In this reflective blog post, Jeff Kaufman revisits his 2017 project on AI risk, analyzing his past thoughts and conversations with AI researchers. He discusses the evolving landscape of AI safety and critiques his initial conclusions, recognizing both the value of his inquiries and shortcomings in understanding AI's trajectory. The post highlights the ongoing debate within the AI community about the importance of addressing risks associated with superintelligence while acknowledging the complexity of predicting future developments in AI.