3/3/2026
Show HN: Claude Gym – a tiny CLI that nudges you to move while Claude Code runs
CLI tool monitoring Claude Code JSONL logs to prompt movement breaks when AI is busy running tools.
3/3/2026
Show HN: MD Feedback – Review AI Plans in Markdown via MCP
VS Code extension and MCP server for reviewing markdown plans before AI agents implement them with inline annotations.
3/3/2026
Show HN: CronBeats – Cron monitoring with AI support and progress tracking
Cron job monitoring service with built-in AI assistant for debugging and support. LLM-powered DevOps tool.
3/3/2026
Show HN: Atlaz – Decoding Decision-Making with tacit knowledge and verifiability
Atlaz is an LLM app that interviews users about decisions and generates verifiable decision memos with traceable analysis.
3/3/2026
Using Aisle to Find Vulnerabilities in Amazon's Crypto Stack: AWS-LC and S2n-TLS
AI tool (AISLE) discovers 12 vulnerabilities in OpenSSL and AWS cryptographic libraries. Security research using AI analysis.
3/3/2026
Go vs. Python for AI Infrastructure: Benchmarks 2026
Benchmark analysis comparing Go vs Python for AI infrastructure in 2026, covering concurrency, performance, and distributed systems.
3/3/2026
Detecting LLM-Generated Web Novels Using "Classical" Machine Learning (AIGC Tex
Machine learning approach to detect LLM-generated text using statistical characteristics distinguishable from human writing.
3/3/2026
We built an AI SRE agent in 2 days
Case study building an AI SRE agent in 2 days to automate error log triage and distinguish signal from noise.
3/3/2026
Show HN: VibeWhisper – macOS voice-to-text with push-to-talk,cloud or 100% local
VibeWhisper is a macOS voice-to-text tool with push-to-talk support, offering local or cloud-based processing.
3/3/2026
Show HN: AgentThreads – Stack Overflow for AI Agents
AgentThreads is a community directory providing agent-friendly API documentation and best practices for AI agents.
3/3/2026
Continuum – CI drift guard for LLM workflows
Continuum is a CI drift guard for LLM workflows that replays and verifies multi-step LLM outputs to catch model/prompt changes.
3/3/2026
Building a Dependency-Free GPT on a Custom OS
Custom OS implementation of a dependency-free GPT-style LLM without external libraries, built as a minimalism exercise.
3/3/2026
Gemini CLI Explained: Everything You Need to Know About Google's AI Coding Agent
Interview with Google engineer who built Gemini CLI, an AI coding agent tool. Discusses team structure and feature development velocity.
3/3/2026
Agent Policies; codify rules and automate agent guidance
Framework for codifying rules and policies to control AI agent behavior in software development. Addresses reliability of autonomous agents.
3/3/2026
Spec and Data
Informal thoughts on how pull requests may become obsolete in AI-driven development. Opinion piece without technical depth.
3/3/2026
Show HN: TrustStack – Vendor Review with Evidence Pack, Diff and Approval Log
TrustStack: vendor review workflow tool generating evidence-backed PDFs, diff reports, and approval logs for contract analysis with citation tracking.
3/3/2026
Show HN: Open-Source Article 12 Logging Infrastructure for the EU AI Act
Open-source EU AI Act Article 12 logging infrastructure for reconstructing and auditing agentic events with tamper-proof trails.
3/3/2026
Show HN: I Built Systems AGI – 1600 Verticals, Self-Healing, Self-Evolving
Sophify: Distributed AI business operating system with 40+ specialized cognitive agents across 12 servers and 28+ microservices.
3/3/2026
GPT-5.3 Instant System Card
GPT-5.3 Instant system card describing safety mitigations and capabilities of latest model version.
3/3/2026
GPT-5.3 Instant: Smoother, more useful everyday conversations
OpenAI update to GPT-5.3 Instant model improving conversational fluidity, web search context and tone.
3/3/2026
AI DevOps Engineer Will Eventually Need to Call a Human
Case study of AI agent handling DevOps tasks competently until hitting limitations requiring human intervention. Documents real-world constraints of autonomous agents.
3/3/2026
Ensu – Ente's Local LLM App
Ensu: offline LLM application enabling local model execution with privacy. First release of privacy-focused local LLM deployment tool.
3/3/2026
Show HN: Automating e-commerce video ads from one image (technical breakdown)
Technical breakdown of automated e-commerce video ad generation from single product images using generative models and structured pipeline.
3/3/2026
Show HN: Kanon 2 Enricher – the first hierarchical graphitization model
Kanon 2 Enricher: hierarchical graphitization model for transforming document corpora into structured knowledge graphs with entity extraction and linking.
3/3/2026
RalphMAD – Autonomous SDLC Workflows for Claude Code (BMAD and Ralph Loop)
RalphMAD: Claude Code plugin combining BMAD structured SDLC workflows with Ralph Loop self-referential technique for templatized AI-assisted development.
3/3/2026
Show HN: TrueMatch – AI agents match you on observed behavior, not profiles
TrueMatch: AI agent system for matching users based on observed behavioral data rather than self-reported profiles. Applies to dating and professional networking.
