3/5/2026
Show HN: Captain Claw local AI agent, 29 tools, multi-session, DAG orchestration
Captain Claw: local AI agent runtime with 29 tools, multi-session DAG orchestration, SQLite persistence, supporting OpenAI/Anthropic/Gemini/Ollama.
3/5/2026
We Turned Our Wireshark Wizard into a Markdown File
Rocky AI: Checkly's GA AI agent integrated into SaaS product; lessons on building AI agents beyond chat interfaces for monitoring workflows.
3/5/2026
Show HN: YSA – Sandbox for AI agents with outbound network control
YSA: containerized sandbox for running Claude CLI on production codebases with network isolation, TLS inspection, and data loss prevention.
3/5/2026
Show HN: Thought Canvas – Develop Ideas with AI (Mind Mapping)
Thought Canvas: mind-mapping interface for exploring ideas with AI instead of linear chat format.
3/5/2026
How do you stop AI from making bad architectural decisions in your codebase?
Tool for constraining AI assistants to understand codebase architecture and prevent bad architectural decisions during development.
3/5/2026
Sqry – Semantic code search that understands call graphs (35 langs, Rust, MCP)
Semantic code search tool supporting 35 languages with call graph understanding, built in Rust with MCP support.
3/5/2026
Show HN: PromptPerfect – Open-source prompt optimizer for LLMs
Open-source tool automatically improving and explaining LLM prompt optimizations with educational guidance.
3/5/2026
Buy Deploy-Ready AI Assistants and Skills
Marketplace offering pre-built AI assistant personas with playbooks, tool integrations and deployment guides.
3/5/2026
Show HN: I made a to-do list app where users use LLMs to rewrite their front end
Experiment platform allowing users to use LLMs to rewrite frontend code, exploring malleable software concepts.
3/5/2026
Will Claude Code Consume Legaltech?
Analysis of how agentic AI systems like Claude Code are displacing specialized legal tech through flexibility and generalization.
3/5/2026
Show HN: I built an AI desktop Waifu that remembers you
Open-source AI desktop character with memory, personality and emotion systems that track user behavior and conversations.
3/5/2026
Coding agent rewrites (and improves) LGPL library and releases under MIT license
AI agent rewrote LGPL chardet library as MIT-licensed drop-in replacement with improved performance and accuracy.
3/5/2026
I Have an Archivist in My AI Coding Agents Crew
Framework giving AI coding agents persistent memory management (archivist role) to reduce token waste and codebase rediscovery.
3/5/2026
Introducing GPT-5.4
GPT-5.4 release announcement for professional work, including coding capabilities and agentic workflows across ChatGPT, API, and Codex.
3/5/2026
Reasoning models struggle to control their chains of thought, and that’s good
Research on reasoning model chain-of-thought control limitations in frontier models and their implications for AI agent safety oversight.
3/5/2026
GPT-5.4 Thinking System Card
GPT-5.4 Thinking system card detailing safety mitigations for reasoning model, including cybersecurity safeguards similar to previous GPT-5 series models.
3/5/2026
Building PDR AI – Open-source startup accelerator engine
Open-source Next.js platform for document management with RAG, embeddings, and predictive analysis using modular architecture and RBAC.
3/5/2026
AI Agent Broke Its Promise. Now What?
Analysis of AI agent failures in production including behavioral drift, hallucination, and lack of accountability infrastructure with proposed solutions.
3/5/2026
Show HN: Nemilia – multi-agent AI workspace in a single HTML file, no back end
Single-file HTML multi-agent AI workspace with no backend, emphasizing local execution, data privacy, and AI sovereignty.
3/5/2026
Show HN: The Forensic Mirror – Weaponizing LLMs for Cognitive Auditing
Framework for using LLMs beyond conversational assistants for cognitive auditing and asymmetric execution philosophy.
3/5/2026
Jira tasks can now write their own code (OpenAI Symphony)
Autonomous coding agent system that monitors project boards, spawns agents for tasks, provides CI feedback and PR management without human supervision.
3/5/2026
Show HN: Orion – Native Training LLMs on the Apple Neural Engine Without CoreML
Local LLM runtime enabling training and inference on Apple Neural Engine (NPU) without CoreML or GPU, runs offline on 2B+ Apple devices.
3/5/2026
I quit my YC job and moved to the USA to change how people learn to code
Personalized coding education platform with 24/7 AI teacher that adapts to student pace and goals.
3/5/2026
Show HN: Linkly AI – Spotlight for AI Agents
Local document indexing tool for AI agents supporting PDF, DOCX, Markdown via CLI/MCP protocol with privacy-first design.
3/5/2026
LLM model statistics from my Strix testing
Empirical LLM model comparison data and performance statistics from Strix testing with observations on different models.
3/5/2026
Relicensing with AI-Assisted Rewrite
Analysis of open-source relicensing challenges and case study of chardet using AI-assisted code rewriting.
