How to Actually Pick Skills That Last Even When Everything Around You Changes
In 1990, there were over 200,000 travel agents in America
who had built years of real expertise regarding airline systems
and destination knowledge.
When Expedia launched in 1996, over 100,000
of those jobs vanished within 15 years.
Their knowledge didn’t become less accurate; it became unnecessary.
They built in a direction the world was quietly moving away from.
To avoid becoming the modern equivalent of a travel agent,
you cannot simply guess the future.
You must run every potential skill through a specific three-part filter
before investing years into learning it.
Level One: The Lindy Filter
The first filter comes from philosopher
and mathematician Nassim Taleb,
who popularized the concept of the Lindy Effect.
The idea is simple: for non-physical things,
the longer something has already survived,
the longer it is likely to keep surviving.
A book read for a thousand years will likely
be read for another thousand.
Mathematicians use a formula called the “hazard function”
to make this precise—the longer something has lasted,
the lower its probability of expiring soon.
- High-Risk Skills: Search Engine Optimization (SEO) has existed for roughly 25 years, but AI search is currently making the technical knowledge of SEO experts obsolete. AI prompt engineering has existed for about three years, and the biggest AI companies are actively building systems to make prompting unnecessary. Investing years into prompt engineering right now is a high-risk bet.
- Low-Risk Skills: Writing has been valuable for thousands of years. Every new technology—the printing press, email, social media, and AI—has made good writing more valuable. The hazard rate is near zero.
This doesn’t mean you should only build ancient skills,
but rather apply ancient skills to modern formats.
Writing YouTube scripts is still writing;
digital marketing is still persuasion.
Before building a skill,
ask how long the underlying capability has been valuable.
If the answer is decades or centuries, you are safe.
If it is only a few years, be very careful.
Level Two: The Combination Advantage
Passing the first filter only tells you a skill is safe to build;
it doesn’t tell you how valuable it will actually be.
Value doesn’t come from being the best at one single thing;
it comes from being the only person at a specific intersection of things.
Most people think about skills one at a time.
However, when you start combining them,
the results are mathematically surprising.
Adding ten skills doesn’t produce ten possible combinations;
it produces over three million.
Every skill you add multiplies your uniqueness.
For example, Tim Urban built the blog Wait But Why into
a massive success read by hundreds of millions of people.
His individual skills were not world-class—he was a decent writer,
a decent researcher, and a decent illustrator.
But combined into one person making long-form,
illustrated deep dives on complex topics,
the combination was completely unique. He created his own category.
The most valuable combination right now is genuine expertise
in any traditional field combined with a real understanding
of how AI works.
Before building your next skill, ask what rare combination
it creates with everything you already have.
Level Three: The Leverage Filter
The third filter is the most important and the most overlooked.
As Naval Ravikant stated, the most important question
about any skill isn’t how much it pays;
it is whether the returns compound or stop when you stop working.
- Linear Skills: You apply the skill, you get paid, and then you stop. The income stops with you. The ceiling is always the number of hours you have available.
- Leveraged Skills: You do the work once, and it keeps working without you. A professor teaches 30 students in a room, while a YouTuber explains the same concept once in a video that reaches three million people while they sleep, earning money for years.
When two skills pass the first two filters,
always choose the one where the returns do not stop when you do.
