Re: Guidelines for pre-trained ML model weight binaries

[Andreas Enge]

Hello,

related to this thread, I just came across an entry in Cory Doctorow's blog:
   
https://pluralistic.net/2023/08/18/openwashing/#you-keep-using-that-word-i-do-not-think-it-means-what-you-think-it-means

It is already interesting in its disection of the terms "open" vs. "free",
which is quite relevant to us (but just echoes the sentiment I had anyway).
The end can be seen as an invitation to *not* package neurol network
related software at all: by packaging "big corporation X"'s free software,
but which is untrainable on anything but big corporations' hardware, we
actually help big corporation X to entrap users into its "ecosystem".

Andreas