Some conventions are load-bearing. They’re doing real cognitive or operational work, and breaking them breaks something important. Others are inherited habits, still standing because nobody thought to push.

Under each convention sits a belief that makes it feel necessary. This is where it gets uncomfortable. You’re not asking what people do. You’re asking why it would feel crazy to do otherwise.

“Software is sold through enterprise sales teams” sits on top of “purchasing decisions are too complex for self-service.”

“Gyms charge monthly memberships” sits

... See more

We’re living through one of those moments when assumptions expire in bulk.

AI isn’t changing a few industries. It’s dissolving beliefs across all of them. What counts as expertise. Who gets to create. What requires human judgment and what doesn’t. The stories we’ve told ourselves about work, about skill, about the barriers between having an idea and... See more

For decades, humans had to learn the machine’s language to bridge the gap.

Large language models changed that. Now the machine can learn ours.

Expertise ossifies assumptions into certainty.

This is the trap. The more you know a category, the harder it is to see its assumptions. Deep knowledge doesn’t reveal hidden beliefs. It encodes them as instinct. The veteran who “just knows” how the market works is often the last to notice when the underlying logic shifts.

They weren’t stupid. They were experts. And their expertise was a prison they couldn’t see the walls of.

Spotify asked a different question: why does anyone need to own music at all?

Every industry runs on invisible assumptions about what it is, who it serves, what counts as value, what’s possible and what isn’t. These assumptions calcify into conventions. The conventions get followed so long that they stop looking like choices. They start looking like physics.

But they’re not physics. They’re agreements. And agreements expire.