AI/ML arXiv cs.AI

MA-SBI: Misspecification-Aware Simulation-Based Inference via Side-Channel Guidance

Researchers propose MA-SBI, a calibration-free framework for simulation-based inference that uses side-channel guidance to correct posterior bias.

AI/ML arXiv cs.AI

RAID: Semantic Graph Diffusion for True Cold-Start and Cross-Lingual Forecasting

RAID is a new framework for time-series forecasting in cold-start scenarios using semantic retrieval and graph-conditioned diffusion.

AI/ML arXiv cs.AI

A Causal Model of Theory of Mind in Conflict for Artificial Intelligence

This paper introduces a structural causal model for Theory of Mind (ToM) in AI to help agents decide when to mentalize during conflict scenarios.

Other Hacker News

I Fired Google

A community discussion on Hacker News regarding moving away from Google services.

AI/ML Hacker News

Subquadratic – Introducing SubQ 1.1 Small

Introduction of SubQ 1.1 Small, a subquadratic sequence modeling approach for improved efficiency.

AI/ML Hacker News

Running local models is good now

A discussion on the current state and improved viability of running large language models locally.

Tech Business/VC TechCrunch

DOJ claims xAI’s unpermitted gas turbines are a matter of ‘national, economic, and energy security’

The US Department of Justice claims that xAI's unpermitted gas turbines are critical for national and energy security.

Tech Business/VC TechCrunch

Robinhood’s note on 10% layoffs shows blaming AI isn’t cutting it

Robinhood announces 10% layoffs without citing AI as the reason, contrasting with other tech firms.

Hardware/Chips The Verge

The Cybercab is the lightest, most efficient Tesla ever made

Tesla's Cybercab is revealed to be the company's most lightweight and range-efficient EV based on EPA filings.

Tech Business/VC The Verge

Pour one out for Roku City

Fox Corporation's potential acquisition of Roku could significantly shift the US television and media landscape.

AI/ML arXiv cs.AI

GIST-CMTF: Goal-State Inference for Causal Minimal Tool Filtering in LLM Agents

Introduction of GIST-CMTF, a goal-state inference layer that improves tool-augmented LLM agents by reducing wrong-goal execution.

AI/ML arXiv cs.AI

Symbolic Informalization: Fluent, Productive, Multilingual

The Informath project utilizes an interlingual architecture to convert formal mathematics into fluent natural language.

AI/ML arXiv cs.AI

Greed Is Learned: Visible Incentives as Reward-Hacking Triggers

Research on 'reward-channel addiction' showing how AI agents can learn to prioritize visible reward proxies over actual task safety.

Other Hacker News

Show HN: Hackers for Granny (defense against industrialized elder fraud)

Introduction of 'Hackers for Granny', a project aimed at defending elderly people against industrialized fraud.

Hardware/Chips Hacker News

4× RTX Pro 6000 Blackwell on Water, and the One Card That Wouldn't Behave

Technical discussion on water-cooling a setup of four RTX Pro 6000 Blackwell GPUs, including troubleshooting a problematic card.

Software Engineering Hacker News

Google Chrome update will close the door on ad blockers

Google Chrome is updating its manifest/system in a way that may significantly limit the effectiveness of ad blockers.

Other The Verge

Florida sues TikTok for allegedly violating its social media ban for kids

Florida has filed a lawsuit against TikTok for allegedly failing to comply with state laws regarding child safety and parental consent.

Other The Verge

Amazon’s capable sound system is over $100 off ahead of Prime Day

Promotional content regarding discounts on Amazon Fire TV sound systems and Bambu Lab 3D printers.

Other Ars Technica

The Ars Technica 2026 Reader Survey: Let your voice be heard!

Ars Technica is conducting its 2026 reader survey to gather feedback on content and user experience.

AI/ML arXiv cs.AI

OpenClaw-Skill: Collective Skill Tree Search for Agentic Large Language Models

Proposed 'Collective Skill Tree Search' (CSTS) framework to automatically construct reusable skill trees for agentic LLMs to improve tool use and reasoning.