mark_k: AI alignment is meaningless unless it means doing exactly what the user intends

mark_k · x · 2026-09-14

markk argues that AI alignment is "bullshit" unless it is defined as doing exactly what the user intends — a pointed take that reframes alignment around instruction-following rather than abstract value alignment, likely to spark debate.

Original post →

More from AGI Musings

AGI Musings channel →