Unpredictable Patterns #100: Invariances

Predicting 2025 by looking at what will stay the same.

Dear reader, we are back. After a two year hiatus, I find that I would like to start writing more, and this letter provides a great opportunity to learn, think and write about things that I want to understand better. There are likely to be changes to the letter, its periodicity and more - I have some ideas - but we will let them emerge over the course of the year and see where we end up. As always: comment, let me know what you think would be interesting to hear more about and tell me when I am, inevitably, wrong.

Invariances are underestimated in future studies. Our focus is often drawn to what changes¹, rather than what stays the same - but what stays the same sets the limits and framework of what might actually change. By looking at constraints and invariances, we get a sense for what will not happen — and that is valuable in itself.² Furthermore, if we identify something as an invariance and it does change, well, then we know that something we believed was fundamental turned out not to be and that is a great signal for when to update our beliefs.

What is the right horizon for studying invariances? Should we speak about what is not likely to happen next year, or should we expand the horizon to, say, five years? Ten? The longer you can make your horizons when studying the variables that will stay stable in a model - the better it is. One way to deal with this is to try to look at the probability curve of an invariance over time, and guess at what rate it declines. I might say that I think that X will hold the same over ten years with 90 percent probability, and after that I might allow for a rapid or gradual drop in probability as the decades pile up.

So let’s look at a few different invariances, and predict for how long they will stay the same. As always, with predictions, all the usual caveats apply — but look at this as an example of how to approach predicting invariances; it is naturally always best if you predict your own (and add them in the comments).

China’s population will continue to shrink for the coming decade. Demographics is always interesting (remember - demography is destiny) and we know that China’s population has started to shrink, and once that happens it is really hard to reverse the trend. This is an interesting invariance - a trend that continues - and it is likely that it will continue for the coming decade, so this is one we can count on. But it is also an invariance that is often over-emphasized in geopolitical analyses. The reason I think it is overvalued is that the population decline needs to be understood against the background of urbanization rates and slowing world population growth. If we toy around with a simple model that states that geopolitical competitiveness is produced primarily in the cities³, and that AI will allow for automation of agriculture with maintained productivity growth, then the current urbanization rates in China - from around 65% to 75% in the next decade - might well offset the effects of population decline.⁴

The number of states in the world will stay between 190-200 for the next decade. This might seem a peculiar fact to point out, but it is an interesting invariance. It is unlikely that the number of countries will vary significantly over the coming ten years, and this implies that some of the logic of influence in geopolitics will stay the same — and continue to play out in international arenas. The number of states has been fairly stable for the last 20 years, growing from around 50 at the first world war to the roughly 200 we have today - which seems to be a stable number.⁵

The energy efficiency of computing will continue to grow by 100x per decade. This is a bolder prediction, but looking back we have seen amazing growth in energy efficiency across computational devices. With the new, promising technologies that we are looking at now, it seems unlikely to slow down. This is not to say that the efficiency gains will offset the energy need — but it is an interesting trend, and there is plenty of room before we reach the Landauer limit.⁶ Granted, it will be harder than before with thermal limits and there are some signs that efficiency gains are slowing down - but I will be optimistic on this one.

Privacy debates will still be open and engaging in the next decade. This is one that is interesting; one way to predict or work with public policy is to figure out how different issues play out over the long run. Some issues will stay the same, some will resolve. In contrast, I suspect that copyright issues will resolve the way they always have: litigation, legislative change and licensing — the three Ls. But privacy issues will continue to be top of mind in a society where we are recorded at higher resolution than ever before — with agents that can guide and advise us. If anything I suspect that the nature of these debates might change - from secrecy and confidentiality to autonomy; raising the question of how we retain individual free will in a world where excellent advice⁷ might be abundant.

We will still not have found life anywhere else in the universe. This is a constant that I think we should think more about than we do, but it is worth pointing out, since at some point we will have to face the fact that we might be it. I don’t believe that we will have any evidence of life anywhere else in the universe in the next 10 years - and I think this realization is likely to fuse with the climate change debate at some point for a more constructive debate.⁸

The US will still outspend at least the next 5-7 countries in defense. Here is where we are today:

This is an interesting invariance, because as long as it holds this spending - the 916 billion USD - will have to be directed at the best available hypothesis of what will increase or maintain military competitiveness. This makes this a source of potential funding for anyone who believes that we are facing a technological revolution in military affairs in the next decade.

The number of internet users will continue to grow, for at least the next 5 years. This is one that you had probably almost forgotten! But the fact that more and more people are connecting is still a key driver of change in many parts of the world - and can be harnessed as a massive force for good. Here are the predicted stats:

And these are just some of the invariances — I am sure that you can come up with more of them, and they do set the framework within which change happens - so they are worth studying. What does not change is as important as what does, the limits and constraints are key to understanding any problem. Predicting limits - while notoriously hard, perhaps harder than predicting change - is a worthwhile practice.

Happy new year and do share your invariances in the comments or with me!

Nicklas

1. Especially in technology futures!

2. Some even suggest that this is a viable investment strategy — see https://fs.blog/staying-the-same/ detailing how Jeff Bezos and Warren Buffett are interested in this question.

3. For an interesting discussion of if this will continue to hold - an invariance - see https://carnegieendowment.org/research/2024/12/will-major-cities-continue-to-shape-global-economic-governance?lang=en

4. See https://www.statista.com/statistics/270162/urbanization-in-china/

5. Indeed, it is interesting to try to figure out what would have had to happen for us to end up with 50 nations in 2035 — war? Climate disaster? Aliens? And conversely - let’s say that we end up with 400 countries - where a significant amount of them might be former US states, for example - with their own nuclear weapons — what would that world look like?

6. See https://en.wikipedia.org/wiki/Landauer%27s_principle — where it says that current computers are at operating at a billion times as much energy per operation.

7. One can be worried about manipulation, but I think that Hugo Mercier showed fairly clearly that manipulation is overvalued as a risk (see Mercier, H., 2020. Not born yesterday: The science of who we trust and what we believe. Princeton University Press). What I do worry about is the abudance of excellent, fantastic advice that always guides us to greater pleasure and satisfaction - and where the free will we retain is the freedom to make slightly worse and awkward choices.

8. I quite liked the way this argument is explored in Godfrey-Smith, P., 2024. Living on Earth: forests, corals, consciousness, and the making of the world and Green, J., 2023. The Possibility of Life: Science, Imagination, and Our Quest for Kinship in the Cosmos. Harlequin.