arXiv:2608.08382v1 Announce Type: new Abstract: As LLM inference shifts to multi-tenant GPU clusters, co-batching improves throughput but obscures…
Query-Only Backdoor Attacks on Self-Evolving Skills via Trajectory Poisoning
arXiv:2608.08303v1 Announce Type: new Abstract: Agentic skills improve large language model (LLM) agents by encoding reusable procedures for complex…
Mitigating Over-Personalization in LLMs via Structured Memory
arXiv:2608.08300v1 Announce Type: new Abstract: Conversational assistants increasingly rely on persistent long-term memory to personalize responses across…
Fair on the Surface? Benchmarking Hidden-Output Fairness Gaps in LLM Recommenders
arXiv:2608.08284v1 Announce Type: new Abstract: Fairness audits for LLM-based recommenders have largely focused on observable outputs, implicitly assuming…
Spotify will label ‘AI Persona’ profiles and exclude their music from recommendations
Spotify is introducing “AI Persona” labels for artist profiles that represent AI-generated identities and will exclude their music from editorial,…
StructReward: Efficient Structured Process Rewards for Self-Correcting Multimodal Reasoning
arXiv:2608.08326v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as an effective approach for improving…
Exploring LLM Capabilities for Situational Understanding and COLREG compliance on real-world maritime navigation scenarios
arXiv:2608.08281v1 Announce Type: new Abstract: Recently, Large Language Models (LLMs) have shown considerable capability for situational understanding,…
OBLIVION: Workflow-Level Operational Skill Unlearning for Deployed Agents
arXiv:2608.08264v1 Announce Type: new Abstract: Large language model agents are becoming operational interfaces to files, memories, registries, and…
Anthropic’s planned mega-IPO faces investor skepticism over Chinese rivals and political headwinds
Anthropic is preparing an IPO for September or October, according to the Wall Street Journal, potentially the largest ever. During investor meetings, the…
FemWear: A Specialized Wearable Foundation Model for Women’s Health
arXiv:2608.08244v1 Announce Type: new Abstract: General wearable foundation models are pretrained across broad sensor streams and populations, but are not…
Anthropic Watermarks Claude Text Output to Meet EU Transparency Rules
Anthropic has confirmed it will embed imperceptible watermarks into text generated by Claude models launched in the EU on or after August 2, 2026 and…
Your Prompt Is Not the Only Prompt: How Much Do LLMs Weight Structured-Output Schema Descriptions?
arXiv:2608.08254v1 Announce Type: new Abstract: Structured output, where an LLM populates a predefined JSON schema, has become a default mechanism for…
AI is Already Here. The Real Challenge Is Trust
For years, discussions around artificial intelligence have centered on capability. Can AI write better content? Can it automate customer interactions? Can…
SuperLocalMemory 4.0: The Governed Memory Operating System for AI Agents
arXiv:2608.08253v1 Announce Type: new Abstract: AI agents are becoming shared infrastructure, yet durable memory is commonly assembled from separate…
AI News Brief Hourly Summary 2026-08-11 15h : 16 posts
16 posts were published in the last hour 12:34 : Metanormative Theory for RL-Based Moral Agents 12:33 : Illusion of Alignment: Detecting Hidden Disagreement in Collaborative Dialogue 12:33 : LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems 12:33 : Anthropic…
Metanormative Theory for RL-Based Moral Agents
arXiv:2608.08220v1 Announce Type: new Abstract: The overlapping disciplines of machine ethics and value alignment are concerned with designing artificial…
Illusion of Alignment: Detecting Hidden Disagreement in Collaborative Dialogue
arXiv:2608.08210v1 Announce Type: new Abstract: Collaborative dialogue can end with apparent agreement while participants still differ on goals,…
LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems
arXiv:2608.08236v1 Announce Type: new Abstract: Multi-agent LLM systems often fail not for lack of candidate answers, but because they have no persistent…