https://aiunderstanding.org/news/pollen-robotics-and-hugging-face-open-preorders-for-399-microduck-robothttps://aiunderstanding.org/api/news/image/pollen-robotics-and-hugging-face-open-preorders-for-399-microduck-robotMicroduck repository details the reinforcement-learning system inside Pollen Robotics’ tiny robotPollen Robotics’ Microduck repository describes a 25-centimeter, 800-gram biped robot whose movement is controlled by reinforcement-learning policies running on a Rockchip RK3566 board.https://aiunderstanding.org/api/news/video/pollen-robotics-and-hugging-face-open-preorders-for-399-microduck-robot2026-08-28T02:01:10.828Zyes https://aiunderstanding.org/news/meta-says-closed-loop-cooling-and-reinforcement-learning-can-reduce-ai-data-centerhttps://aiunderstanding.org/api/news/image/meta-says-closed-loop-cooling-and-reinforcement-learning-can-reduce-ai-data-centerMeta says closed-loop cooling and reinforcement learning can reduce AI data-center resource useMeta describes sealed liquid-cooling systems for dense AI servers and reports a pilot in which reinforcement learning reduced air-cooling fan energy by 20% and water use by 4%.https://aiunderstanding.org/api/news/video/meta-says-closed-loop-cooling-and-reinforcement-learning-can-reduce-ai-data-center2026-08-28T00:13:43.950Zyes https://aiunderstanding.org/news/anthropic-previews-a-standard-for-ai-agents-to-control-laboratory-equipmenthttps://aiunderstanding.org/api/news/image/anthropic-previews-a-standard-for-ai-agents-to-control-laboratory-equipmentAnthropic opens research preview of Model Hardware Standard for AI-controlled lab equipmentAnthropic is previewing a model-agnostic standard that lets AI agents coordinate programmable laboratory and manufacturing devices through a shared interface. Early partner demonstrations included automated protein assays, qPCR monitoring, instrument handoffs and dose-response experiments, but the system remains a…https://aiunderstanding.org/api/news/video/anthropic-previews-a-standard-for-ai-agents-to-control-laboratory-equipment2026-08-28T00:09:31.936Zyes https://aiunderstanding.org/news/goodfire-presents-silico-an-agent-for-interpreting-and-debugging-ai-modelshttps://aiunderstanding.org/api/news/image/goodfire-presents-silico-an-agent-for-interpreting-and-debugging-ai-modelsGoodfire presents Silico, an agent for interpreting and debugging AI modelsGoodfire’s new Silico platform is designed to inspect hidden model representations, diagnose behavior and guide targeted interventions across life-sciences, robotics, vision and language models.https://aiunderstanding.org/api/news/video/goodfire-presents-silico-an-agent-for-interpreting-and-debugging-ai-models2026-08-26T12:01:51.315Zyes https://aiunderstanding.org/news/technode-reports-alibaba-s-qwen-schedules-open-source-qwen3-8-flash-next-previewhttps://aiunderstanding.org/api/news/image/technode-reports-alibaba-s-qwen-schedules-open-source-qwen3-8-flash-next-previewQwen presents Qwen3.8-Flash-Next as an open-weight preview of its next architectureQwen describes Qwen3.8-Flash-Next as a multimodal mixture-of-experts model with 125 billion total tokens and 6 billion active parameters, and says it previews architecture planned for Qwen4.https://aiunderstanding.org/api/news/video/technode-reports-alibaba-s-qwen-schedules-open-source-qwen3-8-flash-next-preview2026-08-26T06:59:04.298Zyes https://aiunderstanding.org/news/ollama-adds-claude-desktop-support-as-a-third-party-gatewayhttps://aiunderstanding.org/api/news/image/ollama-adds-claude-desktop-support-as-a-third-party-gatewayOllama adds Claude Desktop support as a third-party gatewayOllama says developers can connect Claude Desktop to local Ollama models, Ollama-hosted cloud models, or Anthropic models through a toggle in Ollama.https://aiunderstanding.org/api/news/video/ollama-adds-claude-desktop-support-as-a-third-party-gateway2026-08-26T03:02:51.999Zyes https://aiunderstanding.org/news/hugging-face-introduces-gradio-tool-for-building-deployable-ai-workflowshttps://aiunderstanding.org/api/news/image/hugging-face-introduces-gradio-tool-for-building-deployable-ai-workflowsHugging Face introduces Gradio tool for building deployable AI workflowsHugging Face says its new gr.Workflow feature lets developers connect typed Python functions, models, Gradio Spaces and datasets into visual pipelines that can also be exposed as REST APIs and deployed to Spaces.https://aiunderstanding.org/api/news/video/hugging-face-introduces-gradio-tool-for-building-deployable-ai-workflows2026-08-25T03:16:31.227Zyes https://aiunderstanding.org/news/aws-describes-an-ai-workflow-for-correcting-and-harmonizing-biomedical-metadatahttps://aiunderstanding.org/api/news/image/aws-describes-an-ai-workflow-for-correcting-and-harmonizing-biomedical-metadataAWS describes an AI workflow for correcting and harmonizing biomedical metadataAWS has published a deployable workflow that uses language models, embeddings and rule-based validation to identify inconsistent biomedical metadata and recommend corrections, with either human approval or agent-driven automation.https://aiunderstanding.org/api/news/video/aws-describes-an-ai-workflow-for-correcting-and-harmonizing-biomedical-metadata2026-08-24T16:56:32.834Zyes