It's Just Old

People age forward. Ideas age in reverse.

You carry one intuition about age, built entirely from watching living things: the older something is, the closer it is to the end. It is true for you, your dog, a carton of milk. So we quietly apply it to everything — a five-year-old startup, a forty-year-old book, a two-thousand-year-old idea — and we get it exactly backwards.

There are two kinds of things, and they run on opposite clocks. Perishables — bodies, machines, anything with a built-in lifespan — burn down: every year lived subtracts from the expected years left. Non-perishables — ideas, books, technologies, institutions — do the opposite: surviving is the evidence that they are robust, so every extra year of survival lengthens the future you should expect. Drag the age below and watch the two clocks pull apart.

Read where the lines cross. While both are young, the person wins — a child has a whole life ahead, the idea barely a footnote. But the person's curve is doomed to bend down to zero, while the idea's just keeps climbing. Somewhere in the middle they swap places, and after that the gap only widens. The forty-year-old book is not “probably about done.” It is, by the only evidence that exists, just getting started.

“For the perishable, every additional day in its life translates into a shorter additional life expectancy. For the nonperishable, every additional day may imply a longer life expectancy.”

Nassim Nicholas Taleb · Antifragile

That sounds like a paradox or a rhetorical trick. It is neither. It is what one specific, well-behaved distribution does — and you can watch it fall out of pure sampling, with no hand on the scale. That's Game Two.

Survival is the only forecast you need

Here is the whole mechanism, computed honestly. Sample thousands of lifespans from a power law — the distribution of things with no natural scale, where most die young but a few run astonishingly long. Now stand at some age t and look only at the survivors: everything still alive at t. Ask how much life a typical survivor has left. The answer is not a constant. It grows — and it grows in a straight line with t.

Below: a fat histogram of the remaining life of every survivor at age t. Slide t to the right and the whole survivor population shifts — the green line (the typical life still left) marches right with it, in lockstep, exactly as the formula says. Then change α: make the class more fragile and the multiple shrinks; make it more durable and a single extra year of survival buys far more expected future. Nothing here is asserted — it's measured from the sample each time. Redraw the population a few times: the median barely flinches, while the average jumps around wildly — a fat tail has no stable average to settle on.

This is the engine under the paradox. A power law has no built-in “normal” lifespan to decay toward, so age stops being a countdown and becomes a credential. Having lasted is the strongest available signal that the thing is the durable kind — and the longer it lasts, the stronger that signal gets. Mandelbrot worked out the mathematics — life expectancy proportional to age; Taleb made the consequence concrete:

“If a book has been in print for forty years, I can expect it to be in print for another forty years. But, and that is the main difference, if it survives another decade, then it will be expected to be in print another fifty years.”

Nassim Nicholas Taleb · Antifragile

A rule that simple is also a decision rule. If age predicts survival, then when you have to choose what to read, learn, build on, or bet your scarce years against — you already have the only feature that matters. That's Game Three.

Bet on what has already survived

You only get so many hours to spend learning a tool, adopting a standard, or building on a foundation. Every choice is a bet that the thing will still be here when you need it. The new and shiny screams for attention; the old and boring asks for none. But run the clock forward and watch what the Lindy survival curve does to a shelf of candidates of wildly different ages.

Each bar is one thing you could commit to, at its real current age. Push the horizon out — five years, fifty, two hundred — and each bar shows the chance it's still standing that far ahead, under the same power-law survival from Game Two: P = (age / (age + horizon))^α. The hyped two-year-old framework evaporates almost immediately. The thing that's already 2,000 years old barely notices the future passing.

This is not nostalgia, and it is not a ban on the new. It is a prior. When two options compete and you cannot tell which is better — and most of the time you genuinely can't — the one that has already survived longer is the safer place to plant your time, because it has passed more tests you'll never see. The new thing has to earn its way past that prior, not merely look exciting. Most of the noise never does. The few things that survive a long time are quietly telling you they intend to keep doing it.