7 min read
N02

Mind · No. 13

Bet on Old Things

The best predictor of how long something will last is how long it has already lasted.

· · 7 min · 1,532 words

In short

  • The Lindy effect applies to non-perishable things like ideas, books and technologies, which age in reverse compared with living things.
  • Survival over long periods is evidence of robustness, because time has already tested what is old.
  • Reading old books and building on proven technologies are ways of borrowing that filtering for free.
  • New things should be adopted at the edges of a system, with old things at its core.
  • The Lindy effect is a default, not a law, and it does not tell you which new things will become the old things of the future.

In the middle of the twentieth century, comedians in New York would gather late at night at a delicatessen called Lindy's on Broadway. They ate cheesecake and talked shop, and one thing they talked about was how long a career could last.

The theory that emerged around those tables, written up by the writer Albert Goldman in a 1964 magazine article, went something like this. A comedian has a limited amount of material. The more often they appear on television, the faster they burn through it. So you could predict how much longer someone would last by how much they had already appeared.

It was a half-joke about show business. Decades later, the mathematician Benoit Mandelbrot and then the essayist Nassim Nicholas Taleb turned it into something much sharper. In his book Antifragile, Taleb popularised the name it has today, the Lindy effect, along with a formulation that sounds wrong the first time you hear it.

For things that do not die of old age, every additional year of life increases their expected remaining life.

Things that age in reverse

Living things work the opposite way. A ninety-year-old person has fewer expected years left than a twenty-year-old, because bodies wear out. Your car, your phone and your knees are the same. They are perishable. Age is a cost.

But many important things are not perishable in that sense. A book does not wear out because people keep reading it. An idea does not get tired. A technology like the wheel or the alphabet has no internal clock ticking down. For these things, age means something else. It is information.

If a book has been in print for ten years, a reasonable guess is that it will be in print for roughly another ten. If it has been read for five hundred years, it will likely be read for centuries more. The rough rule is that the expected future is proportional to the past.

That sounds like a strange superstition until you think about what survival actually means.

An old thing has passed a test that a new thing has not even taken yet.

Time as a filter

Every year, thousands of new books are published. Almost all of them disappear. The same goes for new frameworks, diets, management theories, fashions, apps and opinions. Most new things fail not because they are bad, but because the world is crowded and attention is limited. Only a small number make it through.

So when you look at something that has lasted, you are looking at a survivor of a brutal selection process. It survived changes in taste, technology, language and politics. It survived the death of everyone who first loved it. Whatever qualities it has, they were robust enough to keep earning new people across generations.

Consider al-Mutanabbi, who wrote his poems in the tenth century. People across the Arab world still quote him, and schoolchildren still memorise his verses. Thousands of poets were writing in his time, and most of their work is gone. The fact that his lines are still being recited more than a thousand years later tells you something that no critic could: they keep working on new readers. A reasonable bet is that they will keep doing so for a very long time.

Or take something less literary. People in this region have been brewing coffee with cardamom, pouring it from a dallah into small cups and offering it to guests for centuries. Coffee chains will come and go. That ritual will very likely outlast all of them.

What to read

The most practical use of the Lindy effect is deciding what to read.

New books have to fight for your attention with marketing, launch buzz and the feeling that you are behind if you have not read them. Most of them will be forgotten in a few years, and if you wait, time will tell you which ones mattered for free.

Old books have already been filtered. If people are still reading Marcus Aurelius, Ibn Khaldun or Euclid, the reason is probably not marketing. Something in them keeps being useful across completely different worlds.

This does not mean you should read only old books. It means your reading diet should have a heavy base of things that have lasted, with newer work on top. If a book has been in print for fifty years, it is probably worth your time. If it came out last month, you are taking a much bigger gamble, and you should know it.

The same goes for news, which is the least Lindy kind of information there is. Most of what fills your feed today will be irrelevant within a week. If you want to understand the world, a good history book usually teaches you more than a year of headlines.

What to build on

Software has its own version of this, and every experienced engineer learns it eventually, usually the hard way.

Some technologies have been around for decades and are still everywhere. SQL dates back to the 1970s and is still how a huge share of the world's data is queried. Unix ideas from that same era sit underneath the servers that run most of the internet and inside the phone in your pocket. Email is older than the web, and businesses still run on it. The C language is half a century old and still compiles the operating systems everything else depends on.

Meanwhile, the history of software is full of exciting frameworks that were everywhere for a couple of years and are now abandoned, leaving teams stuck rewriting code built on top of them.

A Lindy-minded engineer still uses new tools, but puts old, boring, proven technology at the core of a system, where failure is expensive, and experiments with new things at the edges, where failure is cheap. Your database should probably be older than your intern. Your experimental feature can use whatever you like.

Jeff Bezos once made a related point about strategy. People often ask him what will change in the next ten years, he said, but rarely ask what will not change, and that second question is the more useful one, because you can build a business around things that stay stable. Customers will still want low prices and fast delivery a decade from now. You can invest in that with confidence.

Things that have lasted a long time are the closest thing you will find to things that will not change.

What to believe

The Lindy effect also applies to ideas about how to live.

Advice like keeping your word, being careful about who you spend time with, taking care of your body, being generous with guests and not borrowing more than you can repay has been repeated by grandparents across cultures for a very long time. It has been tested by millions of lives.

A new productivity method or life philosophy that went viral last year has been tested by almost nobody. It might be great. But the burden of proof is on it, not on the old advice.

There is a kind of humility in this. It means assuming that people who came before you were not fools, and that the habits and customs they kept for centuries probably exist for reasons, even when those reasons are not written down.

The old advice is not old because people were simple. It is old because it kept working.

Where it breaks

Like any rule of thumb, the Lindy effect can be misused. Three limits matter.

First, it applies only to non-perishable things. Do not use it on people, machines or anything else that wears out.

Second, it is a statement about probability, not certainty. Old things do die. Long-lasting empires fell. Industries that had existed for centuries vanished in a decade when the conditions under them changed. Age protects against ordinary shocks, not against every shock.

Third, it says nothing about which new things are good. Every old thing was once new. The printing press was new. The internet was new. If everyone had refused anything young, nothing would ever have grown old. The Lindy effect helps you set a default and size your bets. It does not replace judgment about which new things are worth backing early.

The sensible position is a balance. Build your foundations on what has lasted. Put a smaller, deliberate share of your time and money into new things you have a real reason to believe in. Expect most of them to fail, and be glad when one does not.

The long view

There is something calming about the Lindy effect, especially when the world feels like it is changing every week.

Most of what demands your attention right now is new, loud and temporary. Underneath it, a slower layer of the world keeps going: languages, stories, crafts, a few great books, the basic ways people trust each other and the tools that have outlived everyone who invented them.

When you are deciding where to put your time, you can ask a simple question. Has this survived anything yet?

If not, it might still be wonderful. But the things that have already outlived their makers are usually the ones that will outlive you too.

Further reading

  • Antifragile — Nassim Nicholas Taleb
  • The Clock of the Long Now — Stewart Brand
  • How Buildings Learn — Stewart Brand

More from the shelf

N06 · Mind · 7 minThe Map Is Not the TerritoryEvery metric, résumé and org chart is a map. The trouble starts the day you forget it. N07 · Mind · 6 minInvert, Always InvertCharlie Munger once gave a graduation speech on how to be miserable. It is still the best advice he ever gave. O02 · Origins · 6 minThe Man Inside Every AlgorithmEvery app on your phone carries the name of a ninth-century Baghdad scholar most people have never heard of.

See the whole library →