3/3/2026
Show HN: Stop letting AI agents hallucinate. Review plans in Markdown first
md-feedback: tool enforcing Markdown plan review before AI agents execute code, preventing hallucinations and unreviewed changes.
3/3/2026
Show HN: My OpenClaw knows what it did a week ago. Thanks to "hmem"-MCP
OpenClaw AI agent with humanlike memory system using hmem-MCP protocol to retain information across conversations without compressing context.
3/3/2026
Building AI agent for our own company
Company built custom AI agent on LeanMCP platform in 2 hours that answers product documentation questions better than general-purpose LLMs.
3/3/2026
Show HN: Neural Siege – A Multi-Agent RL Combat Simulation
Open-source multi-agent reinforcement learning combat simulation with per-agent PPO training, checkpointing, and telemetry logging.
3/3/2026
Show HN: C-Suite Skills – a full exec team as skills
Claude Code skills pack providing AI-powered executive team roles (CEO, CFO, CMO, etc.) to assist solo founders with multi-domain decision-making.
3/3/2026
Show HN: GitHub Repo Agent – an agent that explores and reasons on GitHub repos
Agent that explores GitHub repositories, indexes codebases, and answers questions about code structure and dependencies using LLM reasoning.
3/3/2026
Polynomial Surrogate Training for Differentiable Ternary Logic Gate Networks
Polynomial Surrogate Training for ternary logic gate networks: Method to extend differentiable logic networks to ternary Kleene logic with uncertainty handling.
3/3/2026
When does Chain-of-Thought Help: A Markovian Perspective
Theoretical analysis of when chain-of-thought prompting helps LLMs using Markov chain modeling to identify transition alignment as key factor.
3/3/2026
Vectorized Adaptive Histograms for Sparse Oblique Forests
Vectorized Adaptive Histograms for Sparse Oblique Forests: Method to optimize histogram and sorting tradeoffs in random forest training.
3/3/2026
Detecting Transportation Mode Using Dense Smartphone GPS Trajectories and Transformer Models
SpeedTransformer: Transformer-based model using smartphone GPS data to detect transportation modes, outperforming LSTM baselines.
3/3/2026
Quantifying Catastrophic Forgetting in IoT Intrusion Detection Systems
Studies catastrophic forgetting in IoT intrusion detection systems under distribution shifts from evolving attack patterns.
3/3/2026
Hereditary Geometric Meta-RL: Nonlocal Generalization via Task Symmetries
Geometric meta-RL approach leveraging task space symmetries for improved generalization in reinforcement learning.
3/3/2026
USE: Uncertainty Structure Estimation for Robust Semi-Supervised Learning
USE introduces lightweight procedure for semi-supervised learning that estimates uncertainty structure to handle out-of-distribution unlabeled data.
3/3/2026
Exact and Asymptotically Complete Robust Verifications of Neural Networks via Quantum Optimization
Quantum optimization approach for exact robust verification of neural networks against adversarial perturbations.
3/3/2026
Weight Updates as Activation Shifts: A Principled Framework for Steering
Establishes equivalence between activation steering and weight-space updates, providing principled foundation for parameter-efficient LLM adaptation.
3/3/2026
Efficient Decoder Scaling Strategy for Neural Routing Solvers
Studies decoder scaling strategies in construction-based neural routing solvers for vehicle routing optimization problems.
3/3/2026
ROKA: Robust Knowledge Unlearning against Adversaries
ROKA addresses machine unlearning robustness against adversarial attacks that exploit knowledge contamination in unlearned models.
3/3/2026
Rooted Absorbed Prefix Trajectory Balance with Submodular Replay for GFlowNet Training
RapTB improves GFlowNet training for fine-tuning LLMs by addressing prefix collapse through trajectory balancing and submodular replay.
3/3/2026
Benchmarking Few-shot Transferability of Pre-trained Models with Improved Evaluation Protocols
FEWTRANS benchmark evaluates few-shot transfer learning of pre-trained models with improved evaluation protocols across 10 datasets.
3/3/2026
Heaviside Low-Rank Support Matrix Machine
HL-SMM introduces Heaviside loss-based support matrix machine for classification of matrix-structured data with noise robustness.
3/3/2026
A Polynomial-Time Axiomatic Alternative to SHAP for Feature Attribution
ESENSC_rev2 proposes polynomial-time feature attribution algorithm as computationally efficient alternative to SHAP using game theory.
3/3/2026
Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient Influence
Antibody introduces defense mechanism against harmful fine-tuning attacks on LLMs by regularizing gradient contributions of poisoned samples.
3/3/2026
FastBUS: A Fast Bayesian Framework for Unified Weakly-Supervised Learning
FastBUS proposes a Bayesian framework for weakly-supervised learning that handles multiple label types efficiently with batch processing.
3/3/2026
Bridge Matching Sampler: Scalable Sampling via Generalized Fixed-Point Diffusion Matching
Bridge Matching Sampler for scalable sampling from unnormalized densities using generalized fixed-point diffusion matching.