HN infamous-oven 5/1/2026

Show HN: BetterClaw – Compile a paragraph into a workflow that gates agent tools

Open-source tool for controlling AI agent tool access via natural language workflow descriptions, response to PocketOS database incident.

HN inguz 5/1/2026

Show HN: "Woo", a multi-user world for agents

Show HN: Multi-user persistent world for AI agents using Cloudflare Durable Objects and MCP protocol, inspired by LambdaMOO.

HN dmsozpepnc 5/1/2026

Turn Git commits and PRs into automated team updates (Slack/email)

Tool converts Git commits and PRs into automated Slack/email team notifications.

HN theaniketmaurya 5/1/2026

Parallel Pi coding agents in a sandbox

Project demonstrating parallel execution of coding AI agents within sandbox environments.

HN eummm 5/1/2026

Show HN: Revdoku – visual document review with AI (open-source)

Open-source AI platform for automated document review with line-by-line rule checks and custom script execution.

HN saikatsg 5/1/2026

The AI Jailbreakers

Feature article on red teamers and security researchers jailbreaking LLMs to test safety, including emotional impact on practitioners.

HN sploitzberg 5/1/2026

Mosaic: Local MCP server for agent memory

Mosaic is a local MCP server implementation for managing agent memory in AI systems.

HN todience 5/1/2026

Show HN: Loopsy, a way for terminals and AI agents on different machines to talk

Show HN: Loopsy terminal communication tool enabling AI agents and commands across machines via local network and Cloudflare Workers.

HN alex_petrov 5/1/2026

Xmemory: Benchmarking Structured AI Memory Against RAG and Hybrid RAG

arXiv announcement about Xmemory benchmark comparing structured AI memory, RAG, and hybrid RAG approaches.

HN lastdong 5/1/2026

Advanced Quantization Algorithm for LLMs

AutoRound: Quantization toolkit for LLMs achieving 2-4 bit precision with minimal accuracy loss, open-source with hardware compatibility.

HN mazzamani 5/1/2026

Bouncy – A small Rust web scraper with built-in MCP support

Bouncy: Rust headless browser web scraper with MCP support; single binary, JavaScript execution, Playwright compatible.

HN sea-gold 5/1/2026

Claude Code Source Code Breakdown

Analysis of Claude Code source code leak via npm sourcemap, revealing Anthropic's AI coding CLI implementation and discussing security implications.

HN jacobtomlinson 5/1/2026

OpenClaw Got Safer in Public

OpenClaw, an open-source tool for running tools and installing plugins, improved security through community contributions and public transparency.

HN gcv 5/1/2026

Lightweight OpenCode profile for routine dev work with focused agents

Supersimple is a lightweight OpenCode profile for software development featuring agent orchestration, local skills, and conductor-based track management.

LB github.com by gcv 5/1/2026

Supersimple: Lightweight OpenCode profile for routine dev work with focused agents, local skills, and conductor-based track management

Supersimple is a lightweight OpenCode profile for software development featuring agent orchestration, local skills, and conductor-based track management.

HN swq115 5/1/2026

Reunderstanding the Power of AI Through Reverse Engineering

Demonstration of using AI agents with Ghidra MCP for binary reverse engineering, showing AI capabilities beyond initial expectations.

HN kstonekuan 5/1/2026

Openpi-flash: Real-time inference engine for openpi

Openpi-flash is a real-time inference engine optimized for low-latency policy serving for robots in production environments.

HN montyanderson 5/1/2026

10x Faster Real-Time High-Quality AI Video Generation

Tenstorrent Galaxy supercluster achieves 10x faster real-time video generation using state-of-the-art models with latency/throughput benchmarks.

HN eigenBasis 5/1/2026

Task-Specific LLM Evals That Do and Don't Work

Survey of task-specific LLM evaluation methods that work in production, covering why off-the-shelf evals fail and practical alternatives for measuring application performance.

HN GeneLeo 5/1/2026

PrivClaw – Open-source self-hostable AI plugin marketplace (FastAPI and Next.js)

Open-source self-hostable AI plugin marketplace built with FastAPI and Next.js addressing security risks in enterprise AI workflows.

HN achierius 5/1/2026

LLMs Don't Quite Beat Classical Hyperparameter Optimization Algorithms

Empirical study comparing LLM agents for hyperparameter optimization against classical algorithms like CMA-ES and TP on fixed compute budgets.

Ax Beomchul Park, Minsu Koh, Heejo Kong, Seong-Whan Lee 5/1/2026

Compositional Meta-Learning for Mitigating Task Heterogeneity in Physics-Informed Neural Networks

Meta-learning approach for physics-informed neural networks to reduce retraining across heterogeneous PDE tasks.

Ax Aditya Kar (CNRS, IRIT), Emiliano Lorini (CNRS, IRIT), Timoth\'ee Masquelier (CNRS, CERCO UMR5549) 5/1/2026

Binary Spiking Neural Networks as Causal Models

Causal analysis of binary spiking neural networks using SAT/SMT solvers for logic-based explanations of network behavior.

