OpenAI’s Astra rumors revive the question: how soon could strong AGI arrive?
cuttheblue · reddit · 2026-08-04
The poster asks for a realistic timeline to strong AGI in light of an alleged OpenAI Astra model that reportedly made progress on several hard math, quantum, and cryptography problems.
- The thread defines strong AGI as systems that could do the hardest intellectual jobs, like top CEO or military commander work.
- It notes that progress appears strongest in math and coding, which could accelerate recursive self-improvement.
- The author says the headlines may be overstating the result: Astra reportedly solved some problems and merely advanced on others, while also failing on others like Millennium Prize problems.
- The core question is whether top jobs vanish in months or years, given how surprising recent model jumps have been.
More from AGI Musings
- Quanta traces attention mechanism origins to Bahdanau's translation intuition — CatAstro_Piyush · 2026-10-16
- Why you shouldn't press the 'replace this universe' button: an alignment debate — jd_pressman · 2026-10-16
- Meta scales back AI agent plans as staff clean up after the machines — SumitGup · 2026-10-16
- RFdiffusion3 debuts in Science: generative AI designs proteins and nucleic acids atom by atom — AllThingsApx · 2026-10-16
- Researcher: 'Guru-like' rhetoric fuels claims that AI safety is a cult — jd_pressman · 2026-10-16
- Comparing AI data centers to 1900s power buildout: the anti-DC craze will look silly in a decade — GabGarrett · 2026-10-16