17 articles in this category
Ten years is a long time to be wrong. And if you're building a model that reaches that far, you will be wrong— eventually. The question is whether you...
Back in 2018, a team at a mid-sized think tank built a population model that looked rock-solid. They had census data, migration flows, birth rates—eve...
Most predictive models are built on a silent assumption: the future will look a lot like the past. Train on last year's data, predict next quarter. Wo...
In the early 1990s, a climate model developed by Soviet scientists kept running on a supercomputer in Siberia—two years after the USSR dissolved. The ...
Here's a question nobody asks when they're tuning hyperparameters: Does my model care about people born in 2070? Most long-horizon models—climate simu...
You run a 30-year commodity price forecast for a mining firm. The model says copper demand grows smoothly at 2.1% per year. But the model's ecosystem ...
Pick a number. Ten years? Fifty? A century? That decision might sound technical, but it's deeply ethical. Every long-horizon model embeds our assumpti...
Imagine you're funding a climate model that aims to predict the year 2500. The project will cost millions, take a decade to build, and rely on institu...
Here is a number that should make you uncomfortable: a 0.1 percentage point increase in the discount rate used by the U.S. Social Security trustees sh...
You've just shipped a model that nails the test set. RMSE down 12%. Leadership is happy. Then the market shifts, a regulation lands, or a sensor fails...
You are building a model that claims what the world will look like in 2073. Your stakeholders nod politely. Then someone asks: How do we know you are ...
Imagine a rural health district trying to plan for a potential hospital closure. They commission a predictive model that forecasts patient volume, sta...
Two years ago I sat in a room with a dozen ecologists and one community elder from a coastal village in Alaska. The ecologists were proud: a 50-year s...
Every six months, another long-term forecasting model hits output. Six months later, it's quietly retired. Not because it was off—but because it was b...
Every forecast is a bet. A decade-long forecast is a bet wrapped in layers of assumptions, compounding errors, and deferred reckoning. And yet, most a...
You trained a model on data from 2010–2020. It works beautifully. But now it's 2025, and the world has changed. Your training data is a historical art...
Nobody sets out to build a forecasting model that locks out the next generation. But that is exactly what happens when you choose a horizon that is ei...