AI Infrastructure

Auto Added by WPeMatico

Google AI Releases DiffusionGemma, a 26B MoE Open Model Using Text Diffusion for Up to 4x Faster Generation

ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Language Model, Large Language Model, Machine Learning, New Releases, Open Source, Software engineering, Staff, Tech News, Technology, Uncategorized

Google AI team including the Google DeepMind researchers have just released DiffusionGemma, an experimental open model for text generation. It uses text diffusion instead of standard autoregressive decoding. The model ships under a permissive Apache 2.0 license. Google positions it for devs and researchers exploring speed-critical, interactive local workflows. Examples include in-line editing, rapid iteration, […]

Google AI Releases DiffusionGemma, a 26B MoE Open Model Using Text Diffusion for Up to 4x Faster Generation Read More »

NVIDIA cuTile Python Tutorial: Building Tiled GPU Kernels for Vector Addition, Matrix Addition, and Matrix Multiplication in Colab

ai, AI (Artificial Intelligence), AI Infrastructure, Applications, Artificial Intelligence, Editors Pick, Staff, Technology, Tutorials

In this tutorial, we implement an advanced hands-on workflow for NVIDIA cuTile Python, a tile-based GPU programming interface for writing efficient CUDA-style kernels directly in Python. We start by preparing a Colab-friendly environment, checking the available GPU, driver, CUDA, and cuTile installations before running any kernel code. We then build tiled examples for vector addition,

NVIDIA cuTile Python Tutorial: Building Tiled GPU Kernels for Vector Addition, Matrix Addition, and Matrix Multiplication in Colab Read More »

A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Paper Summary, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Software engineering, Staff, Tech News, Technology, Uncategorized

A new working research from Perplexity and Harvard offers field evidence on what AI agents do to knowledge work. It draws on production data from two Perplexity products: Search and Computer. The setup is a natural comparison. Search is a conversational answer engine. Computer is an agent that plans and executes tasks end to end.

A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search Read More »

Xiaomi MiMo and TileRT Push a 1-Trillion-Parameter Model Past 1000 Tokens Per Second on Commodity GPUs

ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Machine Learning, New Releases, Software engineering, Staff, Tech News, Technology, Uncategorized

Inference speed is becoming a competitive metric for large language models. Xiaomi’s MiMo team just released MiMo-V2.5-Pro-UltraSpeed, built in collaboration with the TileRT systems group. It decodes faster than 1000 tokens per second on a 1-trillion-parameter model. Xiaomi team describes this as a first at trillion-parameter scale. Demos show generation peaks near 1200 tokens per

Xiaomi MiMo and TileRT Push a 1-Trillion-Parameter Model Past 1000 Tokens Per Second on Commodity GPUs Read More »

Meet Harness-1: A 20B Retrieval Subagent Trained With Reinforcement Learning Inside a Stateful Search Harness on gpt-oss-20b

agentic ai, ai, AI (Artificial Intelligence), AI Agents, AI Infrastructure, AI Paper Summary, AI Shorts, Applications, Artificial Intelligence, Editors Pick, New Releases, Open Source, Software engineering, Staff, Tech News, Technology

Most search agents are trained as policies over a growing transcript. The model decides how to search. It must also remember what it saw, which evidence matters, and which claims it checked. A team of researchers from University of Illinois Urbana-Champaign, UC Berkeley, and Chroma argues this asks too much. Reinforcement learning ends up optimizing

Meet Harness-1: A 20B Retrieval Subagent Trained With Reinforcement Learning Inside a Stateful Search Harness on gpt-oss-20b Read More »

NVIDIA garak Tutorial: Build a Complete Defensive LLM Red-Teaming Workflow with Custom Probes and Detectors

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, Applications, Artificial Intelligence, Editors Pick, Security, Software engineering, Staff, Technology

In this tutorial, we analyze NVIDIA garak as a practical framework for defensive LLM red-teaming. We start by setting up Garak, then move through plugin discovery, dry runs, real-model scans, multi-probe evaluations, report analysis, custom probe creation, custom detector creation, and AVID export. Instead of running only a single scan, we use Garak end-to-end to

NVIDIA garak Tutorial: Build a Complete Defensive LLM Red-Teaming Workflow with Custom Probes and Detectors Read More »

Google DeepMind Releases Gemma 4 QAT Checkpoints: Q4_0 and a New Mobile Format Cut On-Device Memory

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, edge ai, Editors Pick, Language Model, Large Language Model, New Releases, Software engineering, Staff, Tech News, Technology

Google DeepMind released Quantization-Aware Training (QAT) checkpoints for the Gemma 4 family. The release targets local deployment on edge devices and consumer GPUs. It follows the Gemma 4 launch in April and a 12B model two days earlier. We compared the available Gemma 4 edge-model formats using only published numbers. The goal was simple. Show

Google DeepMind Releases Gemma 4 QAT Checkpoints: Q4_0 and a New Mobile Format Cut On-Device Memory Read More »

NVIDIA AI Releases Dynamo Snapshot: A CRIU-Based Fast Startup System for AI Inference on Kubernetes

ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Artificial Intelligence, Editors Pick, Language Model, Large Language Model, Machine Learning, New Releases, Software engineering, Staff, Tech News, Technology

In production inference deployments, demand fluctuates over time, requiring inference replicas to scale elastically. Cold-starting inference workloads on Kubernetes can take several minutes. During that time, GPUs are allocated but idle, generating no tokens and serving no requests. ‘Cold start’ means the full sequence a model server must complete before serving any request: pulling the

NVIDIA AI Releases Dynamo Snapshot: A CRIU-Based Fast Startup System for AI Inference on Kubernetes Read More »

Perplexity AI Introduces Hybrid Local-Server Inference Orchestrator for Personal Computer: Automatic On-Device and Cloud Task Routing

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Artificial Intelligence, Editors Pick, New Releases, Staff, Tech News, Technology

Perplexity AI announced what it calls the first hybrid local-server inference orchestrator at Computex 2026. The system is designed to automatically route AI tasks between a user’s local device and cloud-based frontier models without requiring the user to decide in advance. The feature is expected come to Perplexity Computer in July 2026. What is Hybrid

Perplexity AI Introduces Hybrid Local-Server Inference Orchestrator for Personal Computer: Automatic On-Device and Cloud Task Routing Read More »

Meet OpenJarvis: A Local-First Framework for On-Device Personal AI Agents with Tools, Memory, and Learning

ai, AI (Artificial Intelligence), AI Infrastructure, AI Paper Summary, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Language Model, Machine Learning, New Releases, Software engineering, Staff, Tech News, Technology

Researchers at Stanford University and Lambda Labs, have published the research paper for OpenJarvis, an open-source framework that runs inference, agents, memory, and learning entirely on-device. The open-weight models configured through OpenJarvis land within 3.2 percentage points of the best cloud model on average, at roughly 800× lower marginal API cost per query and roughly

Meet OpenJarvis: A Local-First Framework for On-Device Personal AI Agents with Tools, Memory, and Learning Read More »