Google DeepMind Pilots Double-Blind AI Evaluations
Google DeepMind disclosed it is piloting what it describes as the world's first double-blind AI evaluations . Details on methodology, participating models, and results were not available in the initial disclosure.…
NVIDIA Dynamo Adds Qwen 2.4T Serving on GB300, GB200
NVIDIA's AI Dynamo project released v1.4.0-qwen-3.8-2.4t-dev.1, an experimental snapshot build adding serving support for Qwen/Qwen3.8-2.4T-A95B-FP8 on both vLLM and SGLang backends . The build, committed August 27, includes Kubernetes…
Databricks Previews Lakebase, Streaming Research for VLDB 2026
Databricks announced it will present research on Lakebase, streaming, and lakehouse innovations at the VLDB 2026 conference . The presentations cover multiple technical developments underpinning the Databricks platform .…
Red Hat AI Inference 3.5 Ships Priority Queuing for Shared GPU Pools
Red Hat AI Inference 3.5 includes a generally available flow-control feature in llm-d that adds priority-aware admission and tenant fairness scheduling for GPUs, according to a Red Hat Developer post . The capability allows platform teams…
ALL BRIEFS
RANKED · 5 THIS DAYHarness layer splits from model as AI agents' real composition boundary
The agentic harness, the execution loop that handles context, tool wiring, memory, and human approval, is separating from the model itself to become the primary surface practitioners build, govern, and swap. OpenAI's open-sourcing of Codex…