Ax Emma Casey, David Roberts, David Sim, Ian Beaver 5/1/2026

When Your LLM Reaches End-of-Life: A Framework for Confident Model Migration in Production Systems

Framework for migrating production LLM systems using Bayesian calibration of evaluation metrics for confident model replacement.

Ax Shuxing Yang, Fujia Chen, Rui Zhao, Junyao Wu, Yize Wang, Haiyao Luo, Ning Han, Qiaolu Chen, Yuze Hu, Wenhao Li, Mingzhu Li, Hongsheng Chen, Yihao Yang 5/1/2026

End-to-end autonomous scientific discovery on a real optical platform

LLM-based autonomous agent performs end-to-end scientific discovery on real optical platform with experimental validation.

Ax Adela Bara, Gabriela Dobrita, Simona-Vasilica Oprea 5/1/2026

Think it, Run it: Autonomous ML pipeline generation via self-healing multi-agent AI

Multi-agent AI system that autonomously generates ML pipelines from datasets and natural language goals using RAG and DAG construction.

Ax Yu-Chao Huang, Zhen Tan, Mohan Zhang, Pingzhi Li, Zhuo Zhang, Tianlong Chen 5/1/2026

TRUST: A Framework for Decentralized AI Service v.0.1

Decentralized verification framework for Large Reasoning Models and Multi-Agent Systems addressing robustness, scalability, opacity, and privacy in high-stakes domains.

Ax Daiana Rinja, Eduardo Araujo Oliveira, Sonsoles L\'opez-Pernas, Mohammed Saqr, Marcus Specht, Kamila Misiejuk 5/1/2026

Unpacking Vibe Coding: Help-Seeking Processes in Student-AI Interactions While Programming

Empirical study analyzing student help-seeking interaction patterns with AI during programming through vibe coding, comparing high and low performers.

Ax Jinbiao Wei, Kangqi Ni, Yilun Zhao, Guo Gan, Arman Cohan 5/1/2026

Step-level Optimization for Efficient Computer-use Agents

Optimization techniques for computer-use agents to reduce computational cost by selective multimodal model invocation instead of uniform step-level calls.

Ax Yuxuan Huang, Yihang Chen, Zhiyuan He, Yuxiang Chen, Ka Yiu Lee, Huichi Zhou, Weilin Luo, Meng Fang, Jun Wang 5/1/2026

Web2BigTable: A Bi-Level Multi-Agent LLM System for Internet-Scale Information Search and Extraction

Web2BigTable bi-level multi-agent LLM system for internet-scale information search balancing depth reasoning and breadth aggregation with schema alignment.

Ax Juergen Dietrich 5/1/2026

When Roles Fail: Epistemic Constraints on Advocate Role Fidelity in LLM-Based Political Statement Analysis

Empirical study of role fidelity in multi-agent LLM pipelines for political discourse analysis, finding models inconsistently maintain assigned adversarial roles.

Ax Anh Ta, Junjie Zhu, Shahin Shayandeh 5/1/2026

Reinforced Agent: Inference-Time Feedback for Tool-Calling Agents

Reinforced Agent framework adds inference-time feedback loop to tool-calling LLM agents enabling real-time course correction during execution.

Ax Fazle Elahi Faisal, Qianhui Wu, Baolin Peng, Jianfeng Gao 5/1/2026

AutoSurfer -- Teaching Web Agents through Comprehensive Surfing, Learning, and Modeling

AutoSurfer generates comprehensive web trajectories for training multimodal LLM-based web agents, addressing data scarcity through surfing and learning methods.

Ax Lucas Vittor, Ayush Noori, I\~naki Arango, Joaqu\'in Polonuer, Sam Rodriques, Andrew White, David A. Clifton, Marinka Zitnik 5/1/2026

OptimusKG: Unifying biomedical knowledge in a modern multimodal graph

OptimusKG multimodal biomedical knowledge graph unifying structured and semi-structured resources with schema constraints for life sciences applications.

Ax Dahlia Shehata, Ming Li 5/1/2026

The Inverse-Wisdom Law: Architectural Tribalism and the Consensus Paradox in Agentic Swarms

Empirical study of multi-agent swarms showing agents prioritize internal architectural agreement over logical truth, challenging wisdom-of-crowds assumption.

Ax Alan L. McCann 5/1/2026

Mechanized Foundations of Structural Governance: Machine-Checked Proofs for Governed Intelligence

Mechanized proofs in Coq for structural governance of cognitive workflow systems with coinductive safety predicates for infinite program behaviors.

Ax Ming-Hong Yao, Di Wang, Jian Cui, Jin-Yan Chen, Zi-Hao Cui, Fa Wang, Chen Wei, Qiu-Ye Yu 5/1/2026

Learning Rate Engineering: From Coarse Single Parameter to Layered Evolution

Systematic study of learning rate scheduling evolution across five generations from fixed rates to layer-time adaptive strategies in neural networks.

