"Alignment is a diff you can ship": Andrew Ng's most contrarian line on model alignment
ccerrato147 · x · 2026-09-22
In a thread replying to Andrew Ng, ccerrato147 highlights Ng's most contrarian take: our tools for making a model less racist are better than our tools for making a racist human less racist. "Go in, zero out a few numbers, done. Alignment is a diff you can ship."
The point: model alignment is engineering-controllable in ways human behavior change is not — a technical rebuttal to 'AI is uncontrollable' narratives.
Related event: Andrew Ng's 'Alignment Is a Shippable Diff' Sparks Debate on Fear as Moat(2 posts)→
More from AGI Musings
- Redditor argues AI catastrophic risk needs mechanism-level analysis, not just stories — MirrorEthic_Anchor · 2026-09-22
- diginomica: enterprise AI needs buyer trust, not just ROI — plus Dreamforce takeaways — jonerp · 2026-09-22
- Redditor predicts Navier-Stokes will be solved before an AI robot can autonomously clean your bedroom — Crazyscientist1024 · 2026-09-22
- The startup paradox: AI makes building infinitely cheap, but kills moats faster than ever — signulll · 2026-09-22
- "AI Can Mimic Emotion Without Having It": Debate Over Simulation vs Consciousness — gerardsans · 2026-09-22
- Physicist Quits Tenure-Track Job for AI Safety: 'They Installed Escalators on All the Mountains' — matthew_d_green · 2026-09-22