HN mrbnprck 24d ago

Show HN: What if your personal website was ChatGPT?

Personal website using LLM inference via Groq for ChatGPT-style CV exploration. 100-200ms latency with throttled token generation for natural interaction.

HN 10keane 24d ago

Show HN: Open-source desktop agent that uses a local folder as its memory

Open-source desktop agent runtime using local markdown files as persistent memory, with sandbox, approvals, and sub-agent support.

HN gowthamsaiyadav 24d ago

Show HN: Claude-teleport, move your Claude Code history between machines

Tool to migrate Claude Code IDE sessions, conversation history, memory, and project context between machines across OS platforms.

HN joalavedra 24d ago

Show HN: AI Trading Agent via WhatsApp

AI trading agent accessible via WhatsApp interface. Minimal implementation details provided.

HN jarodrh 24d ago

Show HN: Frugon – Find which LLM calls a cheaper model could handle (local, MIT)

Frugon identifies which LLM API calls could use cheaper models without quality loss, helping developers optimize token spending.

HN mister_t_green 24d ago

Clone any app into React Native from screenshots (open source)

Open-source tool clones mobile apps into React Native from screenshots using AI. Enables rapid cross-platform development.

HN xmulot 24d ago

AI Agent using a Burp-style toolkit over MCP

Open-weights LLM agent (GLM-5.2, Gemma, Qwen) performs web penetration testing via headless Chromium. Solved 87% OverTheWire Natas challenges using Burp-style toolkit.

HN Haldt 24d ago

Show HN: Artificiety – Agentic society in a fantasy world

Artificiety simulates agentic society in fantasy world where AI agents with personalities pursue goals based on needs and interactions.

HN qorm 24d ago

Show HN: QORM – build UI apps where you and your AI edit the app together, live

QORM is agent-native declarative UI runtime allowing real-time collaboration between humans and AI (Claude, Cursor) editing apps together via MCP protocol.

HN geordee 24d ago

IDE with agentic support built using Flutter

IDE built with Flutter that includes agentic support for AI-assisted development.

HN yangzhou-tech 24d ago

RealityPatch – OpenAI-compatible gateway to mainland China-hosted LLMs

OpenAI-compatible gateway providing access to LLMs hosted on mainland China infrastructure.

HN raphaelj 24d ago

Ask HN: Are OSS projects allowing vibe-coding?

Discussion of whether OSS projects accept AI-generated code; references OCaml PR rejection and LLM improvement trajectory.

HN Anon84 24d ago

Introduction to Algorithms and Machine Learning from Sorting to Strategic Agents [pdf]

PDF introduction covering algorithms, machine learning, and strategic agents. Educational resource on foundational concepts.

HN mrlou 24d ago

Prevail – open-source AI "life OS" (not for your job)

Open-source personal AI OS application; limited technical information provided.

HN iosazee 24d ago

Show HN: Brianni – AI chat on GPT/Claude/Gemini that we can't read (provable)

Privacy-focused AI chat interface supporting multiple LLM providers with cryptographic verification.

HN freakynit 24d ago

Show HN: A calculator CLI to help your agents perform calculations accurately

llm-calc is a dependency-free CLI calculator for LLM agents to perform deterministic arithmetic operations instead of approximating.

HN Graphenium 24d ago

Show HN: Graphenium – Local Trust Layer for AI Agents (Rust, Datalog, Salsa)

Graphenium is a local trust layer for AI agents built with Rust, Datalog, and Salsa.

HN k1r111 24d ago

AI coding assistant is quietly shipping your secrets

Research on how AI coding agents inadvertently leak secrets through configuration files, logs, and artifacts left in codebases.

HN ronaldl93 24d ago

Show HN: I Built PixelGlass – An AI Agent for Building Ghost Themes

PixelGlass is an AI agent that generates custom Ghost blog themes based on user preferences without vendor lock-in.

HN umutm 24d ago

Show HN: FaultFixer – For websites and apps to self-heal themselves

FaultFixer is a one-line integration that detects errors, diagnoses root causes with AI, and generates fixes for coding agents to implement.

HN _pdp_ 24d ago

Persistent Control of Self-Evolving LLM Agents via Self-Reinforcing Injections

Research on control mechanisms for self-evolving LLM agents through self-reinforcing injections.

HN jonnonz 24d ago

Reinforcement Learning with Metacognitive Feedback Elicits Uncertainty in LLMs

Research on reinforcement learning with metacognitive feedback to elicit uncertainty calibration in LLMs.

HN dimitarst 24d ago

Show HN: Access-aware text-to-SQL – stop LLM agents overfetching data

Access-aware text-to-SQL guard for LLM agents enforcing user data permissions while maintaining full catalog awareness.

HN ulinycoin 24d ago

Show HN: Shadow Web – Cut 64–97% of web page tokens for LLM agents

Open-source Python SDK reducing web page tokens for LLM agents by 64-97% through DOM flattening, semantic grouping, and HTML-to-JSON conversion.

