Does current evidence favor a ‘normal technology’ view of AI?

The original post argues that AI currently looks labor-augmenting, with minimal employment effects and agent incidents whose costs were comparable to or below those of high-profile cyber breaches. Its author says catastrophic-risk arguments must rely on future capabilities or forms of adoption, while acknowledging those possibilities warrant vigilance and safety work. Another user shared the argument with a call to listen to empiricists rather than rationalists.