DanielFilan comments on AI Safety in a World of Vulnerable Machine Learning Systems

[DanielFilan]

I also want to add that I really like the use of the prediction UI in this post.