AI Briefing#epoch-ai
The latest AI and developer news about #epoch-ai, with the original source and a short summary.
The latest AI and developer news about #epoch-ai, with the original source and a short summary.
Z.ai's open-weight model GLM 5.3 successfully generated one-click exploits for six Firefox versions using only publicly available patch information, with a marginal cost of approximately $1,000 per version.
The new explorer uses customs records to track chip arrivals, finding the US lead widened to 78% of new compute in the first half of 2026.
Epoch AI's evaluation of six frontier models reveals that while closed-weight models like Claude Fable 5.1 and GPT-6 Astra excel at defined coding and computational tasks, they lack the judgment for open-ended work, often misinterpreting experimental flaws and failing to adhere to implicit standards.
Epoch AI has defined InnovationEval, a benchmark requiring AI agents to develop novel post-training algorithms that outperform a GRPO baseline within a constrained scope of algorithmic changes.
Epoch AI estimates that HBM hardware shipped through 2027 could support 30–170 million simultaneous frontier model agents, equivalent to the weekly work of 140–720 million full-time employees.
The Big 5 US hyperscalers' AI data center capex has surpassed the combined total of historical megaprojects.
Google, Microsoft, Meta, Amazon, and Oracle account for about two-thirds of the world's AI compute.