The Crowded Mountain -- Can You Still Make Money on the Internet?
Traps and Honey
Go where you get paid, not where you pay. It's obvious. But the moment you ask "where is that, exactly?" -- things get complicated fast.
It used to be simple. Set a trap on the mountain. Paint honey on a tree. Come back the next day and collect what you caught. The internet worked the same way. Put up some content. Build a service. People would come while you slept, and money would appear.
There was a time when that actually worked.
The Mountain Changed
There were barely any schools that taught programming. Hardly anyone even owned a computer. The mountain was quiet, and almost nobody was setting traps. So you caught plenty.
Building websites was hard too. Placing image spacers in Dreamweaver, slicing up designs into table-based layouts -- it was a real craft.
Now? Look around the mountain. Traps everywhere. Every tree dripping with honey. If you saw the same scene on an actual mountain, you'd immediately think, "No chance this is going to work." But on the internet, you can't see it. So people keep setting traps. Trap number 200,000,001, each person convinced theirs is special.
The skills didn't change. The mountain did.
The Moat Has Been Drained
In investing, a "moat" is what keeps competition out -- a skill gap, a barrier, something others can't easily cross. Warren Buffett is often credited with popularizing the term. If you could code and most people couldn't, that was your moat.
It happened before -- tools making something accessible to everyone. Squarespace drained the web design moat. YouTube drained the education moat. Each time a tool made something easy, one skill stopped being special.
But LLMs are different. They're general-purpose moat drainers. Coding, writing, design, translation, data analysis -- all being leveled at once. Not one skill at a time. All of them, simultaneously.
So What Now?
A few options come to mind.
Set a massive number of traps and play the odds. Brute force. Put out 500 traps and maybe you'll catch something. But the competition isn't one person -- it's infinite. You can't win a volume game against infinity.
Find a dangerous mountain nobody else can reach. Somewhere difficult, risky, and hard to imitate. The barrier to entry itself eliminates the competition. Legacy systems, regulated industries, security, low-level engineering -- fields where the cost of failure is high and casual players stay away.
Teach people how to set traps. Move from catching to teaching. There are always fewer teachers than trappers. If you've actually caught things yourself, your materials carry real weight. (Build it once, sell it on repeat as a SaaS -- though Udemy, Codecademy, and Progate already have.) But how do you teach with a straight face? You know the original mountain is already packed, and your students probably won't catch much. But look closer -- teaching is just another mountain.
Who Wants You on the Mountain?
Here's what nobody talks about: the crowding isn't an accident. Somebody profits from it.
Platforms need creators to keep uploading -- that's their inventory. Course sellers need students to believe the mountain still has room -- that's their market. Tool vendors need builders to keep building -- that's their revenue. The entire ecosystem runs on the assumption that your trap is the one that'll work.
They don't need you to succeed. They need you to try.
The Value of Seeing
Every mountain you look at leads to the same conclusion: "This one's already full of traps too."
But the fact that you can see it -- that's rare. Most people set their traps without ever looking at the mountain as a whole. If you can see the crowding, you can make different decisions than those who can't.
What's Left After the Moat Is Gone?
The conventional answer: the only lasting moat is what LLMs can't do. Real-world relationships, trust, accountability, judgment too important to delegate to a machine.
But here's the problem with that answer: those are the exact skills that technical people tend to be worst at. The introverted programmer's dream was always "let the work speak for itself." And for a while, the mountain was quiet enough that it could. Now the mountain is screaming and nobody can hear your work over the noise.
The Opposite Shore
When everyone was in the real world, the edge belonged to those in the internet world. Not programming itself, but being somewhere the crowd wasn't.
Now everyone has come to the internet. The tool shapes what you see -- if you're always in front of a screen, every opportunity looks like an internet one. Another trap on the crowded mountain.
The skill was never the specific mountain. It was the instinct for being where others aren't.
A Different Game
There's one more option nobody lists: stop setting traps entirely.
Every strategy so far assumes the game is trapping -- catching something on the mountain. More traps, harder mountains, teaching others to trap. But what if the game itself is the wrong frame?
Consider investing. It's pattern recognition -- seeing what others don't see. Judging which mountains are crowded and which aren't. That's not trapping. That's reading the terrain. And the ability to see crowding -- the exact skill this whole essay is about -- is the core competency.
And here, LLMs become useful for a different reason. Reading the terrain isn't just thinking -- it requires building instruments. Systems to process data, test hypotheses, model patterns. That used to be bottlenecked by coding. Now LLMs make building those tools trivial. The bottleneck shifts to the part that can't be automated -- knowing what to build and why. The moat was never the code. It was the judgment behind it.
Maybe the answer isn't a different mountain at all. Maybe it's a different game entirely -- not setting traps and waiting, but reading the mountains themselves. Not setting traps, but understanding where the trappers will go next.
For the first time, the path doesn't look crowded.