Free software is dead. AI agents are about to resurrect it.
Richard Stallman fought a war for software freedom in the '90s. He won the legal battle but lost the practical one.
Then SaaS happened.
When your software lives on someone else's servers, "source code access" becomes meaningless. You don't run the software. They do. The four freedoms — run, study, modify, redistribute — became theoretical rights nobody could exercise.
AI coding agents just changed that equation.
The problem Stallman couldn't solve
Free software's original weakness was accessibility. The four freedoms assumed you could read and modify source code. Most people can't.
This was always the elephant in the room. Critics pointed it out for decades: free software licenses are meaningless for users who lack the expertise to exercise those freedoms. The movement narrowed itself to legal requirements and code, effectively limiting its audience to technical enthusiasts.
Then SaaS delivered the killing blow. The GPL required you to share source code when you distributed software. But if you never distributed it — if you just ran it on your servers and let people access it over the web — the license didn't apply.
Companies loved this. AWS built an empire on open source without sharing modifications. Google banned AGPL across the board. The "open source" rebrand stripped out the ethical philosophy and kept only the development methodology.
Software freedom became a nice idea that didn't matter in practice.
Agents change everything
Here's the thing about AI coding agents: they can exercise software freedom on behalf of people who can't code.
When you tell Claude or Codex "make my task manager automatically categorize tweets I save," you're exercising Freedom 1 — the freedom to study and modify — through a proxy. You don't need to understand the codebase. You don't need to know what a GraphQL schema is. You just describe what you want, and the agent exercises the technical freedoms for you.
This bridges the gap between software freedom as an abstract right and software freedom as a practical capability.
Consider the practical difference:
- Proprietary SaaS: "I wish my task manager could do X. Let me submit a feature request and check back in 2032."
- Free software + agent: "Agent, make my task manager do X." Done in ten minutes.
The four freedoms stop being a theoretical right and start being the practical difference between solving your problem and being helpless.
A real example (it's ugly)
Someone recently tried to build a simple workflow: share a tweet, have an AI generate a smart task title, save it to Sunsama with the right category.
Sounds simple. Here's what it actually required:
- A reverse-engineered API (because Sunsama has no official API in 2026)
- Storing their actual password as an environment variable (because no OAuth)
- A serverless function they have to host and maintain
- An Anthropic API key for the title generation
- A manually-built iOS Shortcut (no programmatic way to create them)
- A dependency on a stranger's open source project that could break any time
This is what closed software does to motivated users. Six layers of workarounds for a twenty-minute idea.
Now imagine Sunsama was free software. The agent reads the source code, understands the task data model, modifies the default behavior to match the user's preference. Done. No reverse engineering. No gray-zone APIs. No infrastructure debt.
Why this matters for builders right now
If you're building a product in 2026, pay attention. The choice between open and closed just got a new dimension.
Open source gives your users' agents superpowers. When your codebase is accessible, every user with an AI agent becomes a power user. They can extend, customize, and modify your software to fit their exact needs — without submitting a feature request or waiting for your roadmap.
Closed source creates walls that even smart agents can't climb. No source code means no modification. No API means no integration. The agent hits the same walls every non-technical user hits: dependency and helplessness.
This is a competitive advantage that most founders haven't realized yet. Open source isn't just about community contributions or developer goodwill anymore. It's about giving every user's AI agent the ability to make your software work for them.
The free software movement is coming back
For a long time, the "free vs. proprietary" debate felt academic. SaaS made it irrelevant. Convenience won.
But agents change who can exercise software freedom. It's not just programmers anymore. It's anyone who can describe what they want in plain English.
The four freedoms were written for a world where only humans read code. In 2026, something finally can read it on your behalf — and actually do something about it.
If you're building something, think about this: do you want your users' agents to be able to help them, or do you want them hitting walls?
The answer might determine who wins the next decade of software.