Hugging Face announced Inkling 2 Mini, a small model offering cutting-edge performance for local setups. It targets on-device/local deployment.
AWS published a distributed KV caching agent skill to reduce cost and latency for self-hosted LLM deployments. It is packaged as a reusable agent skill.
DeepSeek's deepseek-harness repository shows 285,027 stars and an update on 2026-11-21, indicating a very large open-source footprint for the harness release.
AWS announced Amazon Textract Insights, adding verifiable generative AI extraction for documents. It couples generative outputs with verification against source documents.
AWS introduced a SageMaker HyperPod Global Inference Gateway providing a single endpoint across regions and clusters. It simplifies multi-region inference routing.
Google launched a new search homepage with AI chat history and no search button, adding an AI sidebar. It is a significant UX shift toward conversational search.
Meta is reconsidering its 2030 net-zero emissions goal as AI data centers drive record emissions. The reversal reflects the energy cost of scaling AI infrastructure.
OpenAI published work on defending models against adversarial distillation, where third parties extract model capabilities via distillation. The post details mitigation approaches.
OpenAI released MedChemBench, a benchmark evaluating whether AI can help scientists identify promising drug candidates. It is a new evaluation resource for medicinal chemistry tasks.