From Programs to Source Code, and Now to Prompts

Table of contents

The Era of Shareware

In the early days of the internet, the “Shareware” model was the gold standard. Independent developers would often build useful tools or games entirely on their own and offer them for trial. You could download them for free, though some required a registration code—obtained through a purchase or donation—to unlock the full features. Others were completely open, relying on voluntary tips. Back then, software was distributed as binary programs: you could use them for free, but the inner workings (the source code) remained a mystery.

The Rise of Open Source

Then came the Open Source movement. Platforms like GitHub made it easier than ever to share and collaborate. The focus shifted from distributing a finished “black box” to sharing the source code itself. Suddenly, you could peek under the hood—auditing the logic, learning from it, and modifying it to fit your needs before sharing it again.

The AI Shift: Prompts as the New Code

Now, as Large Language Models (LLMs) become increasingly powerful, the object of “open source” is shifting again—this time from code to prompts. We are seeing GitHub repositories that contain no actual code at all; instead, they feature a single README file filled with prompts. By feeding these prompts into an LLM, a user can generate the target program on the spot. It’s a leap from binary files and source code to a “Just-In-Time” (JIT) way of creating software.

A Final Thought

We used to need a long, technical process to turn a design into a product. Now, the leap in technology is replacing the very tools we used to build things. It makes you wonder: as technology continues to automate the creative process, at what point does it start to replace the users themselves?