News for hackers. Stuff that matters.
Re:Claude Opus 5 by throw-the-towel on Saturday July 25, 2026 @11:30AM (#49048416)
But can we say that the LLMs are consciously designed? AFAIK nobody really understands why they work.
Re:Claude Opus 5 by amoss on Saturday July 25, 2026 @05:00PM (#49051509)
Whether they are or not is irrelevant to the difference- there is still an explicit division between training and inference, with reinforcement of the statistics of the training set.