Follow
The implication that training an #AI on someone's published work is stealing seems to me to carry the definite implication that training your meat-brain on someone's published work is also stealing.
("Only be sure always to call it please - research!")
Maybe the alleged distinction here is clearer to people whose conceptual brains reside entirely inside their skulls.
@cerebrate I'd argue that the way in which current LLMs learn from what they read is different to the ways in which humans learn, and could be construed as closer to "copying". (Various qualifiers here that wouldn't fit in the character limit.)
But whether that translates to "should be treated as copyright violation" is a harder question. I think the bigger problem here is that "is X copyvio?" is really mostly a proxy debate for "should authors be able to afford food and rent?" because of how the current everything couples those two questions.