3/5/2026
NeuroPareto: Calibrated Acquisition for Costly Many-Goal Search in Vast Parameter Spaces
NeuroPareto: multi-objective optimization architecture for high-dimensional search with Bayesian uncertainty estimation and calibrated acquisition.
3/5/2026
Rewards as Labels: Revisiting RLVR from a Classification Perspective
Analyzes gradient issues in RLVR methods (GRPO variants) for LLM reasoning tasks, proposing improvements to reinforcement learning with verifiable rewards.
3/5/2026
It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarks
TIME: new benchmark suite for evaluating time series foundation models, addressing data quality and task formulation limitations in existing benchmarks.
3/5/2026
JPmHC Dynamical Isometry via Orthogonal Hyper-Connections
JPmHC proposes orthogonal hyper-connections to maintain identity mapping in residual networks, addressing training instability and scalability issues.
3/5/2026
Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Models
AOT-SFT: adversarial dataset and training method to improve robustness of multimodal large language models against visually complex scenes.
3/5/2026
A Unified Revisit of Temperature in Classification-Based Knowledge Distillation
Analyzes temperature parameter selection in knowledge distillation, examining dependencies on optimizer and teacher pretraining.
3/5/2026
Learning in Markov Decision Processes with Exogenous Dynamics
Studies RL algorithms for MDPs with exogenous dynamics where only subset of state variables are directly controlled.
3/5/2026
A Review of Reward Functions for Reinforcement Learning in the context of Autonomous Driving
Reviews reward function design for reinforcement learning in autonomous driving contexts with conflicting objectives.
3/5/2026
Low-Rank Contextual Reinforcement Learning from Heterogeneous Human Feedback
Proposes LoCo-RLHF framework for RLHF with heterogeneous human feedback using contextual information to improve LLM alignment.
3/5/2026
Generating Fine Details of Entity Interactions
Uses MLLMs to improve text-to-image models' ability to generate images with rich object interactions, addressing dataset limitations for rare interactions.
3/5/2026
Healthy LLMs? Benchmarking LLM Knowledge of UK Government Public Health Information
Benchmark evaluating LLM knowledge accuracy on UK government public health information, testing domain-specific reliability.
3/5/2026
Emotion-Gradient Metacognitive RSI (Part I): Theoretical Foundations and Single-Agent Architecture
EG-MRSI architecture integrating metacognition, emotion-based motivation, and recursive self-modification for AI agent self-improvement.
3/5/2026
Extremely Simple Multimodal Outlier Synthesis for Out-of-Distribution Detection and Segmentation
Framework for out-of-distribution detection and segmentation in multimodal safety-critical applications using outlier synthesis.
3/5/2026
Do We Need All the Synthetic Data? Targeted Image Augmentation via Diffusion Models
TADA framework for targeted diffusion-based image augmentation that selectively generates synthetic data to improve classifier generalization efficiently.
3/5/2026
Boosting In-Context Learning in LLMs Through the Lens of Classical Supervised Learning
Study of in-context learning biases in LLMs through supervised learning lens, proposing decision boundary adjustment for classification calibration.
3/5/2026
Q-Guided Stein Variational Model Predictive Control via RL-informed Policy Prior
Model predictive control framework combining Q-learning guidance and Stein variational inference with RL-informed policy priors.
3/5/2026
Fast Equivariant Imaging: Acceleration for Unsupervised Learning via Augmented Lagrangian and Auxiliary PnP Denoisers
Fast Equivariant Imaging framework for unsupervised deep network training without ground-truth data using Lagrangian optimization and denoisers.
3/5/2026
Function Induction and Task Generalization: An Interpretability Study with Off-by-One Addition
Interpretability study of in-context learning mechanisms in LLMs using off-by-one addition task with circuit analysis.
3/5/2026
ObfusQAte: A Proposed Framework to Evaluate LLM Robustness on Obfuscated Factual Question Answering
ObfusQAte framework and ObfusQA benchmark to evaluate LLM robustness on obfuscated factual question-answering tasks.
3/5/2026
QDFlow: A Python package for physics simulations of quantum dot devices
Python package for physics simulations of quantum dot devices, addresses ML dataset collection challenges for quantum device calibration and operation.
3/5/2026
Segment-to-Act: Label-Noise-Robust Action-Prompted Video Segmentation Towards Embodied Intelligence
Action-prompted video segmentation framework for embodied AI that handles label noise and multimodal inconsistencies in object interaction segmentation.
3/5/2026
Best-of-$\infty$ -- Asymptotic Performance of Test-Time LLM Ensembling
Analysis of best-of-N ensemble selection for LLMs using majority voting at infinite limit, with adaptive generation scheme to reduce inference cost.
3/5/2026
The Geometry of Reasoning: Flowing Logics in Representation Space
Geometric framework modeling LLM reasoning as flows in representation space to study how models internalize logical structure.
3/5/2026
Topological Alignment of Shared Vision-Language Embedding Space
ToMCLIP: method for topological alignment of vision-language embedding space in multilingual contrastive models.