in world3 end up in a peak->collapse, but it strains credulity to imagine that similarly applies in the real world.

Why? Many civilizations have come and gone through the collapse cycle already. Why do you suspect we're any different? What makes you think the ecosystem can even tolerate 9-10 billion of us?

[AlotOfReading]

If you pick any particular definition of "collapse", very few will actually meet it (let alone on a timescale of mere years or decades). There's quite a lot of literature on this subject already (e.g. Questioning Collapse), but there's also tens of thousands of years of human history apparently lacking records of anything that could be called a collapse.

As for a specific carrying capacity for the Earth, it should be obvious how impossible such a number is to give without a lot more detail in the question. But if we were to assume 10B global population, it could be done with a population density roughly equivalent to precolumbian California. This is not to suggest that indigenous californians lived in perfect natural harmony, but rather to illustrate how low the numbers actually are. I suspect there's probably many reasonable (though utterly alien) ways of life where that density could be "sustainable". Equally, I suspect there are many ways of life where those numbers are not "sustainable".

[jacquesm]

The eocsystem we were born into as humanity would not sustain even a small fraction of the humans alive today. We modify the ecosystem until it will sustain us (or 'just us', for that matter).

[actually_a_dog]

What makes you think we can continue to do that? Even if we can, what makes you think we can do it sustainably? And if we can do it sustainably, what makes you think we will? History doesn't really provide any encouragement here.

[jacquesm]

I don't think we will be able to do that, sustainably or otherwise. We're locusts, pure and simple.

But we got away with it long enough to a large number of people now believe that this is normal. It isn't. The wake up call will be a very harsh one.

[perfect_kiss]

https://dash.harvard.edu/bitstream/handle/1/37364868/BRANDER...

Just in case, here is another empirical data comparison for world3 predictions, this time from 2020.

[actually_a_dog]

We're already starting to see evidence for the beginnings of global population collapse: https://www.weforum.org/agenda/2021/06/birthrates-declining-...

[nl]

Slowing birthrates is very far from a population collapse. It's also (to generalise vastly) desirable - slower population growth is desirable environmentally.

[actually_a_dog]

This is why I said "starting to." You did notice those two very important words, didn't you? Because this is how it starts.

I agree that decreasing population is desirable from an environmental POV, to the extent that population is more or less directly proportional to consumption. But, capitalism as we know it can't survive it. Here's a good explanation: https://www.axios.com/the-new-threat-to-capitalism-73ff54bd-...

[nl]

> This is why I said "starting to." You did notice those two very important words, didn't you? Because this is how it starts.

"starting to" is a prerequisite, but a reduction in birth rate doesn't necessarily lead to a population collapse, and you don't seem to present any evidence that this is one of the cases where it does.

[actually_a_dog]

What evidence would you consider sufficient, besides globally falling birthrates?

Again, I will remind you, the comment says we are "starting to see evidence," of population collapse, not that population collapse is happening or is inevitable. Globally declining birth rates is certainly evidence that it may be happening.

[jnurmine]

During the 1970s the Club of Rome made "The Limits to Growth" with the World3 model and, whatever it predicts, some scenarios like the Business As Usual (BAU) seems to have tracked quite well for ca. 50 years already.

The latest check-up is by Gaya Branderhorst in 2020, https://dash.harvard.edu/handle/1/37364868.

Of course, one has to understand that such a model doesn't show you "on the year 2028 the population is size 16.2931 billion". That is not the point. The point is to show how something grows, or shrinks, and the interrelations of the tracked variables. As x goes higher, y starts to diminish and soon after z goes through the roof, that kind of thing.

The collapse of, for example, industrial output levels might be a form of a Seneca cliff. That is, the growth of the industrial output becomes harder and harder to achieve, while the difficulties compound and actually keep on growing long after the growth of industrial output itself has stagnated or started to drop slightly. This will dramatically increase the inhibitory effect, thus creating a growth curve which starts to slow down and then drops fast ("collapse").

