Rented GPU clusters are being handed to tenants with container escapes, unrestricted east-west networking, and shared Grafana. Coding-capable models make exploiting stale software cheap. A runnable audit CLI is available to check your own cluster.

• postAMD's inference gap is a software and tooling problem, not a silicon one
• postSemiAnalysis on Jalapeno: OpenAI's chip, measured in tokens per megawatt
• postGB300 NVL72 claimed at 7x perf-per-dollar over H200 for agentic inference
• postGLM-5.3-Flash's 100T daily tokens reportedly served on Chinese chips

• articleMost Neoclouds Suck At SecurityJordan Nanos
• article[AINews] AI Cybersecurity becomes top of mindLatent Space
• articleOpenAI’s accidental cyberattack against Hugging Face is science fiction that happenedSimon Willison

Checking sign-in…

Loading comments…