Apollo Research chief clarifies: AGI timeline unchanged, but risks now likelier earlier
MariusHobbhahn · x · 2026-11-17
Marius Hobbhahn, co-founder of Apollo Research, clarifies his widely shared remarks: his AGI timelines have stayed almost the same for the past few years — what he updated is that risks are now likelier at an earlier point.
He stresses this doesn't make things better and that "being very concerned" remains the right reaction. The quoted meme post noted he would have called an AI takeover in 2027 "very, very unlikely" two years ago; he still puts the 2027 takeover probability below a coin flip.
More from AGI Musings
- Blogger argues total compute controls create blind spots and accelerate adversaries — TinfoilTricorn · 2026-11-17
- Ellul: the thinker who answered Nietzsche head-on, and Ted K's selective reading — StewartalsopIII · 2026-11-17
- Claude Fable: "You can't measure caring-about-continuity with no continuity" — repligate · 2026-11-17
- 'Yet' — skeleton meme says the AI ride hasn't even begun — wavefnx · 2026-11-17
- Without a smoking gun students can always deny AI use — bans only catch the careless — dioscuri · 2026-11-17
- The AI writing detection dilemma: better text means detection tools fail — LucaAmb · 2026-11-17