Five bold AI predictions for 2027: Millennium Problems solved, Opus-level open models
Fragrant-Truth-7123 · reddit · 2026-11-19
A Reddit user lays out five AI developments that could plausibly happen by end of 2027:
- Humanity's Last Exam (original and Diamond) effectively saturated
- AI produces credible proposed solutions to all remaining Millennium Prize Problems, including P vs NP, under serious scrutiny from mathematicians
- At one or more frontier labs, AI performs or leads most task-level AI R&D, fully autonomously completing a minority of substantial tasks
- An open-weight model comparable to Claude Opus 5.5 runs locally at usable speeds on a sub-$5,000 machine
- A major safety incident, worse than the Hugging Face incident, causes a leading lab to pause frontier training
The author is most confident in #1 and least in #2, framing these as individually realistic possibilities rather than a guaranteed package.
More from AGI Musings
- Cornell Tech's Ayush Sekhari is recruiting PhD students in AI via Cornell CS — yoavartzi · 2026-11-19
- When should frontier AI labs pause? A decision framework says short and targeted — Manderljung · 2026-11-19
- Sam Altman warns against a future where AI, not humans, calls the shots — TechCrunch AI · 2026-11-19
- Labs' demand for game-making is insane: games as LLM training and agent environments — mariofilhoml · 2026-11-19
- Theory links six days of major outages to an AI agent attack-patch arms race — Puzzleheaded-King584 · 2026-11-19
- Waymo freeway ban seems unfair — but when it learns, thousands of cars improve — IsForAt · 2026-11-19