Astra makes a weird dashboard mistake at just 44% context usage
eigenron · x · 2026-09-09
eigenron shares a screenshot of Astra making a weird mistake on a dashboard with only 44% context used — showing the failure isn't caused by context exhaustion but is an inherent capability issue.
More from Models
- Gaperon paper shows late test-set contamination recovers benchmark scores without hurting generation — burny_tech · 2026-09-09
- OpenAI's Navier–Stokes run: train-while-deploying and 10,000 coordinated agents — DataLearnerAI · 2026-09-09
- GPT-6 Astra rumored to hit ~11.6-hour p80 task horizon, tracking AI-2027 curve — haider1 · 2026-09-09
- NVIDIA open-sources gold-medal IMO system Nemotron with models, datasets and 200 new problems — kuchaev · 2026-09-09
- Two years from o1-preview to superhuman math: RL scaling now cracks open research problems — jam3scampbell · 2026-09-09
- Meta's AI assistant outperforms Instinct and ChatGPT Work on a 5-task test, claims reviewer — garrytan · 2026-09-09