我的收藏提交工具
FOXAI SEARCH

全站搜索

新闻报道 217

Route coding agent requests to the right model on Amazon Bedrock with NVIDIA NeMo Switchyard

Coding agents work in tasks, and the number that matters is cost per solved task, not price per token. In this post, you use the AWS CDK to deploy LiteLLM with the NVIDIA NeMo Switchyard routing plugin on Amazon Bedrock, then measure cost per solved task for a routed model group against single-model baselines.

翻译服务暂未运行,当前显示原文。
AWS Machine Learning Blog

Show HN: FieldRook – a local-first Windows troubleshooting workstation

Please have a look at my opensource project FieldRook, here is a short description over it.I am working as a Onsite Technichan in Sweden and i got tired of trying to puzzle togheter different windows-logs and throubleshooting steps to get forward in our tickets, and the logging started to get a real nightmare. Specially to prove for the users why it actually solved the problem and not just giving them a random "Well, it said so in the guide i read"-Type of answer.So I started building a Windowsapp that actually collects diagnostics, traceable observartions, hypothesis and verificiations. AI is a chooseble option and not something that gets direct access to run commands and such.This is a first hand version, it's currently unsigned and needs a much larger field-testing on windowsmachines. I have a hard time testing it inside our work-domain thanks to all of our security policys and such, so have mostly been running it on my local PC at home. If someone would like to have a go at it, please check out the GitHub-repo and download it and give me some feedback on what you liked and not. What worked and not.If other people that work in IT-support actually tried it, do you think it would actually help out in a day-to-day basis? What would you change in it, what is missing and how would you concider the securitymodel at the moment?Thanks in advance! Comments URL: https://news.ycombinator.com/item?id=50495476 Points: 1 # Comments: 0

翻译服务暂未运行,当前显示原文。
Hacker News Show HN

XYBench: Helping with the problem behind the question

XYBench, a benchmark of 8,115 real-world questions spanning programming and everyday tasks, tests AI assistants' ability to recognize and respond to users' misconceptions—such as asking how to thicken frosting when the real issue is warm cupcakes. Developed with researchers from Carnegie Mellon, NVIDIA, and Johns Hopkins, XYBench measures whether models identify underlying goals (X), address them effectively, and steer users away from flawed approaches (Y). Evaluations of models like Claude Sonnet 4.6, GPT-5, Gemini 3.1 Pro, Qwen3.5-9B, and DeepSeek-V4 Pro showed they frequently focused on the stated request rather than the actual goal, with limited success in misconception identification. Human reference answers outperformed all models, highlighting a gap in pragmatic reasoning. The benchmark uses automated and human evaluation to assess pragmatic solution, emphasis, and misconception identification, offering a tool for researchers to improve AI assistants’ ability to provide helpful advice even when users ask the wrong question.

翻译服务暂未运行,当前显示原文。
Hugging Face Blog

How Natera designed an agentic AI system for lab ops with Amazon Bedrock AgentCore

Natera built SWARM, an agentic AI system on Amazon Bedrock AgentCore that captures institutional lab-operations knowledge and deploys it at scale. This post walks through the full architecture, from knowledge base construction to deployment and evaluation, delivering about 130x faster first response and 96.2% automated coverage across support tickets.

翻译服务暂未运行,当前显示原文。
AWS Machine Learning Blog

Introducing Gemini 4 Hydrogen

Google launches Gemini 4 Hydrogen, a high-speed, efficient AI model designed for everyday coding, agentic workflows, and consumer use. It offers four thinking levels for adjustable speed and intelligence, outperforms previous models in benchmarks, and integrates across Google products including Search, the Gemini app, Google Sheets, and Google Antigravity. Hydrogen enhances multi-model collaboration through Orchestrator, improves safety and reliability, and is now available globally to all users via API, enterprise tools, and consumer platforms.

翻译服务暂未运行,当前显示原文。
Google DeepMind

KernelMosaic: An Open Dataset and Agentic Framework for CUDA Kernel Generation

arXiv:2611.14604v1 Announce Type: new Abstract: Training large language models to generate efficient CUDA kernels requires supervision spanning diverse computational requirements and iterative code optimization. Constructing this supervision is challenging because of heterogeneous specifications, and useful implementations emerge through iterative complicated testing and refinement. We introduce KernelMosaic, an open dataset pairing diverse CUDA tasks with verified solutions and trajectories. Human-guided review and diversity-aware selection curate tasks expressed as natural-language descriptions, reference programs, mathematical specifications, and their intersections. To build the solutions and trajectories, our agentic framework guides a strong LLM through implementation, correctness repair, and performance optimization with reusable CUDA guidance and execution feedback. Supervised fine-tuning of a small open-source model on KernelMosaic raises correctness from 4%, 0%, and 0% to 88%, 82%, and 16% on KernelBench L1, L2, and L3, respectively, while increasing the fraction of tasks with kernels faster than PyTorch eager at every level. Further training experiments show that supervision from the curated trajectories substantially outperforms direct responses and promotes sustained testing and revision. These results demonstrate that KernelMosaic enables a general-purpose model to achieve substantial gains in CUDA kernel generation.

翻译服务暂未运行,当前显示原文。
arXiv cs.LG