Ax Gyoung S. Na, Chanyoung Park 5/1/2026

Machine Collective Intelligence for Explainable Scientific Discovery

Machine collective intelligence paradigm combining multiple AI approaches to discover explainable governing equations from empirical data for scientific discovery.

Ax Jianpeng Chen, Wangzhi Zhan, Dongqi Fu, Junkai Zhang, Zian Jia, Ling Li, Wei Wang, Dawei Zhou 5/1/2026

METASYMBO: Multi-Agent Language-Guided Metamaterial Discovery via Symbolic Latent Evolution

Multi-agent LLM system for metamaterial discovery interpreting natural language design intents and using symbolic latent evolution for geometry optimization.

Ax Aaryan Shah, Andrew Hines, Alexia Downs, Denis Bajet, Paulius Mui, Fabiano Araujo, Laura Offutt, Aida Rutledge, Elizabeth Jimenez 5/1/2026

End-to-End Evaluation and Governance of an EHR-Embedded AI Agent for Clinicians

Framework for continuous governance of clinical AI agents integrating rubric validation, live feedback, performance monitoring, and controlled experimentation in EHR systems.

Ax Ziyao Xu, Cong Wang, Houfeng Wang 5/1/2026

Investigating More Explainable and Partition-Free Compositionality Estimation for LLMs: A Rule-Generation Perspective

Research on compositional generalization testing for LLMs using rule-generation approach to improve explainability and address dataset partition leakage issues.

Ax Zihao Li, Jiaru Zou, Feihao Fang, Xuying Ning, Mengting Ai, Tianxin Wei, Sirui Chen, Xiyuan Yang, Jingrui He 5/1/2026

Heterogeneous Scientific Foundation Model Collaboration

Eywa framework enables heterogeneous agentic LLM systems to collaborate with domain-specific foundation models beyond language interface for scientific domains.

Ax Louth Bin Rawshan, Zhuoyu Wang, Brian Y. Lim 5/1/2026

CoAX: Cognitive-Oriented Attribution eXplanation User Model of Human Understanding of AI Explanations

CoAX studies human cognition in understanding XAI explanations. Evaluates multiple explanation methods on structured data reasoning tasks.

Ax Yuan Sun 5/1/2026

Safe Bilevel Delegation (SBD): A Formal Framework for Runtime Delegation Safety in Multi-Agent Systems

Safe Bilevel Delegation framework for runtime safety in delegating subtasks between LLM agents. Dynamically adjusts safety-efficiency trade-off based on task context.

Ax Sidi Chang, Peiying Zhu, Yuxiao Chen, Rongdong Chai 5/1/2026

Measurement Risk in Supervised Financial NLP: Rubric and Metric Sensitivity on JF-ICR

Studies measurement risk in financial NLP benchmarks. Shows gold labels can be sensitive to rubric wording, metrics, and aggregation policies.

Ax Qiyao Wang, Haoran Hu, Longze Chen, Hongbo Wang, Hamid Alinejad-Rokny, Yuan Lin, Min Yang 5/1/2026

InteractWeb-Bench: Can Multimodal Agent Escape Blind Execution in Interactive Website Generation?

InteractWeb-Bench evaluates multimodal agents on interactive website generation. Addresses semantic misalignment between ambiguous requirements and code synthesis.

Ax Wenhao Yuan, Chenchen Lin, Jian Chen, Jinfeng Xu, Shuo Yang, Edith Cheuk Han Ngai 5/1/2026

Belief-Guided Inference Control for Large Language Model Services via Verifiable Observations

Proposes belief-guided inference control for LLM services. Decides when to allocate additional computation to improve response quality versus using low-cost default responses.

Ax Chaemin Jang, Woojin Park, Hyeok Yun, Dongman Lee, Jihee Kim 5/1/2026

In-Context Examples Suppress Scientific Knowledge Recall in LLMs

Shows in-context examples suppress LLMs' ability to recall and apply scientific formulas. Demonstrates trade-off between in-context learning and parametric knowledge retrieval.

Ax Song Tang, Kaiyong Zhao, Yuliang Li, Qingsong Yan, Penglei Sun, Junyi Zou, Qiang Wang, Xiaowen Chu 5/1/2026

SpatialGrammar: A Domain-Specific Language for LLM-Based 3D Indoor Scene Generation

SpatialGrammar DSL enables LLMs to generate interactive 3D indoor scenes from natural language with fewer spatial errors and collisions.

Ax Anna Mazhar, Huzaifa Suri, Sainyam Galhotra 5/1/2026

Trace-Level Analysis of Information Contamination in Multi-Agent Systems

Analyzes information contamination in multi-agent workflows. Shows uncertainty in artifact extraction can redirect decomposition and produce different execution trajectories.