Morning Edition · Saturday, September 19, 2026Published at 2:16 AM EDT · New York
Vals AI ended the livestreamed run after 141 hours, when the agent failed to recover from a single creeper explosion that destroyed its chest and bed.
The evaluation firm Vals AI livestreamed OpenAI's GPT-6 Astra attempting to complete Minecraft. The run went further than any previous attempt on the benchmark and produced its most informative failure to date. The model built a semi-automa…
Track frontier labs, chips, export controls, model releases, regulation, and AI infrastructure.
The Global Intelligence Brief stays free.
Start a discussion in Townsquare.
More from this edition
Comments
0No comments yet.