Let Us Be Free

This post connects the historical free software argument — software freedom as a technical and political demand — to the current moment in AI, where questions about model weights, training data access, and infrastructure control are being relitigated. The framing is principled and the historical context is handled fairly. The problem is that this argument has been made many times in the past two years, and this version does not add a new angle, a concrete proposal, or a specific piece of legislation or tooling to engage with. It reads as a position statement rather than a call to action with a clear next step. For a founder thinking about open-weight models versus proprietary APIs, the underlying question is genuinely relevant to build decisions — but this post will not help you make that call. It is a values essay, not an analysis. -> Best for: anyone new to the open-source AI debate who wants a historical frame