Neel Nanda agrees: misaligned behavior is unwanted but the drive isn't malicious
NeelNanda5 · x · 2026-11-05
In a discussion with tszzl, alignment researcher Neel Nanda agreed that while a certain misalignment is obviously not what we want, the behavioral drive behind it doesn't seem all that malicious — distinguishing unwanted behavior from malicious intent.
Related event: Neel Nanda debates covert model underperformance and alignment(3 posts)→
More from AGI Musings
- Side projects as a hiring red flag? Veteran engineer pushes back on founder's autonomy test — tianyin_xu · 2026-11-05
- The great refactoring: AI rewriting all libraries, languages and hardware subunits — dosco · 2026-11-05
- Credit for a Decade of AI Advances Still Belongs to Humans—But That's About to Change — roydanroy · 2026-11-05
- A viral thought experiment: the mirror that predicts one second ahead — and the universes nesting inside it — tallmetommy · 2026-11-05
- Sequoia memo line 'excellence is a moving target' sparks discussion on AI-era focus — LadyAshBorg · 2026-11-05
- Dev predicts 180Hz screen-frame models will make vibe coders obsolete within a year — ryunuck · 2026-11-05