AI/ML arXiv cs.AI

Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation

Researchers propose using Transformers as Bayesian In-Context Experimenters to improve the efficiency and precision of average treatment effect estimation.

AI/ML arXiv cs.AI

Gated Multi-Graph Fusion via Graph Attention Networks for Alzheimer's Disease Detection

A Multi-View Gated Graph Attention Network is developed to detect Alzheimer's Disease by analyzing non-linear structural disruptions in spontaneous speech.

AI/ML arXiv cs.AI

Distilling Temporal Coherence into 2D Networks for Transrectal Ultrasound Prostate Video Segmentation

A Temporally Consistent Learning Framework is introduced to enable real-time, temporally coherent prostate video segmentation using 2D networks.

AI/ML arXiv cs.AI

Can LLMs Imagine Moral Alternatives Beyond Binary Dilemmas?

A study examines whether LLMs can imagine moral alternatives beyond binary dilemmas, introducing the MoralAltDataset for evaluation.

AI/ML Hacker News

ZCode: Claude Code from the Makers of GLM

Introduction of ZCode, a tool likely inspired by or similar to Claude Code, developed by the makers of GLM.

Other Hacker News

Fable 5 Is Back

Discussion regarding the return or update of Fable 5.

Software Engineering Hacker News

Understanding the Linux Kernel: The Scheduler

A technical deep-dive into the Linux Kernel Scheduler, explaining its inner workings.

Other Hacker News

Mortality associated with non-optimal ambient temperatures from 2000 to 2019

A study analyzing the mortality associated with non-optimal ambient temperatures from 2000 to 2019.

Cybersecurity Hacker News

Show HN: Z-Jail – A 130 KB Linux sandbox-C99 with 7 defense layers and zero deps

Showcase of Z-Jail, a lightweight (130 KB) Linux sandbox written in C99 featuring seven defense layers and zero dependencies.

AI/ML Hacker News

Claude Fable 5 Promotional Access

Information regarding promotional access for Claude Fable 5.

AI/ML arXiv cs.AI

PPT-Eval: A Benchmark for Computer-Use Agents on PowerPoint Tasks

Introduction of PPT-Eval, a benchmark for evaluating computer-use agents on multimodal PowerPoint tasks.

AI/ML arXiv cs.AI

One Retrieval to Cover Them All: Co-occurrence-Aware Knowledge Base Reorganization for Session-Level RAG

Proposal for a co-occurrence-aware knowledge base reorganization to improve session-level retrieval in RAG systems.

AI/ML arXiv cs.AI

LLM-Powered Interactive Robotic Action Synthesis from Multimodal Speech, Gestures, and Music

A framework utilizing LLMs to synthesize robotic actions from multimodal inputs including speech, gestures, and music.

AI/ML arXiv cs.AI

ComplianceGate: Classifier-Gated Multi-Tier LLM Routing for Inference in Regulated Industries

ComplianceGate introduces a classifier-gated routing architecture for LLMs to ensure PII compliance and optimize inference costs in regulated industries.

Hardware/Chips Hacker News

Weave Robotics launches Isaac 1, a $7,999 home robot with fall 2026 deliveries

Weave Robotics has announced Isaac 1, a home robot priced at $7,999, with deliveries expected to begin in fall 2026.

Cybersecurity TechCrunch

Apple’s Hide My Email feature has a bug that’s been exposing real email addresses, researcher claims

A security researcher claims that a bug in Apple's 'Hide My Email' feature has been exposing users' real email addresses.

AI/ML Ars Technica

After spooking Trump into safety testing, Anthropic AI models get global release

Anthropic's advanced AI models, Fable and Mythos, are receiving a global release after the US lifted safety testing restrictions.

Tech Business/VC VentureBeat

The Control Gap: Enterprise AI organizations have an ownership problem, not a technology problem — and most are governing it by hand

A VentureBeat report highlights a 'control gap' in enterprise AI, where expansion is outpacing governance, resulting in ownership vacuums and financial failures from autonomous agents.

Other VentureBeat

Digital resilience compounds when AI and human expertise scale together

A sponsored piece explores how agentic AI may erode the traditional apprenticeship model for junior IT/security roles and proposes design choices to maintain human expertise.

AI/ML arXiv cs.AI

What Probing Reveals about Autonomous Driving: Linking Internal Prediction Errors to Ego Planning

Researchers investigate autonomous driving policies, finding that high performance in nominal scenarios can mask flawed internal prediction and planning reasoning.