Post #1647366

RE: https://social.coop/@cwebber/116426152872895154 Christine Lemmer-Webber as per usual really articulated well, something ive been trying to say much less elegantly. (Conclusion quoted here beginning of the sub-thread here: https://social.coop/@cwebber/116426025287444979 ) My fear is with these few possible realistic outcomes: The output of AI is a derivative work of the training data. LLM generated code containing near-verbatim copies of others code is legally unusable.The output of AI code is effectively public-domain; and FLOSS and proprietary software face copyright mexican standoff. With LLMs able to scrub the copyright off code.We are in fact hurtling towards a third: The output of AI code is effectively public-domain, but courts consistently side with big tech in protecting their products in a "let them have their cake and eat it too" style hypocrisy which eventually becomes normalized.