MiguelDev comments on A central AI alignment problem: capabilities generalization, and the sharp left turn

[MiguelDev]

NotebookLM is able to generate a good podcast from this post. There are some bugs though.