Sam Altman declares the takeoff begun: novel insights by 2026, robots by 2027
itsOmSarraf_ · x · 2026-09-29
In his essay "The Gentle Singularity," Sam Altman argues humanity has passed the event horizon and takeoff has started—yet it feels far tamer than expected: no street robots, disease unsolved.
Key claims:
- We've built systems smarter than humans in many ways; the hardest scientific breakthroughs (those behind GPT-4 and o3) are done and will carry us far.
- ChatGPT is already more powerful than any human who has lived, relied on by hundreds of millions—small misalignments multiplied across users carry huge downside risk.
- Timeline: 2025 brings agents doing real cognitive work (coding is changed forever); 2026 likely brings systems generating novel insights; 2027 possibly robots doing real-world tasks.
- More people will create software and art, but experts embracing AI will still vastly outperform novices.
- The biggest gains come from AI accelerating scientific progress and productivity.
More from AGI Musings
- KatjaGrace: Pausing frontier AI is right when needed, don't take AI CEOs' word on it — KatjaGrace · 2026-09-29
- KatjaGrace asks whether the China-race narrative morally justifies risky AI development — KatjaGrace · 2026-09-29
- The math on doomer borrowing: why 50% extinction credence doesn't justify borrowing to spend — trevposts · 2026-09-29
- signulll: VC-backed founders should be 100x more ambitious and risk-seeking than they think — signulll · 2026-09-29
- Microsoft AI CEO Mustafa Suleyman on rising AI alarm and his human-first approach — shiringhaffary · 2026-09-29
- Steven Pinker's open letter: why I'm not an AI doomer — Usual_Program_7167 · 2026-09-29