AI/ML arXiv cs.AI

Fantastic Adaptive Taxonomies and How to Use Them

AdaMAST creates an automated, adaptive failure taxonomy from agent execution traces to improve system diagnostics and runtime performance.

Hardware/Chips arXiv cs.AI

Automated Hardware Validation Test Plan Generation for Large Scale AI Datacenter Platforms Using a Generative AI Multi-Agents Architecture

A new multi-agent AI architecture automates the generation of hardware validation test plans for large-scale AI datacenters.

AI/ML arXiv cs.AI

Think, Plan, Paint: Layout-Aware Reasoning for Controllable Image Generation in Unified Models

ATLAS is a unified framework for controllable image generation that uses a 'Think, Plan, Paint' layout-aware reasoning paradigm.

Cybersecurity arXiv cs.AI

Signal-based Model Access Risk Analysis for AI System Operations Security

The SMART taxonomy classifies attacker access to AI models based on information signals to better analyze evasion attack risks.

Cybersecurity Hacker News

OpenAI and Hugging Face partner to address security incident

OpenAI and Hugging Face have partnered to address a security incident.

Software Engineering Hacker News

Show HN: Justif – Knuth-Plass justification and microtypography for the web

Justif is a new tool implementing Knuth-Plass justification and microtypography for web content.

Other Hacker News

Show HN: Read the Tape – Wordle for daytrading, five blind S&P 500 charts a day

Read the Tape is a game that applies Wordle-style mechanics to daytrading using S&P 500 charts.

Software Engineering Hacker News

Zero-dependency streaming tar parser and writer for JavaScript

A new zero-dependency streaming tar parser and writer has been released for JavaScript.

AI/ML VentureBeat

Google's Gemini 3.6 Flash model cuts AI agent token costs by up to 65% on long horizon engineering tasks —and 3.5 Pro is on the way

Google released Gemini 3.6 Flash and 3.5 Flash-Lite, focusing on token efficiency and reduced costs for AI agents, along with a specialized Gemini 3.5 Flash Cyber model for security.

AI/ML arXiv cs.AI

Clarify Before Executing: A Self-Evolving Agent for Resolving Intent Asymmetry in 3D Tool Orchestration

CLARE is a self-evolving 3D agent that resolves vague user instructions through strategic dialogue to improve 3D asset creation.

AI/ML arXiv cs.AI

A Predict-then-Correct Loop Based on Few-Shot Continuous Contextual Bandit for Demand Forecasting

Researchers proposed a predict-then-correct framework using few-shot continuous contextual bandits to improve retail demand forecasting.

AI/ML arXiv cs.AI

PhysAgent: Reflective Agentic Physics Control for Physically Plausible Video Generation

PhysAgent is a reflective agentic framework that uses physics simulation and verification loops to generate more physically plausible videos.

Cybersecurity arXiv cs.AI

Reliable Remediation Impact Prediction for Black-Box Security Ratings

A new surrogate-based approach is proposed to predict the impact of remediation actions on black-box security ratings without revealing scoring engine details.

AI/ML arXiv cs.AI

A Quantum-Classical Hybrid Framework for Multivariate Time-Series Forecasting Complexity-Fidelity Trade-offs and Limitations

A quantum-classical hybrid framework for multivariate time-series forecasting is introduced, featuring Quantum Reservoir and Variational Quantum Forecasters.

Cybersecurity Hacker News

'VPNs are lawful technical tools,' says EU Court in landmark copyright ruling

The EU Court has ruled that VPNs are lawful technical tools, a landmark decision regarding copyright law.

Tech Business/VC Hacker News

Judge approves $1.5B Anthropic settlement for pirated books used to train Claude

A judge has approved a $1.5 billion settlement for Anthropic regarding the use of pirated books for training Claude.

Software Engineering TechCrunch

Jack Dorsey is taking on Slack with Buzz, a group chat platform for teams and their AI agents

Jack Dorsey is launching Buzz, a new group chat platform designed for teams and their AI agents.

Tech Business/VC Ars Technica

Anthropic’s $1.5B copyright settlement approved; only 350 authors opted out

Anthropic's $1.5 billion copyright settlement was approved, though some authors opted out at the last minute.

AI/ML VentureBeat

Google's Gemini Flash 3.6 model cuts AI agent token costs by up to 65% on long horizon engineering tasks —and 3.5 Pro is on the way

Google releases Gemini 3.6 Flash and 3.5 Flash-Lite, offering significant token efficiency and lower costs for agentic AI tasks.

AI/ML arXiv cs.AI

DMFNet: Dual-Backbone Multiscale Fusion Network for Urban Scene Classification

DMFNet is introduced as a dual-backbone multiscale fusion network designed to improve urban scene classification in remote sensing.