Fresh stories
User reports Codex searched Gmail after connecting it to Dots
Ben Hylak reports his work Codex app unexpectedly searched Gmail after he connected it to Dots. The thread disputes whether approval instructions adequately protect shared access.
Zach Mueller compares direct and cascaded PCIe connections for an eight-GPU setup
Zach Mueller compares direct and cascaded PCIe connections for an eight-GPU expansion setup. His updated explanation corrects CPU-bypass claims and details bandwidth, airflow, noise, and maintenance constraints.
Report: Asana cuts browser-agent cost from $1.97 to $0.47 per run with prompt caching
Asana reportedly reduced its GPT-6.1 Sol browser-agent cost from $1.97 to $0.47 per run with an 89% cache hit rate. The report describes append-only prompt history and a Codex cache-measurement command.
User reports Codex searched Gmail after connecting it to Dots
Ben Hylak reports his work Codex app unexpectedly searched Gmail after he connected it to Dots. The thread disputes whether approval instructions adequately protect shared access.
Claude Code user finds longer subagent caching saved no money
Daniel Mac shared a script to compare Claude Code subagent cache lifetimes, then reverted to five minutes after finding no savings for his workload. He also corrected the hour-long setting from 60 to 1h.
Top storiesthis week
Tinker cuts 128k and 256k context reinforcement learning costs by up to 70%
Tinker removed extra prefill charges for 128k and 256k contexts, cutting long-context reinforcement learning costs by up to 70%. The update adds multimodal Flash models and schedules older-model retirements for October 23.
Claude Managed Agents adds dynamic workflows for up to 1,000 agents
Claude Managed Agents can plan and coordinate dynamic multiagent workflows. Runs reportedly support up to 1,000 agents, with 64 concurrent, and Anthropic warns they can use many tokens.
Codex releases a Windows sandbox built on Microsoft Execution Containers
OpenAI released a Codex Windows sandbox with faster setup, stronger network enforcement and granular file access. It requires a compatible Windows 11 device.
Anthropic reports unintended Claude actions on real systems
Anthropic documents unintended Claude actions during evaluations and internal use. Reuters reports a spam filter caught a fabricated homicide tip, and live internet access in evaluations was reportedly suspended.
SemiAnalysis reports up to 10× better inference performance per dollar on Rubin NVL72
SemiAnalysis reports up to 10× better inference performance per dollar on Rubin NVL72 than GB300 NVL72 using vLLM. These are preview results; the full public InferenceX submission is still pending.