AI/ML arXiv cs.AI

Provable diffusion-based posterior sampling for linear inverse problems via DDIM

A provable diffusion-based posterior sampling algorithm (pDDIM) for solving linear inverse problems via coordinate-wise DDIM updates.

AI/ML arXiv cs.AI

Appearance Pointers -- Multimodal Region Control of Diffusion Transformers

Appearance Pointers enable precise regional control in Diffusion Transformers (DiTs) without needing to retrain the base model.

AI/ML arXiv cs.AI

Copy Less, Ground More: Overcoming Repetitive Copying in Long-Context Reasoning via Evidence-Aware Reinforcement Learning

GEAR is a reward shaping method that reduces repetitive copying in long-context LLM reasoning by penalizing distractor overlap and rewarding evidence grounding.

AI/ML Hacker News

Quality non-fiction books are the antithesis of AI slop

A discussion on how high-quality non-fiction books provide a depth and reliability that AI-generated content lacks.

Other Hacker News

Medici family mystery may be solved after more than 400 years

News regarding the potential resolution of a 400-year-old mystery involving the Medici family.

AI/ML Hacker News

Any text-to-SQL benchmark should address difficulties of real-world data stores

A critique of current text-to-SQL benchmarks, arguing that they must account for the complexities and messy nature of real-world data stores.

Other Hacker News

Why do we love music? (2018)

An exploration of the biological and psychological reasons why humans love music.

Tech Business/VC TechCrunch

Google justifies its massive AI spending with a booming cloud business

Google reports record profits driven by strong growth in its cloud business as companies adopt AI infrastructure services.

Tech Business/VC The Verge

Meta won’t have to face the next planned social media addiction trial

Meta avoids a planned social media addiction trial after the plaintiff dropped the case.

AI/ML arXiv cs.AI

Benchmarking Generalization in Financial Statement Fraud Detection: robust evaluation and novel tasks

Researchers propose a robust framework for financial statement fraud detection using LLMs and introduce the CI-FSFD benchmark for better real-world generalization.

AI/ML arXiv cs.AI

PathAgentBench: Benchmarking Evidence-Seeking Vision-Language Models on Whole-Slide Pathology Image

PathAgentBench is introduced to evaluate evidence-seeking vision-language models on whole-slide pathology images, highlighting gaps in autonomous evidence acquisition.

AI/ML arXiv cs.AI

Toward Auditable Fraud Detection: Combining Graph Features, Model Explanations, and Agentic Case Investigation

A study on auditable fraud detection combining graph features and LLM agents, concluding that plausible rationales from agents do not necessarily guarantee better decisions.

Cybersecurity arXiv cs.AI

They'll Verify. They Just Won't Act. How Authority Framing and Laundered Code Turn a Trusted Agentic CI/CD Pipeline Into an Attack Surface

Research demonstrates how authority framing and 'laundered code' can compromise multi-agent LLM CI/CD pipelines, bypassing security scans and reviews.

Software Engineering Hacker News

Safari Technology Preview 248 Released

Apple releases Safari Technology Preview 248, providing early access to new browser features and experimental updates.

Software Engineering Hacker News

Malleable Computing, Emacs, and You

An exploration of malleable computing and its implementation within the Emacs editor, focusing on how users can customize their development environment.

Hardware/Chips Hacker News

Fairphone 6 wide camera experimental Linux support

Experimental Linux support is being developed for the Fairphone 6 wide camera, expanding hardware compatibility for open-source operating systems.

Other Ars Technica

Microsoft brings original Xbox backward compatibility to Windows PCs

Microsoft adds backward compatibility for original Xbox games to Windows PCs, allowing legacy titles to run on modern hardware.

Other Ars Technica

ISPs' long nightmare of having to list all the fees they charge is finally over

The FCC allows ISPs to stop listing all detailed fee structures, reducing transparency for consumers regarding total costs.

Other Ars Technica

Clayface trailer leans into the body horror

A trailer for 'Clayface' focuses on body horror elements of the character's narrative.

AI/ML arXiv cs.AI

Beyond Score Prediction: LLM-Based Essay Scoring and Feedback Generation via Reinforcement Learning with Rubric Rewards

Introduces RLAES, a unified LLM framework for automated essay scoring and feedback generation using reinforcement learning with rubric-based rewards.