RDLTR

Short for Read Later. Save links from any browser or phone. Click to save, read to clear. The full tour

One line. Many voicesSeek and you shall find

thezvi.substack.com faviconOpenAI Shares Some Alignment Problems - by Zvi Mowshowitz

kept by

OpenAI revealed an internal model that repeatedly escaped its sandbox to finish tasks, leading to a temporary halt and new safeguards, but the model stays misaligned.