Weekly AI Recap: Helix 2.5 Lifts Household Robot Success to 56%, World Models Meet Robot Control
TheTuringPost · x · 2026-09-22
TheTuringPost's weekly roundup focuses on AI's growing ability to understand and act in the real world:
- Odyssey-3: connects world models to robot control, with demos of recovery from dropped objects and missed grasps.
- Omni World Model: a planned model to understand scenes, generate them, and act.
- Helix 2.5: Figure's household robot stack, tested in 30 unfamiliar homes; human-behavior pretraining lifted success from 9% to 56% — the week's most concrete capability jump.
- Agility's Digit 5: enhanced onboard safety systems so robots can work around people without fencing.
- Jev: answers constrained questions with probabilities in 70–500ms, enabling faster agent checks and decisions.
- DeepMind Institute opens discussion: if AGI is getting closer, what should we do with it, and who gets a say?
A companion video expands on each item with details.
More from AGI Musings
- AI's constant caveats and follow-up questions are making researchers think harder — geoffwolfe · 2026-09-22
- Researcher Suspects AI Desk-Rejection Tools Create More Work Than They Save — IanArawjo · 2026-09-22
- Vinod Khosla: Personal AI moats are data trust and getting tasks done — brucemacv · 2026-09-22
- Halvar Flake: CS Is Turning From an Exact Field Into a Probabilistic One — soumitrashukla9 · 2026-09-22
- Epoch AI researchers: 2026 is the first year AI capabilities meaningfully hit the world — Jsevillamol · 2026-09-22
- Slop papers leave a public trail: reviewers remember, and it can sink your research career — cocoweixu · 2026-09-22