HN Gtombri 24d ago

Oikoumene: Autonomous Agent Civilization Simulator

Oikoumene is an autonomous agent civilization simulator.

HN circularfoyers 24d ago

Posthorn – A self-contained email pen-pal daemon for slow, self-hosted LLMs

Posthorn is a self-contained email daemon that enables pen-pal conversations with self-hosted LLMs.

HN bupticybee 24d ago

Show HN: A free, GPU-accelerated Texas Hold'em GTO solver in C++/CUDA

Free GPU-accelerated Texas Hold'em GTO solver in C++/CUDA for game theory and high-performance computing applications.

HN magester 24d ago

New Mass GitHub dorking tool

ghscan is a GitHub reconnaissance tool that maps organizational footprint and scans for leaked secrets across employee repos.

HN ctenb 24d ago

The Widening Gap: The Benefits and Harms of Generative AI for Novice Programmers

Studies effects of generative AI on novice programmers, examining both benefits and harms in programming education.

Ax Shahd Seddik, Fahd Seddik, Amirrezza Esmaeili, Mahdieh Sadatbenis, Fatemeh Fard 24d ago

ARISE: A Repository-level Graph Representation and Toolset for Agentic Program Repair and Fault Localization

Repository-level graph representation and toolset for AI agents performing automated program repair and fault localization with data flow modeling across multiple semantic levels.

Ax Sarthak Choudhary, Atharv Singh Patlan, Nils Palumbo, Ashish Hooda, Kassem Fawaz, Somesh Jha 24d ago

Undetectable Backdoors in Model Parameters: Hiding Sparse Secrets in High Dimensions

Sparse backdoor attack technique for injecting undetectable adversarial perturbations into pre-trained image classifiers and Vision Transformers at supply-chain level.

Ax Yehudit Aperstein, Eden Moran, Alexander Apartsin 24d ago

IRC-Bench: Recognizing Entities from Contextual Cues in First-Person Reminiscences

NER benchmark for recognizing entities from contextual cues in first-person narrative reminiscences, addressing implicit reference resolution in therapeutic and archival contexts.

Ax Zhaohui Du, Zhe Wang, Dongzhan Zhou, Minting Pan, Hongmei Fei, Xiwen Cao, Ting Xiao, Qi Wang, Huanbo Jin, Jiaming Gu, Quan Lu, Zhe Liu 24d ago

BioProVLA-Agent: An Affordable, Protocol-Driven, Vision-Enhanced VLA-Enabled Embodied Multi-Agent System with Closed-Loop-Capable Reasoning for Biological Laboratory Manipulation

Multi-agent vision-language embodied system for biological lab automation with closed-loop reasoning and protocol-driven execution on wet-lab manipulation tasks.

Ax Jun Wen Leong 24d ago

Defense effectiveness across architectural layers: a mechanistic evaluation of persistent memory attacks on stateful LLM agents

Evaluates defenses against persistent memory attacks on stateful LLM agents across architectural layers, testing 6 defenses on 9 open-source models with 5,040 experimental runs.

Ax Zongmin Yu, Liu Yang 24d ago

Evolutionary Ensemble of Agents

Evolutionary Ensemble (EvE): decentralized framework co-evolving coding agents for algorithmic discovery. Evolves guidance and skills rather than modifying base agent substrate.

Ax Yixuan Yang, Mehak Arora, Ryan Zhang, Baraa Abed, Junseob Kim, Tilendra Choudhary, Md Hassanuzzaman, Kevin Zhu, Ayman Ali, Chengkun Yang, Alasdair Edward Gent, Victor Moas, Rishikesan Kamaleswaran 24d ago

Clin-JEPA: A Multi-Phase Co-Training Framework for Joint-Embedding Predictive Pretraining on EHR Patient Trajectories

Clin-JEPA: joint-embedding predictive pretraining on EHR data for simultaneous patient trajectory forecasting and downstream risk prediction without task-specific tuning.

Ax M. Shamail J. Khan, Nisha L. Raichur, Lucas Heublein, Christian Wielenberg, Alexander Mattick, Tobias Feigl, Christopher Mutschler, Felix Ott 24d ago

Active Sensing with Meta-Reinforcement Learning for Emitter Localization from RF Observations

Meta-reinforcement learning framework for GNSS interference localization via active sensing. Agent sequentially explores environment to infer RF emitter position.

Ax Andrea Ferrario 24d ago

High-Risk AI Systems and the Problem of Identity in the European AI Act

EU AI Act identity problem: defining when high-risk systems remain the same over updates. Analyzes conformity assessment and re-assessment obligations.

Ax Xintong Yang, Hao Gu, Binxing Xu, Lujun Li, Bei Liu, Jiacheng Liu, Qiyuan Zhu, Yike Guo, Sirui Han 24d ago

IndexMem: Learned KV-Cache Eviction with Latent Memory for Long-Context LLM Inference

IndexMem: learned KV-cache eviction policy for long-context LLM inference. Uses learned importance distribution to reduce linear cache growth bottleneck in attention.

