---
title: "A capability, democratised"
url: https://raskell.io/notes/a-capability-democratised/
author: Raffael Schneider
date: 2026-10-25
description: "Building software took money, a team, or years of training. Now anyone with an idea and a subscription can do it, and nobody predicted what would follow."
tags: [democratisation, emergence, ai-engineering, history-of-technology, commons]
---

# A capability, democratised

Making software used to be reserved for people with a particular privilege: the money to pay a team, or the team itself, or the years it takes to learn to program well enough to build something real. [The power](https://raskell.io/notes/power-in-the-good-sense/) I wrote about is that privilege, democratised. Anyone with an idea, a subscription, and the patience to describe what they want can now have it built, and the ones who learn to describe well get a great deal.

Democratised capabilities have a history, and the history has a shape. Nobody who built the printing press or the web could say what people would do with them, and what people did was not on anyone's plan. The web's version of that story is in [The Web Was the Training Set](https://raskell.io/articles/the-web-was-the-training-set/): a protocol given away, and thirty years of uses nobody at CERN drew. It emerged, from millions of small decisions, and it was better for almost everyone, collectively and in the end, even where it was hard for the people in the middle of it. I expect the same of this, with the same caveat; [the move to the city](https://raskell.io/notes/the-move-to-the-city/) is the comparison I keep returning to. And I expect the window in which it is an advantage rather than a baseline to be short, for the reason in [the window before the utility](https://raskell.io/notes/the-window-before-the-utility/).