Altman: Alignment isn't just an engineering problem — we must solve the science
tszzl · x · 2026-09-30
Sam Altman argues it's incorrect to treat alignment purely as an engineering problem. While acknowledging the engineering work needed — better sandboxes and monitoring tools — he says that if we're going to create models far smarter than us, we ultimately have to solve the science of alignment.
More from AGI Musings
- India's 6M IT jobs under AI pressure: TCS cuts 20,000 as hiring lags revenue — mixtapedmonk · 2026-09-30
- Blogger: When AI takes the job, you get the 40 hours back — that's what nobody is pricing in — rand_longevity · 2026-09-30
- Ken Griffin gifts over $3 billion to Carnegie Mellon, the largest single donation in US higher education history — vaibhavbetter · 2026-09-30
- Blogger predicts superintelligence hits intelligence's limit by 2029, fully human-understandable — davidpattersonx · 2026-09-30
- AI is like a microwave: a video take on AI adoption in healthcare — bondegezou · 2026-09-30
- Oxford philosopher Carissa Véliz: using AI to predict is like reading animal entrails — CarissaVeliz · 2026-09-30