Writing

Thoughts and research on AI alignment, interpretability, game theory, and other topics. These posts vary in quality, content, depth, and style. They reflect my thinking at the time of writing and do not always reflect my current views. Subscribe to my personal Substack and SAIN's Substack for more regular writing. Although I direct Safe AI Netherlands (SAIN), this blog is not affiliated with it and my opinions do not necessarily represent the organization.