I've seen a viewpoint that goes something like "#FOSS should reclaim LLM's through open #LLM + training data":
https://writings.hongminhee.org/2026/01/histomat-foss-llm/
That said:
1. training data consists of copyrighted works. Exposing this exposes the theft
2. #GPL is enforceable because of the copyright system; copyright is the GPL's substrate.
As of now, it just reads like proponents want the benefits of the #copyright system but without any of the drawbacks.
(and yes, not all FOSS is GPL but this is still glaring)
Hong Minhee on Things
Histomat of F/OSS: We should reclaim LLMs, not reject themA few days ago, I came across a blog post titled On FLOSS and training LLMs that articulates a growing frustration within the free and open source software…
