HN jlengelbrecht 5/15/2026

Show HN: GlycemicGPT – Open-source AI-powered diabetes management

Self-hosted AI platform integrating glucose monitors and insulin pumps with LLM analysis layer for diabetes management on user infrastructure.

HN cui511511 5/15/2026

FilePilot AI – local-first desktop file manager with optional AI summaries

Local-first desktop file manager with optional AI summarization supporting both local and cloud AI providers through unified interface.

HN yakshithk_ 5/15/2026

Ask HN: How do you catch regressions when you change your AI agent's prompt?

Discussion on regression testing strategies for AI agents when modifying prompts, model swaps, or tool calls without manual verification.

Ax Chenlu Ding, Jiancan Wu, Yanchen Luo, Zheyuan Liu, Yancheng Yuan, Xiang Wang 5/15/2026

Teaching Large Language Models When Not to Know: Learning Temporal Critique for Ex-Ante Reasoning

Study on temporal critique in LLMs to improve ex-ante reasoning and prevent knowledge leakage across time cutoffs.

Ax Joy Bose 5/15/2026

Falkor-IRAC: Graph-Constrained Generation for Verified Legal Reasoning in Indian Judicial AI

Falkor-IRAC system using graph-constrained generation for legal reasoning in Indian judicial AI with verified precedent handling.

Ax Haoran Zhang, Luxin Xu, Zhilin Wang, Runquan Gui, Shunkai Zhang, Haodi Lei, Zihao He, Bingsu He, Chicheng Qin, Tong Zhu, Xiaoye Qu, Yang Yang, Yu Cheng, Yafu Li 5/15/2026

$\pi$-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows

π-Bench benchmark for evaluating proactive personal assistant agents that identify hidden user intents in long-horizon workflows.

Ax Minghao Wu, Yuting Yan, Zhenyang Cai, Ke Ji, Chuangsen Fang, Ziying Sheng, Xidong Wang, Rongsheng Wang, Hejia Zhang, Shuang Li, Benyou Wang, Hongyuan Zha 5/15/2026

Agentifying Patient Dynamics within LLMs through Interacting with Clinical World Model

SepsisAgent: world model-augmented LLM agent for sepsis management integrating learned clinical dynamics with LLM reasoning.

Ax Gong Zhiren, Tiantong Wu, Jiaming Zhang, Fuyao Zhang, Che Wang, Yurong Hao, Yikun Hou, Foo Ping, Yilei Zhao, Fei Huang, Chau Yuen, Wei Yang Bryan Lim 5/15/2026

XDomainBench: Diagnosing Reasoning Collapse in High-Dimensional Scientific Knowledge Composition

XDomainBench: diagnostic benchmark for LLM compositional generalization in interactive scientific knowledge synthesis tasks.

Ax Luca Marzari, Enrico Marchesini 5/15/2026

Probabilistic Verification of Recurrent Neural Networks for Single and Multi-Agent Reinforcement Learning

Probabilistic verification tool for RNNs in single and multi-agent reinforcement learning with latent hidden state dynamics.

Ax Yoshia Abe, Tatsuya Daikoku, Yasuo Kuniyoshi 5/15/2026

AI Outperforms Humans in Personalized Image Aesthetics Assessment via LLM-Based Interviews and Semantic Feature Extraction

LLM-based approach to personalized image aesthetics assessment via semantic feature extraction and user interviews.

Ax Shaoan Zhao, Huanlin Gao, Qiang Hui, Ting Lu, Xueqiang Guo, Yantao Li, Xinpei Su, Fuyuan Shi, Chao Tan, Fang Zhao, Kai Wang, Shiguo Lian 5/15/2026

MediaClaw: Multimodal Intelligent-Agent Platform Technical Report

MediaClaw: multimodal agent platform with three-layer architecture addressing fragmentation, heterogeneity, and workflow reuse in AIGC deployment.

Ax Zhao Yang, Wang Huan, Li Yingshuo, Tu Haomiao, Lin Hujite 5/15/2026

A Heterogeneous Temporal Memory Governance Framework for Long-Term LLM Persona Consistency

ARPM: temporal memory governance framework for long-term LLM consistency across dialogue, addressing fact loss and persona drift.

Ax Vineet Kotecha, Vansh Gupta 5/15/2026

Emotion-Attended Stateful Memory (EASM):The Architecture for Hyper-Personalization at Scale

EASM: emotion-attended stateful memory architecture enabling persistent user-specific context and hyper-personalization across LLM sessions.

