From: lynnmcguire5@gmail.com   
      
   On 11/1/2025 10:46 AM, Paul S Person wrote:   
   > On Fri, 31 Oct 2025 14:10:26 -0500, Lynn McGuire   
   >  wrote:   
   >   
   >> On 10/31/2025 11:44 AM, Paul S Person wrote:   
   >>> On Thu, 30 Oct 2025 21:50:02 -0500, Lynn McGuire   
   >>>  wrote:   
   >>>   
   >>>    
   >>>> “The lawsuit, which is part of the first wave of AI copyright cases,   
   >>>> revolves around three main arguments. Firstly, the authors claim that   
   >>>> OpenAI’s training of AI models on copyrighted books constitutes   
   >>>> infringement. Secondly, they allege that the company engaged in the   
   >>>> practice of pirating books from shadow libraries, regardless of whether   
   >>>> these books were used for training purposes. Lastly, the plaintiffs   
   >>>> argue that the answers generated by ChatGPT, OpenAI’s chatbot, are   
   >>>> substantially similar to the books on which they were trained.”   
   >>>   
   >>>    
   >>>   
   >>>> So does this mean that all book reviewers are infringing the author’s   
   >>>> copyright ?  I think not.   
   >>>   
   >>> Book reviewers are covered by fair use. Which does not encompass   
   >>> copying entire books into the review (I suspect).   
   >>>   
   >>> But it would be infringement (I suspect) if one asked the AI for a   
   >>> /review/ and it instead produced /the entire work/.   
   >>>   
   >>> On another newsgroup, we heard complaints from a person who had   
   >>> laboriously restructured one of JRRTs works into a different format.   
   >>> (I forget the details.)  The Estate successfully claimed it would   
   >>> infringe if publshed and they would not approve it.   
   >>>   
   >>> He was advised by another person who had had a similar experience to   
   >>> do what that person had done: include so much of his own material that   
   >>> the JRRT bits were "incidental" and so satisfied "fair use".   
   >>>   
   >>> OTOH, if they regularly used online libraries of /illegal copies/ of   
   >>> books, then they may have committed copyright infringement regardless   
   >>> of whether they used them for training purposes or not. At $150K per   
   >>> occurrence.   
   >>   
   >> I am getting a lot of downloads of the five software manuals and the   
   >> marketing materiel that my people and I have written over the long   
   >> years.  1,500 pages of highly technical info, I am surprised that the AI   
   >> programmers find it useful.   
   >>     https://www.winsim.com/doco.html   
   >> and   
   >>     https://www.winsim.com/newsletters.html   
   >   
   > Perhaps they find the grammar and syntax so good that it helps.   
   >   
   > Or maybe they just don't care.   
      
   Hah, the joke is on him.  Our manuals and marketing materials are   
   written in typical programmerese.   
      
   Lynn   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)