Urgent.News

What's breaking now, across thousands of outlets.

AI

Nvidia PAIR enables local AI cluster construction

Nvidia has introduced a free software tool that allows users to create AI inferencing clusters using separate computers on a single network. This utility provides a unified interface to manage distributed hardware for local processing. The software supports multiple operating systems and focuses on privacy and efficiency. Distributed Hardware Connectivity and Integration The new software is…

Nvidia has unveiled a free software tool called the Nvidia Personal AI Router (PAIR) that enables users to construct local AI inferencing clusters using separate computers on a single network. This tool provides a unified interface for managing distributed hardware, focusing on privacy and efficiency by keeping sensitive data within the user's private environment.

Available as a beta release, PAIR supports multiple operating systems, including Windows, macOS, and Linux, making it a versatile option for diverse hardware environments. The primary function of PAIR is to facilitate AI inferencing workloads, reducing the time required to complete tasks by using multiple machines to handle complex data processing.

The software directs tasks to run in parallel across the available hardware, leveraging Nvidia GPUs to accelerate artificial intelligence calculations.

Brief written by urgent.news from Dev.to's own syndicated text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

Demystifying LLM Context Windows: How AI Memory Works (and Why It Fails)

Imagine asking an AI coding assistant to help refactor a complex application. At first, it gives sharp, accurate responses.

  • Context window limits LLMs' active memory during interactions
  • Tokenization and embeddings create numerical representations of text
  • Context window complexity scales quadratically with length

SWE-Gate: Why Passing Tests Isn't Enough for Agent-Generated Code

Coding agents pass tests but fail code review. Repository-level benchmarks measure test passage but ignore review acceptance criteria. This is the blind spot in every benchmark from SWE-bench onward.

  • SWE-Gate introduces a second evaluation gate beyond test success to assess review compliance.
  • 34% of agent-generated patches failed review constraints despite passing functional tests.
  • SWE-Gate separates functional correctness evaluation from constraint compliance evaluation.

Anthropic Drops Claude Sonnet 4.5: Extended Thinking Hits 200K Context for Agentic Coding

Anthropic just shipped Claude Sonnet 4.5, and the headline number isn't the benchmark score, it's the 200K token context window combined with what they're calling "extended thinking mode." For…

  • Claude Sonnet 4.5 introduces 200K token context window for agentic coding.
  • Extended thinking mode enables interleaving reasoning with action in tasks.
  • Benchmark shows 30% reduction in hallucinations and 77.2% GitHub issue resolution.

More from Saturday 5 September →