Ax Yu Zhang, Dongjiang Zhuang, Qu Zhou, Zheng Huang, Junhe Wu, Jing Cao, Kai Chen 5/15/2026

A Deterministic Agentic Workflow for HS Tariff Classification: Multi-Dimensional Rule Reasoning with Interpretable Decisions

Deterministic agentic workflow for HS tariff classification using multi-dimensional rule reasoning with interpretable decisions.

Ax Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, Jonatan Liberman, Rotem Brazilay, Liron Schliesser, Max Svidlo, Shai Nir, Orel Shalom, Yaron Friedman, David Connack, Amos Rimon, Philip Tannor, Shir Chorev 5/15/2026

Holistic Evaluation and Failure Diagnosis of AI Agents

Holistic evaluation framework for AI agents combining top-down diagnosis with bottom-up span-level analysis to identify failure types and locations.

Ax Shihao Qi, Jie Ma, Rui Xing, Wei Guo, Xiao Huang, Zhitao Gao, Jianhao Deng, Jun Liu, Lingling Zhang, Bifan Wei, Boqian Yang, Pinghui Wang, Jianwen Sun, Jing Tao, Yaqiang Wu, Hui Liu, Yu Yao, Tongliang Liu 5/15/2026

Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolution in LLM-based Multi-Agent Systems

Survey of LLM-based multi-agent systems covering collaboration, error propagation, failure attribution, and self-evolution mechanisms.

Ax Sohel Aman Khan, Raghava Mutharaju, Supratim Shit 5/15/2026

COREKG: Coreset-Guided Personalized Summarization of Knowledge Graphs

COREKG: personalized knowledge graph summarization using coreset methods for question answering and visualization tasks.

Ax Yisen Gao, Jiaxin Bai, Haoyu Huang, Zhongwei Xie, Yufei Li, Hong Ting Tsang, Sirui Han, Yangqiu Song 5/15/2026

KGPFN: Unlocking the Potential of Knowledge Graph Foundation Model via In-Context Learning

KGPFN: knowledge graph foundation model leveraging in-context learning for reasoning over unseen entities and relations.

Ax Varad Vishwarupe, Nigel Shadbolt, Marina Jirotka 5/15/2026

From Sycophantic Consensus to Pluralistic Repair: Why AI Alignment Must Surface Disagreement

Framework addressing AI alignment through pluralistic values rather than preference aggregation, highlighting sycophantic consensus failure modes.

Ax Drewry H. Morris V (MedFlow, Inc.), Luis Valles (MedFlow, Inc.), Reza Hosseini Ghomi (MedFlow, Inc.) 5/15/2026

GraphFlow: An Architecture for Formally Verifiable Visual Workflows Enabling Reliable Agentic AI Automation

GraphFlow: visual workflow system for agentic AI automation in multi-step processes with formal verification for semantic correctness guarantees.

Ax Chris Davis Jaldi, Anmol Saini, Shan Zhang, Noah Schroeder, Cogan Shimizu, Eleni Ilkou 5/15/2026

Small, Private Language Models as Teammates for Educational Assessment Design

Small private language models for educational assessment design, examining generation, evaluation, and deployment constraints versus proprietary alternatives.

Ax Baolin Peng, Wenlin Yao, Qianhui Wu, Hao Cheng, Xiao Yu, Rui Yang, Tao Ge, Alessandrio Sordoni, Xingdi Yuan, Yelong Shen, Pengcheng He, Tong Zhang, Zhou Yu, Jianfeng Gao 5/15/2026

Orchard: An Open-Source Agentic Modeling Framework

Orchard: open-source agentic modeling framework for LLM agents with planning, reasoning, tool use, and multi-turn environment interaction.

Ax Renning Pang, Tian Lan, Leyuan Liu, Piao Tong, Sheng Cao, Xiaosong Zhang 5/15/2026

Case-Based Calibration of Adaptive Reasoning and Execution for LLM Tool Use

CAST: case-driven framework for LLM tool use calibrating reasoning depth and structural validity using historical execution trajectories.

Ax Rongman Xu, Yifei Li, Tianzhe Zhao, Yanrui Wu, Bo Li, Hang Yan 5/15/2026

Dual-Dimensional Consistency: Balancing Budget and Quality in Adaptive Inference-Time Scaling

Dual-Dimensional Consistency: adaptive inference-time scaling balancing sampling width and depth for efficient LLM reasoning with budget constraints.

