AI News

China-Linked UNC3569 Exploited Sogou Input Method Flaw to Deploy GRAYRABBIT Backdoor (thehackernews.com)

One Biosciences reçoit un financement de 3,3 millions d'euros dans le cadre d'i-Démo 5 de ... (www.joplinglobe.com)

August Was World's Joint-Hottest Month On Record, Scientists Say (news.slashdot.org)

Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages (www.marktechpost.com)

One Biosciences reçoit un financement de 3,3 millions d'euros dans le cadre d'i-Démo 5 de ... (fr.finance.yahoo.com)

Conti ransomware gang member sentenced to 4 years in prison (www.bleepingcomputer.com)

DeepSeek v4.1 Flash Uncensored (huggingface.co)

PaperCut Replaces Emergency Patches With Fixes for Two Actively Exploited Flaws (thehackernews.com)

One Biosciences reçoit un financement de 3,3 millions d'euros dans le cadre d'i-Démo 5 de ... (www.businesswire.com)

Sakana AI Launches Fugu Max and Fugu Ultra v2 for Cheaper, Stronger Multi-Agent Orchestration (www.marktechpost.com)

Astra for Coding: Why Are We Doing This Again? (lucumr.pocoo.org)

Cisco FMC Flaws Exploited to Steal Credentials and Deploy Qilin Ransomware (thehackernews.com)

Anthropic blocks 'malicious use' of AI that could develop biological weapons (www.bbc.com)

Google Research Releases ToolGrad: Answer-First Framework Hits 99.8% Pass Rate for Tool-Use Data Generation (www.marktechpost.com)

GPT‑Live‑1 in the API (openai.com)

We Replaced MMAP with Io_uring in Our Rust Query Engine. It Got Slower (www.conviva.ai)

Water and ethics: AI and data centre concerns growing in Metro Vancouver - CTV News (www.ctvnews.ca)

Veille quotidienne — 11/09/2026

The Gemini app is now available for Windows (blog.google)

AI cybersecurity is a cat and mouse game (stackoverflow.blog)

LiveGrid (www.producthunt.com)

MetroLLM-Bench: Evaluating Language Models as Transit Kiosk Runtimes (arxiv.org)

RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems (arxiv.org)

Are We Really Doing Few-Shot Learning? A Critical Examination of Pre-Training Assumptions (arxiv.org)

When Synthetic Data Hurts: On Catastrophic Forgetting in Skill Retrieval for LLM Agents (arxiv.org)

RobustSGPO: Search-Space Control for Agent Harness Evolution (arxiv.org)

Building py-kvcache: A Performance Characterization of External KV Caching for vLLM with NVMe SSDs (arxiv.org)

BRACE: Anchored Bellman-Residual Correction for Stale Critics in Asynchronous RL (arxiv.org)

Dynamic language model representations for multi-objective reaction optimisation (arxiv.org)

Fidelity-Aware Scheduling of Quantum Circuits on Multi-QPU Systems (arxiv.org)

Forward-Free LLM Depth Pruning via Weight Redundancy (arxiv.org)

Learnware and AI Model Management System (arxiv.org)

Subagents vs Agent Skills: Executing Reusable Knowledge for Long-Horizon Agentic Tasks (arxiv.org)

AdamX: Cosine similarity meets gradient descent (arxiv.org)

What Makes Adversarial Examples Transfer Across Deepfake Detectors? (arxiv.org)

Project Qualia: Recovering Experiential Music Structure from Session Co-occurrence Data (arxiv.org)

Improving Cross-Lingual Token Representations by Adding a Pinch of SALT (arxiv.org)

CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search (arxiv.org)

SIRF: A Spec-Internalized Risk Foundation Model for Industrial Content Risk Control (arxiv.org)

CausalArena: Benchmarking Causal Discovery in the Foundation Model Era (arxiv.org)

Thinking with Looped Flows (arxiv.org)

Agent-Based ML-LLM Fusion with Self-Optimizing Prompts for Plateau Weather Alerts (arxiv.org)

An Empirical Measurement of Jailbreaking Evaluators (arxiv.org)

Reflex-Informed Neuromuscular Reinforcement Learning for Muscle-Driven Locomotion (arxiv.org)

Model-Aware Schedules Improve Generation via Fiberwise Optimal Transport (arxiv.org)

Albedo Estimation via Latent Bridge Matching (arxiv.org)

TART: A Modular Tool for Technique-Aware Audio-to-Tablature Guitar Transcription (arxiv.org)

RiLM: Parameter-Efficient Language Modeling via Geodesic Decoding (arxiv.org)

Structurally Speaking: Motif-Oriented Graph Captioning through Bidirectional Graph-Text Translation (arxiv.org)

Generative Replay Mitigates Sample Starvation in Quantum Architecture Search (arxiv.org)