Inkhaven

Hi reader! I will be collecting all my posts for the Inkhaven Residency in November 2025 on this page. The posts will be distributed across my Substack newsletter, longform posts on my X profile, the SPYLab blog, this website, and wherever else is fitting.

Recent posts:

• Things I hated at first but now like and also things you should just try (Nov 15)
• Obvious ways RL can fail (Nov 14)
• Inkhaven’s first retraction (Nov 13)
• How well AIs do AI research is not that important right now (Nov 12)
• Travel checklist (Nov 11)
• Immortality and reproduction are incompatible in the long term future (Nov 10)
• The two types of LLM preferences (Nov 9)
• Prioritize the base of Maslow’s hierarchy (Nov 8)
• Slow takeoff for AI R&D, fast takeoff for everything else (Nov 7)
• October 2025 AI safety news: Adaptive attacks, Tokenization, Impossible tasks (Nov 6)
• You should publish your Stack (Nov 5)
• You are going to get priced out of the best AI coding tools (Nov 4)
• Some hidden time sinks of the writing process (Nov 3)
• Why I am writing 30 posts in 30 days (Nov 2)
• A/B testing could lead LLMs to retain users instead of helping them (Nov 1)