Build a self-service LLM gateway with LiteLLM on AWS Fargate that puts Amazon Bedrock foundation models behind an OpenAI-compatible API. Developers get instant access through a single API key, while platform teams gain per-user cost controls, quota isolation, and observability.
Your Amazon Bedrock inference bill can keep climbing even after you tune prompts and application logic. Four request-level choices (cache duration, service tier, reasoning effort, and maxTokens) change how many tokens the service processes and the rate applied. We show how we tested each and how to verify the effect.
Article URL: https://github.com/erdemozkan/kiln Comments URL: https://news.ycombinator.com/item?id=50474384 Points: 3 # Comments: 0
Hey guys. I recently watched the latency reports for my local voice assistant, which show a waveform plus silero probability for voice activity. I noticed silero lags visibly on the speech off signal.There had to be a method to get this signal closer to what I felt is optimal. My idea was to use a 10 ms energy detector with a threshold that adapts to environment or room noise. Which worked amazingly well. It created more false positives but in my speculative generation pipeline that would not matter. If it's wrong and the user continues speaking I can abort the pipeline fast.I benchmarked my solution over ~3200 turns from my assistant logs and it came out ~110 ms faster initially. Still 100 ms now after some a fix which was needed for quiet speech.I made it public if you want to have a look. If you measure it on your data I really appreciate any feedback. There's a benchmark script in the repo.Thx for reading, have a good day folks. Comments URL: https://news.ycombinator.com/item?id=50474586 Points: 2 # Comments: 0
Every generative AI team keeps rebuilding the same components: RAG pipelines, document processors, and agent scaffolding. This post shows how to access a curated catalog of 23 production-ready, open source reusable assets from the AWS Generative AI Innovation Center, with real delivery results from DoorDash, Lyft, and Ricoh.
英国高等教育监管机构学生办公室（OfS）发布报告，指出当前AI检测工具难以准确判断学生是否使用人工智能完成作业，建议高校减少无监督考核，转向监督式评估。报告强调，随着生成式AI能力提升，识别AI内容愈发困难，应通过过程性评估、口试和笔试等方式重构评价体系。同时，报告警示AI使用可能加剧教育不平等，并影响学术诚信与残障学生权益。全球多地已采取限制措施，但技术应用与教育公平间的平衡仍待探索。
豆包工作于11月12日更新，新增Kimi K3.1Pro、GLM5.3、DeepSeek V4.1Flash等第三方大模型接入，并在模型选择页面标注额度消耗倍率，实现成本透明化。平台同步推出限时优惠活动，部分模型在指定时段使用可享额度折扣，且豆包2.1Lite与DeepSeek V4.1Flash的消耗倍率下调，提升额度使用效率。新用户登录电脑版可免费领取30天订阅权益。此次更新标志着AI办公工具从封闭模型向开放市场演进，强调任务调度灵活性与单位成本效率。
Anthropic Fellows Program 团队发布 MAMUBench 基准，测试显示在多用户多智能体场景下，由单一智能体统筹的协调者模式表现全面优于各用户独立分配智能体的点对点团队模式。研究揭示了多智能体团队存在行动迟缓、越权干涉、信息捏造及竞态条件等问题，提出通过明确责任人、设定优先级以及引入平台级防护机制可有效缓解协作失灵，强调未来需从平台设计与智能体训练层面构建更安全高效的多智能体协作生态。
Anthropic 于11月12日为 Claude API 添加了 Beta 功能，允许开发者在消息流中随时调整 model effort 参数，动态控制模型回答的深度与资源消耗。通过插入空白 system 消息并设置 output_config.effort，可在不重开会话或丢失缓存的前提下切换档位，新设置从下一条用户消息开始生效，不影响已生成内容。该功能使模型能根据对话阶段灵活调整思考深度、工具调用频率和输出复杂度，适用于长链路智能体场景，实现算力与成本的精细化管理。
Anthropic宣布对Claude的交互式连接器进行重大升级，支持用户在对话旁直接打开并全屏操作工具，实现AI实时修改内容、并排工作流及免提示词调用。首批支持Asana、Canva、Confluence、Notion等主流工具，Figma、Harvey和Shopify正在开发集成。该功能基于MCP Apps构建，面向开发者开放，企业管理员可灵活控制权限。
