Simon Lermen
AI alignment researcher
I write about AI safety, alignment, and the trajectory of advanced AI.
Read & subscribe on Substack →View my papers on Google Scholar →
Selected writing
-
June 2026
Where does the race to automate AI research end?
Why automating AI research could trigger rapid, unrecoverable alignment failures.
-
February 2026
Large-Scale Online Deanonymization with LLMs
How language models can identify anonymous online users with high precision.
-
January 2026
On Owning Galaxies
Why property rights may not survive an AI singularity.
-
November 2025
Will We Get Alignment by Default?
A debate on whether current alignment methods will scale to superintelligence.