@j3j5@blogdiva@KathyReid this appears to be only an opinion, which 1) doesn't appear very informed to me, and 2) only mentions code-related issues, and does not address them.
More precisely, clearly when talking of code, this is not part of the process used:
"Diffusion models like Stable Diffusion and Midjourney take these inputs, add “noise” to them, corrupting them, and then train neural networks to remove the corruption."
Neither is this:
"This is because using the digitized books as part of the database provided information about the books and did not use them for their creative content"
Also, regarding the court's inverse question, it seems to me that this is an extremely valid concern:
"On this point, the court wrote that the better question to answer was not how much of the works [company] copied, but instead how much was available to users". Due to the granularity of SO, this could easily be 100%, even though this would need to be illustrated (=proven) as in the NYT case.
Could anyone identify "substantial, non infringing uses" of the SO data? Here using CC might make it difficult to establish what was the original interest of the contributor.
Final note, to me it remains a bit unclear what is the relevance of the Oracle Vs Google case according to the author.