Why dismiss the model? The BAU scenario seems to reflect reality for over 50 years now, so is it not rational to try to understand it?

The Limits to Growth is fundamentally a realization that the world has finite resources. It is not illogical to consider this as a truth. At some point there won't be new resources to obtain, and hence growth in systems based on capturing resources will simply slow down (and perhaps even drop fast, "collapse", because difficulties compound). Circular economy and recycling are good but they cannot expand the available resource pool of e.g. raw materials.

[jshmrsn]

I assume these are references to Isaac Asimov’s Foundation.

[shoo]

the references to have a read through are not science fiction, but are system modelling studies done in the 70s:

> Meadows, Dennis L., William W. Behrens, Donella H. Meadows, Roger F. Naill, Jørgen Randers, and Erich Zahn. Dynamics of Growth in a Finite World. Cambridge, MA: Wright-Allen Press, 1974.

> Meadows, Donella H., Dennis L. Meadows, Jorgen Randers, and William W. Behrens. The Limits to Growth. New York 102, no. 1972 (1972): 27.

there's lots of people who dismiss this as "doom and gloom". it's worth getting hold of the books and reading through it, and making up your own mind. do you think the modelling assumptions seem reasonable? even if some of the parameters seem difficult to estimate from observed real world data, do you reckon the overall system dynamics behaviour of predicted "overshoot and collapse" seems plausible, even if the timing may be very difficult to predict?

[rileyphone]

I'm reading Donella Meadows' Thinking in Systems right now which I think is a good overview of this line of reasoning. Of course models are idealized, but they're also one of the best ways to understand complex systems. The human brain does something very similar in forming intuitions about the world, but we should recognize that in order to make a better model we will have to make one in silico. Another interesting approach is agent-based modeling, which explores how the decisions of agents with limited information can create emergent patterns.

[canjobear]

Has their model made accurate predictions on held-out data? That's all that matters.

[dredmorbius]

For those curious about the term "held out data", it's a modelling / validation method.

https://people.duke.edu/~rnau/three.htm

[jacquesm]

Wrong domain.

[dredmorbius]

How so?

I'm presuming that held out data is used in gradient descent machine learning artificial intelligence, and isn't applicable to systems theory models.

And that other options, including backcasting or parameter adjustment (as was done with different scenarios tested under World3) are.

[onethought]

Ability to predict a past future doesn’t give any strength to the ability to predict the future future.

[betwixthewires]

I'm guessing you're using an analog to "past performance doesn't predict future performance" which is true.

However, seeing as the default likelihood of predicting any future with certainty is zero, a track record of a model of being entirely unable to predict the future at all does very much indicate that such a model will continue to fail, more often than not.

Predicting the future based on what has happened is not reliable, but predicting the future based on models that have predicted the future before should be reliable, if the models themselves are based in reality.

[onethought]

how did everyone’s model of 2019-2021 pan out on any given metric? Even if they successfully predicted that same metric correctly for the last 100 years.

I guess weather/climate models stay somewhat consistent. But they also don’t predict very far forward.

[bryanrasmussen]

Ok so assuming you had a system that you could give it a bunch of inputs and it would predict what the results from those would be 10 years out, and it was correct 60% of the time from past observations, and you had run it on thousands of simulations (this is obviously a hypothetical) is it then your contention that if you gave it inputs from our current inputs and had it predict results ten years from now that those predictions would not have a 60% chance of being correct?

I mean sure, since there is such a thing as an inductive gap in theory, but it seems to me from your statement that you think it to be an insurmountable obstacle in practice as well?

[cryptonector]

It's obvious enough. Conditions change. Systems are dynamic. A model of human society that "works" for the next 10 years may not work for the subsequent 10 years.

[baq]

and yet we model complex stuff with some success.

[onethought]

Like how many computer chips we’ll need during a pandemic. (Sorry couldn’t help myself)