Episode Details

OpenAI details an agent security incident, makes the case for custom inference silicon, and Nvidia is reportedly pursuing Hugging Face. Plus: why automated fact-checkers need cross-domain tests.

Covered today:
- OpenAI’s account of the Hugging Face incident and its security response
- OpenAI’s Jalapeño custom inference chip results
- Research on cross-benchmark robustness in automated fact-checking
- Reported Nvidia acquisition of Hugging Face
- IBM Granite 4.2 open-weight models
- Qwen3.8-Flash-Next
- Nvidia NVLink Fusion and NVHBM memory

Source links:
- OpenAI, Hugging Face incident: https://openai.com/index/hugging-face-incident-and-the-road-ahead
- OpenAI, Jalapeño: https://openai.com/index/jalapeno-first-results
- arXiv, automated fact-checking evaluation: https://arxiv.org/abs/2608.25934v1
- The Decoder, Nvidia and Hugging Face: https://the-decoder.com/nvidia-snaps-up-hugging-face-for-12-9-billion-as-closed-ai-labs-pull-away/
- Ars Technica, IBM Granite 4.2: https://arstechnica.com/ai/2026/08/ibms-new-granite-4-2-models-ride-the-wave-of-interest-in-local-llms/
- Simon Willison, Qwen3.8-Flash-Next: https://simonwillison.net/2026/Aug/26/qwen38-flash-next/
- Nvidia, NVLink Fusion and NVHBM: https://blogs.nvidia.com/blog/nvlink-fusion-nvhbm-custom-high-bandwidth-memory/