There's a useful distinction in forecasting: signal versus noise. Noise is the churn of daily headlines — mostly random, mostly forgotten within a week. Signal is a slow-moving current backed by hard-to-fake data — demographics, physics, technology diffusion — that keeps flowing regardless of any single news cycle. This piece is about signal: not predictions of exactly what will happen, but the deep currents worth understanding, presented with the genuine uncertainty and open debate that honesty requires.
Fertility rates have dropped below the replacement level in most of the world — not just wealthy countries but increasingly middle-income ones too. This is one of the most reliable trends in social science, because the people who will be adults in 20 years are already born and counted today.
Artificial intelligence and automation are already changing how many jobs are done. There's broad expert agreement this continues; there's genuine, active disagreement about how fast, which jobs, and whether new roles will replace old ones as fast as history's past automations did.
The costs of solar, wind, and battery storage have fallen dramatically over the past decade, and their share of new electricity generation has grown accordingly — a well-documented technological and economic trend. How fast the overall energy system shifts, and what role different sources play alongside it, remains a genuinely contested policy and engineering question.
Economic and military weight has been spreading across more countries and regions than the more single-centered world of the late 20th century — a trend widely noted across the political spectrum, even as analysts disagree sharply about its pace, its consequences, and which powers will matter most.
Tools like CRISPR gene editing and mRNA vaccine platforms (proven at global scale during COVID-19) signal a shift: biology is increasingly something we can precisely read and edit, not just observe. This is early-stage but moving quickly, with real momentum in medicine and agriculture.
Global urbanization has been rising for over a century and is projected to continue; meanwhile, time spent on digital devices and platforms has grown sharply within a single generation. Both trends are well-measured. Their effects on wellbeing, community, and attention are actively studied and debated rather than settled.
The point of tracking signal isn't to predict the future with certainty. It's to stop being blindsided by things that were, in hindsight, entirely visible in advance.
How to hold all of this well
A few honest caveats, in the spirit of the rest of this library:
Trends bend; they don't just extend. Every trend above could accelerate, plateau, or reverse due to a war, a breakthrough, a policy shift, or a crisis nobody currently foresees — that's exactly what happened with COVID-19's effect on remote work, or the sudden cost collapse in solar power. Treat these as the current of the river, not a guarantee of exactly where the boat ends up.
Uncertainty is not the same as "nobody knows anything." Demographics 20 years out are unusually predictable because the people involved already exist. Technology diffusion speed and geopolitical shifts are far less certain. Hold your confidence in proportion to the evidence for each trend, not evenly across all of them.
These are not verdicts on what's good or bad. Reasonable, well-informed people disagree sharply about whether each of these trends is a net positive, negative, or a mix — and that disagreement usually reflects real, different values and priorities, not just different facts. This piece aims to describe the currents, not to tell you how to feel about them.
Several slow-moving, well-evidenced currents are shaping the next two decades: an aging, slower-growing population; AI and automation reshaping work; a transitioning energy system; a less single-centered geopolitical order; biology becoming programmable; and deepening urbanization and digital life. Each carries genuine, contested trade-offs — the goal isn't to predict exact outcomes, but to notice currents early enough to think and plan clearly, rather than being surprised by what, in hindsight, was visible all along.
Try this: pick the one trend above most relevant to your own work or life, and ask what it implies for a decision you'll face in the next few years. Signal is only useful if it changes what you do with it.
Scope. Directional trends draw on demographic, economic, and technological data from sources like the UN Population Division, IEA, and IMF; the pace, causes, and value judgments around each are subjects of genuine, ongoing expert and public debate, and this piece aims to present multiple reasonable perspectives rather than a single forecast.
- 🌐 Our World in Data — long-run trend data across most topics above
- 📘 Superforecasting — Philip Tetlock & Dan Gardner (how to think probabilistically about the future)
- 🌐 UN World Population Prospects — the demographic data behind trend one
- 🌐 IEA World Energy Outlook — for the energy-transition trend, updated annually