List of AI safety challenges and meta-level ideas for addressing some

Let me know what “challenges of today” you’d add to the list. I put this together recently and wrote it up and shared it with a few friends.