The blog post critiques the current state of AI safety efforts, arguing that they prioritize theoretical discussions over real-world applications, leading to significant harm. It contrasts AI safety with established safety practices in industries like automotive and healthcare, highlighting that AI safety is often treated as a cost center rather than a primary focus. The author calls for real engineering solutions to mitigate harms caused by AI deployments, emphasizing the urgent need for effective safety measures in light of recent tragedies linked to AI usage.