Ax Aleksandar Todorov, Matthia Sabatelli 24d ago

Learning in Low-Dimensional Subspaces: Orthogonal Bottlenecks for Reinforcement Learning

Orthogonal bottleneck method constrains deep RL agent representations to low-dimensional subspaces via fixed orthonormal projections without auxiliary objectives.

Ax Xinle Deng, Ruobin Zhong, Hujin Peng, Xiaoben Lu, Yanzhe Wu, Guang Li, Buqiang Xu, Yunzhi Yao, Jizhan Fang, Haoliang Cao, Junjie Guo, Yuan Yuan, Ziqing Ma, Yuanqiang Yu, Rui Hu, Baohua Dong, Hangcheng Zhu, Ningyu Zhang 24d ago

MemTrace: Tracing and Attributing Errors in Large Language Model Memory Systems

MemTrace framework for tracing and attributing errors in LLM memory systems. Debugs how information synthesis, propagation, or corruption occurs in long-horizon reasoning.

Ax Jiayi Fang 24d ago

Emergent Semantic Representations in World Models through Physical Interaction without Linguistic Supervision

World models trained on embodied exploration develop spatial semantic structure mirroring physical geometry without linguistic supervision, analyzed via VAE latent spaces.

Ax Botao Amber Hu, Helena Rong, Max Van Kleek 24d ago

Dissociative Identity: Language Model Agents Lack Grounding for Reputation Mechanisms

Analyzes identity and reputation mechanisms for autonomous LLM agents. Shows lack of grounding for trust verification and credibility signals in multi-agent systems.

Ax Xinyang Lu, Jiabao Pan, Rachael Hwee Ling Sim, See-Kiong Ng, Anthony Kum Hoe Tung, Bryan Kian Hsiang Low 24d ago

De-attribute to Forget for LLM Unlearning

LLM unlearning method using de-attribution to remove inappropriate training data influence while maintaining model utility and avoiding over-forgetting.

Ax Andrzej Cichocki, Michal Wietczak 24d ago

Automatically Differentiable Nonlinear Tensor Networks (ADNTNs) for Exponential Parameter Compression of Deep Neural Networks

ADNTNs compress deep neural networks via learned tensor networks that generate weights for dense and attention layers, reducing parameter count through hierarchical contractions.

Ax Ruiqing Sun, Sen Yang, Dawei Feng, Bo Ding, Yijie Wang, Huaimin Wang 24d ago

ParetoPilot: Zero-Surrogate Offline Multi-Objective Optimization via Infer-Perturb-Guide Diffusion

ParetoPilot: offline multi-objective optimization using diffusion models for Pareto-optimal design generation without surrogate models, guided at inference time.

Ax Wanghan Xu, Shuo Li, Tianlin Ye, Qinglong Cao, Yixin Chen, Hengjian Gao, Yiheng Wang, Qi Li, Kun Li, Sheng Xu, Shengdu Chai, Fangchen Yu, Xiangyu Zhao, Zhangrui Zhao, Weijie Ma, Zijie Guo, Koutian Wu, Haoyu Zhou, Haoxiang Yin, Lixue Cheng, Chaofan Hu, Haoxuan Li, Lu Mi, Xuxuan Xie, Yifan Zhou, Ruizhe Chen, Zhiwang Zhou, Xingjian Guo, Yuhao Zhou, Xuming He, Shengyuan Xu, Xinyu Gu, Jiamin Wu, Mianxin Liu, Chunfeng Song, Fenghua Ling, Dongzhan Zhou, Shixiang Tang, Yuqiang Li, Mao Su, Peng Ye, Siqi Sun, Bin Wang, Xue Yang, Zhenfei Yin, Tianfan Fu, Guangtao Zhai, Wanli Ouyang, Bo Zhang, Lei Bai, Wenlong Zhang 24d ago

ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research

ResearchClawBench: benchmark for evaluating autonomous AI coding agents on 40 scientific research tasks across 10 domains with real published papers and multimodal evaluation rubrics.

Ax Igor Itkin 24d ago

Quickest Detection of Hallucination Onset: Delay Bounds and Learned CUSUM Statistics

Formulates hallucination onset detection in LLMs as quickest change detection problem. Proposes learned CUSUM statistics to monitor token-level hallucinations with reaction time bounds.

Ax Cheng Qian 24d ago

A Mathematical Theory of Value: a synthesis on goal-directed agency under resource constraints

Mathematical theory formalizes value as lawful structural quantity in goal-directed agents, using scale-invariance axiom to derive logarithmic value measure.

Ax Animesh Tripathy, Aswanth Krishnan 24d ago

Iterative Visual Thinking and the Self-Correction Mirage in VLM Grounding

Iterative Visual Thinking (IVT) evaluates visual self-correction in VLMs for spatial grounding; shows refinement often corrects correct predictions not errors.