Free software's four freedoms, run, study, modify, and distribute, became largely symbolic once software moved to the cloud. You can't modify what runs on someone else's server. SaaS didn't kill free software ideology. It just made it irrelevant to most users' daily lives.
AI agents change the calculus. When software actively acts on your behalf, reads your files, sends your emails, and executes decisions autonomously, the question of who controls that agent's code becomes urgent again. A closed-source agent is not just a tool you use. It is a proxy operating inside your life with rules you cannot inspect.
The original article is worth reading for how it reconstructs the free software argument from first principles for the agent era, not just nostalgia for GNU. The core tension it identifies: the more capable and autonomous AI agents become, the more catastrophic it is to have no access to their internals.
[READ ORIGINAL →]