Revisiting LessWrong’s AI assumptions

The critiques challenge both descriptions of AI minds and expectations about AI development. One post argues that claims of incomprehensible AI thinking are empirically wrong and serve as in-group signals. Another calls for revising LessWrong doctrine, rejecting “no fire alarm” for artificial general intelligence, commercial applications arriving only just before a rapid intelligence takeoff, and “AGI in a basement.” It is more tentative about “alien minds,” calling the idea arguable and citing “convergent representations.”