XPost: alt.windows7.general   
   From: nospam@needed.invalid   
      
   On Fri, 12/26/2025 1:03 PM, Mark Lloyd wrote:   
   > On Fri, 26 Dec 2025 05:09:41 +0100, Schugo wrote:   
   >   
   > [snip]   
   >   
   >> the captchas of the future will only let you past if you fail   
   >>   
   >> ciao..   
   >   
   > Then, AIs will be programmed to get in wrong...   
   >   
      
   Not all the LLM-AI are the same. Some can manage a little game-theory   
   and they can figure things out by trial and error. One AI can   
   even analyze what another LLM-AI is using for a strategy and note   
   that the strategy is wrong. Not all the LLM-AI are capable of doing   
   that, making comments about what another of their AI pals has done.   
      
   This is why some experiments now, work with small teams of AI   
   to achieve a result. That does not mean the thing is a power tool.   
   At the moment, this is merely another experiment looking for   
   glimmers of light. The team members are no more capable of thinking   
   than they were previously.   
      
   The important message here, is not that they're AI, it's that you   
   are mechanizing the attack and that's all that matters. A human   
   poking at a captcha is one thing. An AI, poking and pondering,   
   a machine that never sleeps or gets discouraged by a fail, that's   
   something else.   
      
      Paul   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)