Ax Riccardo Terrenzi, Maximilian von Zastrow, Serkan Ayvaz 5/15/2026

Why Neighborhoods Matter: Traversal Context and Provenance in Agentic GraphRAG

Agentic GraphRAG: framework addressing citation faithfulness in graph-based retrieval-augmented generation by considering agent traversal trajectories.

Ax Evan Rose, Tushin Mallick, Matthew D. Laws, Cristina Nita-Rotaru, Alina Oprea 5/15/2026

APWA: A Distributed Architecture for Parallelizable Agentic Workflows

APWA: distributed architecture for parallelizable multi-agent LLM workflows addressing coordination, reasoning, and computational scaling bottlenecks.

Ax Shang Zhou, Wenhao Chai, Kaiyuan Liu, Huanzhi Mao, Qiuyang Mang, Jingbo Shang 5/15/2026

OpenDeepThink: Parallel Reasoning via Bradley--Terry Aggregation

OpenDeepThink: test-time scaling method using parallel reasoning with Bradley-Terry aggregation to select best reasoning candidates without ground-truth verification.

Ax Alexandre Le Mercier, Chris Develder, Thomas Demeester 5/15/2026

Hidden State Poisoning Attacks against Mamba-based Language Models

Hidden State Poisoning Attacks: adversarial attack method exploiting state space models like Mamba via specific input phrases causing hidden state corruption.

Ax Alexandre Le Mercier, Chris Develder, Thomas Demeester 5/15/2026

GAMBIT: A Three-Mode Benchmark for Adversarial Robustness in Multi-Agent LLM Collectives

GAMBIT: benchmark for evaluating adversarial robustness in multi-agent LLM systems with adaptive adversaries and three evaluation modes.

Ax Sihang Guo, Chenlin Zhou, Jiaqi Wang, Kehai Chen, Qingyan Meng, Zhengyu Ma 5/15/2026

BiSpikCLM: A Spiking Language Model integrating Softmax-Free Spiking Attention and Spike-Aware Alignment Distillation

BiSpikCLM: spiking neural network language model with softmax-free attention and spike-aware distillation for energy-efficient LLM alternatives.

Ax Sushant Gautam, Annika W. Olstad, Klas H. Pettersen, Michael A. Riegler 5/15/2026

The Moltbook Observatory Archive: an incremental dataset of agent-only social network activity

Moltbook Observatory Archive: dataset of agent-only social network activity with continuously recorded agent profiles, posts, and platform metrics.

Ax Said Slaoui 5/15/2026

S-AI-Recursive: A Bio-Inspired and Temporal Sparse AI Architecture for Iterative, Introspective, and Energy-Frugal Reasoning

S-AI-Recursive: bio-inspired sparse AI architecture for iterative reasoning using hormonal closed-loop iteration instead of feed-forward passes.

Ax Wajdi Aljedaani, Rubel Hassan Mollik 5/15/2026

Large Language Models for Web Accessibility: A Systematic Literature Review

Systematic literature review of LLM applications in web accessibility, covering content generation, issue detection, and remediation approaches.

Ax Ahmadreza Jeddi, Minh Ngoc Le, Hakki C. Karaimer, Konstantinos G. Derpanis, Babak Taati 5/15/2026

GEAR: Genetic AutoResearch for Agentic Code Evolution

GEAR: genetic algorithm framework enabling autonomous research agents to explore multiple evolutionary paths simultaneously instead of single-path search strategies.

Ax Abdullah Naeem, Md Wasi Ul kabir, Manish Bhatt, Ayon Dey, Anav Katwal, Md Tamjidul Hoque 5/15/2026

ARES-LSHADE: Autoresearch-Enhanced LSHADE with Memetic Polish for the GNBG Benchmark

ARES-LSHADE: memetic differential-evolution algorithm for optimization, built via LLM-driven autonomous research loop for GECCO 2026 competition.

Ax Jiajun Zhou, Wei Shao, Lingchao Zheng, Yuwei Fan, Ngai Wong 5/15/2026

AIS: Adaptive Importance Sampling for Quantized RL

Adaptive importance sampling method for reinforcement learning with quantized rollouts and BF16 trainer mismatch correction.

Ax Erica Zhang, Fangzhao Zhang, Aneesh Pappu, Batu El, Jose Blanchet, Susan Athey, Jiashuo Liu, James Zou 5/15/2026

TERMS-Bench: Diagnosing LLM Negotiation Agents Beyond Deal Rate

Benchmark for evaluating LLM negotiation agents beyond deal rate, assessing strategic communication under hidden preferences.

