Other Hacker News

Modder Makes GTA: Vice City Playable Inside GTA 3 Inside GTA: San Andreas

A modder creates a complex recursive game-within-a-game, making GTA: Vice City playable inside GTA 3, which is itself inside GTA: San Andreas.

Tech Business/VC The Verge

Electric air taxis go to war

Archer Aviation and Anduril collaborate to develop 'Thunder', an electric aircraft for both military and commercial air taxi use.

AI/ML arXiv cs.AI

AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations

Introduces AgentRedBench, a dynamic redteaming benchmark and AgentRedGuard model to protect LLM agents from indirect prompt injection in SaaS integrations.

AI/ML arXiv cs.AI

AuAu: A Benchmark for Auditing Authoritarian Alignment in Large Language Models

Presents AuAu, a benchmark for auditing authoritarian tendencies and alignment in Large Language Models.

AI/ML arXiv cs.AI

GeoRouteNet: A Geometry-Aware Non-Autoregressive Neural Solver for the Euclidean Traveling Salesman Problem

Introduces GeoRouteNet, a geometry-aware non-autoregressive neural solver for the Euclidean Traveling Salesman Problem.

Tech Business/VC AI News

US public health agencies to test OpenAI and Anthropic AI models

US public health agencies will test generative AI tools from OpenAI and Anthropic via the PULSE programme.

AI/ML AI News

Kimi K3 open-weight model: China’s biggest AI is a bet on memory, not compute

Moonshot AI releases Kimi K3, a 2.8 trillion parameter open-weight model focusing on memory efficiency over compute.

Open Source Hacker News

Stop Using OpenCode

A community discussion regarding the use of OpenCode, likely focusing on critiques or alternatives.

Cybersecurity Hacker News

Hacker wipes Romania's land registry database

A hacker successfully wiped the land registry database of Romania, highlighting critical security vulnerabilities in government infrastructure.

Hardware/Chips Hacker News

ECC and DDR5

Technical discussion regarding the implementation and differences between ECC and DDR5 memory.

Other Hacker News

Satan's 19th-Century Bank Note

A non-technical story about a 19th-century bank note associated with Satan.

AI/ML Hacker News

Inertia-1: An Open Exploration to a Unified Motion Foundation Model

Introduction of Inertia-1, a unified motion foundation model for physical movement and motion analysis.

Hardware/Chips Hacker News

Codex is wearing out our devices

An exploration of how the use of Codex and similar AI tools may be accelerating the wear and tear on physical hardware devices.

Other The Verge

Avengers: Doomsday’s first trailer puts everyone on high alert

Marvel releases the first trailer for Avengers: Doomsday featuring Robert Downey Jr. as Doctor Doom.

AI/ML arXiv cs.AI

Diagnosing Overhead in Dispatch Operations: Cross-architecture Observatory

Research on diagnosing dispatch overhead in Mixture-of-Experts (MoE) parallelism across different architectures including MHA, MLA, and Mamba-2.

AI/ML arXiv cs.AI

The Terminal Representation in Reinforcement Learning

Introduction of the Terminal Representation (TR) in reinforcement learning as a computationally efficient alternative to Successor and Default representations.

AI/ML arXiv cs.AI

memorywire: A Vendor-Neutral Wire Format for Agent Memory Operations

Introduction of memorywire, a vendor-neutral JSON-Schema wire format for agent memory operations to standardize how AI agents manage long-term storage.

Other Hacker News

Sensing warm and cool: how the body detects temperature changes

An article exploring the biological mechanisms of how the human body detects temperature changes.

AI/ML arXiv cs.AI

When Perplexity Lies: Generation-Focused Distillation of Hybrid Sequence Models

Introduces GenDistill, a pipeline for distilling Transformers into efficient Hybrid-KDA models, reducing KV cache memory by 75% and improving time-to-first-token.

AI/ML arXiv cs.AI

Ruling Out to Rule In: Contrastive Hypothesis Retrieval for Medical Question Answering

Proposes Contrastive Hypothesis Retrieval (CHR), a framework for medical QA that reduces hard-negative contamination in RAG systems by modeling what to avoid.