x
Can Large Language Models Identify Novel Threats? Part 1: Mirror Life and the Classification Gap — LessWrong