You’ve probably looked at a massively successful idea and thought, “I could have done that.”
It’s a comforting lie we tell ourselves. When an innovation is stripped of its mystery by time, it feels as natural as breathing. We look at Google’s PageRank today—counting links as votes—and think it’s just basic common sense. A child could code it.
But there’s a famous story about the artist Damien Hirst. A critic looked at his pickled shark and scoffed, “I could have done that.” Hirst simply replied, “But you didn’t, did you. I did.”
Hindsight is a liar that makes genius look like common sense.
Let’s rewind to 1996. The internet was a chaotic, exploding frontier. If you wanted to find something, you used AltaVista. And if you were there, you remember the daily frustration: AltaVista was terrible. We were all drowning in a sea of spam and irrelevant garbage.
Millions of us were using the web every single day. We felt the pain. We complained about the search results. Yet, not a single one of us thought to map the web as a graph. We didn’t invent PageRank. We didn’t even come close.
Why? Because the real magic of PageRank wasn’t the math. The algorithm is practically elementary. The true breakthrough was a mental model.
Before you can count links as votes, you have to stop seeing the internet as a collection of isolated text documents. You have to reframe the entire web as a connected graph. You have to treat hyperlinks as distributed relevance votes.
You don’t need a better algorithm; you need a better map of the territory.
That is a first-principles move. It requires tearing down the existing paradigm of keyword-matching and building a completely new structural understanding of how information flows. It’s easy to explain after it’s been done. It was incredibly hard to see before anyone had made the leap.
But here is the part everyone misses when they study this history.
PageRank wouldn’t work today. If you launched the exact same algorithm right now, the web would eat it alive.
The reason PageRank succeeded wasn’t just the elegance of the graph model. It succeeded because it was deployed into a pristine, untouched web. In 1996, nobody was gaming links because PageRank didn’t exist yet. The system was vulnerable to link-farming, but nobody had thought to build a link farm.
Today, the web is a battleground. Every angle is exploited. Every link is bought, sold, or manipulated. The environment has fundamentally changed.
The window for the obvious is always closing. By the time you see it, the game has already moved.
Every era rewards a different first-principles redefinition of relevance. In 1996, the winning move was treating links as votes. Today, the winning move is something else entirely—something hiding in plain sight, waiting for someone to draw a new map.
So stop obsessing over the algorithms of the past. Stop kicking yourself for not inventing PageRank. You didn’t invent it because you didn’t have the mental model.
If you want to build the next breakthrough, don’t look for a more complex formula. Look at the current landscape, find the fundamental assumption everyone is blindly accepting, and shatter it. That’s where the next shark is waiting to be pickled.
FAQ
Q: Wasn't PageRank just basic linear algebra that anyone could have figured out?
A: No. The math was the easy part. The breakthrough was the mental leap of reframing the web from a pile of documents into a connected graph where links acted as votes. You can't solve a problem with an algorithm if you haven't fundamentally redefined the problem first.
Q: What's the practical implication for building tech today?
A: Stop trying to build slightly better versions of existing paradigms. The next massive breakthrough won't come from a more complex algorithm; it will come from someone who looks at the current, polluted state of the web and draws a completely new structural map of how relevance actually works.
Q: If PageRank is so brilliant, why wouldn't it work today?
A: Because the environment has changed. PageRank was deployed into a pristine web where nobody was gaming links. Today, the web is entirely adversarial. The same 'obvious' insight fails because every era rewards a different first-principles redefinition of relevance. The window for that specific solution is permanently closed.