The First Principle of Breaking Your Filter Bubble

This came up in a consulting conversation recently: how do you actually break through your filter bubble?

From platform dependency to a breakthrough

I’d been building more and more projects on Claude AI — iterating prompts, growing knowledge bases, automating workflows. The better it got, the more anxious I became.

I’m sensitive to system vulnerabilities. My dependency kept deepening, but one account ban would wipe everything out. And Claude has a reputation for banning accounts.

Then I stumbled on an essay by Steph Ango, CEO of Obsidian, called “File over App.”

The core idea: if you want to create digital artifacts that last, they need to be files you control, stored in formats that are easy to retrieve and read. It’s also a call to tool makers — acknowledge that all software is temporary. Give data ownership back to users. In the long run, the files you create matter more than the tools you used to create them. Apps fade. Files can endure.

That essay sealed the deal. I migrated everything to local-file AI systems.

Why “consuming more” doesn’t work

Most people try to break their filter bubble the obvious way — follow more people, try more platforms, subscribe to more newsletters.

The logic seems sound. But do the math: if your hit rate is 1%, doubling your intake just moves useful information from 1 to 2 while noise jumps from 99 to 198. You don’t get better at filtering. You just get more tired.

The problem was never about volume.

Your filter bubble is a sorting cost problem

Physics has this thought experiment called Maxwell’s demon — a tiny gatekeeper that can sort fast molecules from slow ones in a random mix.

Your filter bubble is essentially a “missing gatekeeper” problem.

Massive amounts of information hit you every day. Without precise filters, your brain defaults to the cheapest strategy: only process what looks familiar. Recognizing familiar things costs the least cognitive energy.

The bubble isn’t something someone built around you. It’s your cognitive system choosing energy-saving mode.

Why I recognized that essay instantly

Because I’d already spent serious time diagnosing my system’s vulnerability. I knew exactly what was wrong: all my knowledge assets lived on a platform I didn’t control.

That diagnosis installed a dedicated gatekeeper in my brain. One job: scan the information flood for anything related to portability and platform independence.

Tech CEOs post about AI all the time. I scroll past most of it. But my filter was precisely tuned to the problem File over App addresses.

It’s not how much information you see that determines what you notice. It’s how deeply you understand your own system.

Every time you diagnose your system — spot a vulnerability, identify a bottleneck, surface a hidden assumption — you install a new gatekeeper. More gatekeepers, higher precision in the information flood.

The formula

The real leverage for breaking your filter bubble isn’t “consume more.” It’s “understand your system more deeply.”

Depth of system self-knowledge = Precision of information capture.

The deeper you understand your business structure, thinking frameworks, and capability boundaries, the less you need to cast a wide net. You know what you’re looking for. Valuable information surfaces on its own.

The trap inside self-knowledge

When your filters get very precise, you’ll efficiently capture known-type valuable information — while systematically missing an entire category: information that challenges your filtering criteria themselves.

Knowing your circle of competence matters. But so does regularly questioning how you drew that circle.