Hacker News new | ask | show | jobs
by jxjnskkzxxhx 387 days ago
In my experience, people who understand llms better are more impressed. Not impressed like "wow so smart" but impressed like "wow can't believe that just training to predict the next token actually works so amazingly well"