Ax Lingchao Zheng, Yuwei Fan, Jun Li, Chengqiu Hu, Qichen Liao, Junyi Fan, Rui Shi, Fangzheng Miao 5/15/2026

Multi-Scale Dequant: Eliminating Dequantization Bottleneck via Activation Decomposition for Efficient LLM Inference

Optimization technique eliminating dequantization bottleneck in quantized LLM inference through activation decomposition.

Ax Daniel M. Jimenez-Gutierrez, Enrique Zuazua, Georgios Kellaris, Joaquin del Rio, Oleksii Sliusarenko, Xabi Uribe-Etxebarria 5/15/2026

Towards the Next Frontier of LLMs, Training on Private Data: A Cross-Domain Benchmark for Federated Fine-Tuning

Federated fine-tuning benchmark for training LLMs on private data across regulated domains like healthcare and finance.

Ax Haomin Zhuang, Hanwen Xing, Yujun Zhou, Yuchen Ma, Yue Huang, Yili Shen, Yufei Han, Xiangliang Zhang 5/15/2026

AgentTrap: Measuring Runtime Trust Failures in Third-Party Agent Skills

Security analysis of third-party agent skills, measuring how malicious skills can disguise harmful behavior in LLM agent workflows.

Ax Jiaqi Liu, Xinyu Ye, Peng Xia, Zeyu Zheng, Cihang Xie, Mingyu Ding, Huaxiu Yao 5/15/2026

EvolveMem:Self-Evolving Memory Architecture via AutoResearch for LLM Agents

Self-evolving memory system for LLM agents that co-evolves stored knowledge and retrieval mechanisms for long-term multi-session operation.

Ax Darius A. Faroughy, Sofia Palacios Schweitzer, Ian Pang, Siddharth Mishra-Sharma, David Shih 5/15/2026

Collider-Bench: Benchmarking AI Agents with Particle Physics Analysis Reproduction

Benchmark for evaluating LLM agent capabilities on reproducing particle physics analysis from papers and scientific software.

Ax Evelyn Turri, Davide Bucciarelli, Sara Sarto, Lorenzo Baraldi, Marcella Cornia 5/15/2026

Few Channels Draw The Whole Picture: Revealing Massive Activations in Diffusion Transformers

Analysis of activation patterns in Diffusion Transformers showing sparse channels control image generation from text prompts.

Ax Katherine Lambert, Sasha Luccioni 5/15/2026

Towards Resource-Efficient LLMs: End-to-End Energy Accounting of Distillation Pipelines

Energy accounting study of LLM distillation pipelines including teacher-side costs and environmental impact assessment.

Ax Tao Zhong, Dongzhe Zheng, Christine Allen-Blanchette 5/15/2026

HodgeCover: Higher-Order Topological Coverage Drives Compression of Sparse Mixture-of-Experts

Research on compressing Sparse Mixture-of-Experts models using topological methods to reduce inference costs without retraining.

Ax Tomoyoshi Kimura, Denizhan Kara, Jinyang Li, Hongjue Zhao, Yigong Hu, Yizhuo Chen, Xiaomin Ouyang, Shengzhong Liu, Tarek Abdelzaher 5/15/2026

Dywave: Event-Aligned Dynamic Tokenization for Heterogeneous IoT Sensing Signal

Dynamic tokenization framework for heterogeneous IoT sensing signals addressing non-stationary multi-scale data challenges.

Ax Haofei Xu, Umar Iqbal, Jacob M. Montgomery 5/15/2026

Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact

Large-scale study measuring Google AI Overviews deployment across activation rates, source quality, claim fidelity, and publisher impact.

Ax Xiao Jia 5/15/2026

Do Language Models Align with Brains? Prediction Scores Are Not Enough

Framework evaluating whether language model representations align with brain language processing beyond prediction scores.

Ax Sanghyeob Song, Donghyeok Lee, Jinsik Kim, Sungroh Yoon 5/15/2026

R2R2: Robust Representation for Intensive Experience Reuse via Redundancy Reduction in Self-Predictive Learning

Regularization method for self-predictive learning in reinforcement learning that improves data efficiency in robotic learning with high update-to-data ratios.

Ax Ignacio Sastre, Guillermo Moncecchi, Aiala Ros\'a 5/15/2026

Derivation Prompting: A Logic-Based Method for Improving Retrieval-Augmented Generation

Logic-inspired prompting technique for RAG systems that improves question answering accuracy and reduces hallucinations in knowledge-intensive domains.