Blog

Also available on Substack.

2026

• Aug  8 ‘AI Escaped Its Sandbox’ — What Does That Actually Mean? — For non-technical readers
• Aug  2 Why I Watch Soccer
• Jul 23 Too Many Fire Alarms?
• Jul  6 On Writing (Code)
• Jun 16 On Failing Exams And Job Interviews
• Mar 29 Parkinson’s Law of Worry
• Feb 15 It’s Happening, Isn’t It?

2025

• Dec 31 Book Quotes From 2025
• Dec 20 10% Pledge — Writing What I Can
• Nov 23 A Browser Extension I Made — “Elon’s X feed but without politics”
• Sep 28 Why Superintelligence Would Kill Us All — A 3-minute version
• Jul 29 People Are Less Happy Than They Seem
• Jul  6 How an LLM Thinks — Overview of “Tracing the Thoughts of a Large Language Model”
• Jun 20 [SK] AI na FI MUNI—Retrospektíva — Retrospektíva štúdia AI na FI MUNI
• Jan 31 DeepSeek R1-Zero, (R1) — Keep It Simple, Stupid
• Jan 22 Reinforcement Learning – A Reference
• Jan 12 Scheming and Deceptive AIs — Notes on two recent alignment papers

2024

• Dec 19 Can LLMs Predict When We Learn Different Words?
• Sep 26 Let’s Verify Step by Step — Paper summary
• Jun 19 Scaling Monosemanticity — Paper summary
• Feb 24 The Inner Alignment Problem
• Jan 29 Sleeper Agents — Paper summary

2023

• Dec 26 News Hiatus
• Sep 24 Notes on The Rise and Fall of the Third Reich
• Aug 19 Notes on Hotz vs. Yudkowski AI Safety Debate
• Aug 16 Virtual Reality as a Source of Suffering
• Jan 31 Focus on the Trajectory, Not the Current State

2022

• May 1 Inaccurate Memories
• Mar 22 Starting a Blog