3/4/2026
Show HN: Ungrind – the solopreneur CRM that updates itself
CRM tool for solopreneur with self-updating capabilities. Minimal content available.
3/4/2026
Show HN: CUP – MCP but for desktop UI (open spec for computer use agents)
Computer Use Protocol (CUP): universal schema for AI agents to perceive and interact with desktop UIs. Compact text encoding ~97% smaller than JSON for LLM context. Open spec for Windows, macOS, Linux, Web, Android, iOS.
3/4/2026
Which LLMs fold under pressure? We made 6 LLMs argue 300 hard cases to find out
Benchmark testing 6 LLMs under adversarial pressure across 300 cases. Evaluates model resilience in argumentation and agentic workflows beyond standard capability tests.
3/4/2026
Show HN: SFT to convert a base language model into a conversational chat model
Minimal code example demonstrating Supervised Fine-Tuning on Llama-2-7b using OpenAssistant dataset with parameter-efficient techniques to create chat model.
3/4/2026
Donx64mcp-dbg – an injected DLL debugger toolkit with an MCP server for x64 apps
Windows x64 DLL debugger toolkit with MCP server for AI agents. Provides 40+ debug commands for real-time process inspection. Designed for security research and CTF.
3/4/2026
Molmo 2: video understanding, pointing, and tracking
Molmo 2 open-source vision language model with state-of-the-art video understanding, pointing, and tracking capabilities. Hugging Face models available with training code.
3/4/2026
Slack bot coding agent built on pi (mom)
Slack bot AI agent (Mom) powered by LLM. Executes bash commands, manages files, installs tools, configures credentials autonomously. Node.js app with Socket Mode integration.
3/4/2026
LLM Gateway: Budget enforcement, virtual API keys and usage analytics for LLMs
FastAPI-based LLM gateway proxy providing budget enforcement, virtual API key management, and usage analytics across multiple LLM providers.
3/4/2026
Real-AI needs rolling aggregations and AI won't build them for you
Technical analysis of rolling aggregations as essential for real-time AI systems, covering incremental views and sub-millisecond latency approaches.
3/4/2026
APM – Agent Package Manager (Microsoft)
Open-source Agent Package Manager by Microsoft. Dependency manager for AI agents declaring skills, prompts, instructions, and tools via apm.yml configuration files.
3/4/2026
Show HN: Audicia – Generate least-privilege Kubernetes RBAC from audit log
Open-source Kubernetes operator generating least-privilege RBAC policies from audit logs. Automates security policy creation from actual access patterns.
3/4/2026
Show HN: GuardClaw – cryptographically verifiable execution logs for AI agents
GuardClaw implements cryptographically verifiable execution logs for autonomous AI agents using GEF-SPEC-1.0 protocol with append-only, immutable audit trails.
3/4/2026
What I do, and what I delegate to AI
Essay on treating AI as a leverage tool rather than productivity hack, discussing how to reshape work and decision-making with AI.
3/4/2026
Show HN: EU AI Radar – 60-second self-check for EU AI Act exposure
Static HTML quiz tool to help EU companies assess their risk tier under the EU AI Act, with news feed and regulatory tracking.
3/4/2026
AuroriaLink
Real-time team messaging solution with encryption and file sharing. In active development.
3/4/2026
Show HN: Retro – active context curator for coding agents
Open-source Rust tool that manages context for AI coding agents using git hooks and SQLite, analyzing agent conversations to optimize performance on large codebases.
3/4/2026
Nuclear War: An LLM Scenario
Minimal stub post about nuclear war scenarios with LLMs, no content provided.
3/4/2026
Show HN: Nova – AI terminal that writes, fixes, and ships your code
Nova is an AI-native developer workspace that executes code directly, eliminating the iterative chat-paste-error cycle of traditional AI coding assistants.
3/4/2026
NeuroHex: Highly-Efficient Hex Coordinate System for Creating World Models to Enable Adaptive AI
Hexagonal coordinate system for efficient world models in adaptive AI, inspired by grid cells in human brain. Mathematical framework for rotational symmetry and low-cost spatial computation.
3/4/2026
RubricBench: Aligning Model-Generated Rubrics with Human Standards
RubricBench benchmark for evaluating rubric-guided LLM reward models against human standards, addressing discriminative complexity in alignment evaluation.
3/4/2026
Nano-EmoX: Unifying Multimodal Emotional Intelligence from Perception to Empathy
Nano-EmoX framework unifying multimodal emotional intelligence across perception, understanding, and interaction levels with cognitively-inspired hierarchy.
3/4/2026
What Is the Alignment Tax?
Geometric theory formalizing alignment tax as projection in representation space, deriving Pareto frontier for safety-capability tradeoffs in LLMs.
3/4/2026
DeepXiv-SDK: An Agentic Data Interface for Scientific Literature
SDK providing LLM-agents structured data access to scientific literature via agentic interface, reducing token consumption and improving retrieval efficiency.
3/4/2026
A Novel Evolutionary Method for Automated Skull-Face Overlay in Computer-Aided Craniofacial Superimposition
Evolutionary algorithm for automated skull-face overlay alignment in forensic craniofacial superimposition using 3D skull and 2D facial image correspondence.
3/4/2026
Theory of Code Space: Do Code Agents Understand Software Architecture?
Theory of Code Space benchmark evaluating whether AI code agents understand software architecture through multi-file codebase exploration in procedurally generated environments.
3/4/2026
IDER: IDempotent Experience Replay for Reliable Continual Learning
IDER method addressing catastrophic forgetting in continual learning through idempotent experience replay with uncertainty calibration.
3/4/2026
FastCode: Fast and Cost-Efficient Code Understanding and Reasoning
FastCode system for efficient repository-scale code reasoning using selective context retrieval and compression for cost-effective LLM-based software engineering.
3/4/2026
QIME: Constructing Interpretable Medical Text Embeddings via Ontology-Grounded Questions
Proposes QIME framework for interpretable biomedical text embeddings using ontology-grounded natural language questions for clinical decision-making.
3/4/2026
Solving Inverse PDE Problems using Minimization Methods and AI
Compares numerical methods with physics-informed neural networks for solving direct and inverse PDE problems in physical/engineering systems.
3/4/2026
Rethinking Policy Diversity in Ensemble Policy Gradient in Large-Scale Reinforcement Learning
Studies policy diversity in ensemble policy gradient methods for large-scale RL, analyzing exploration-exploitation tradeoffs across parallel environments.
3/4/2026
Hyperparameter Trajectory Inference with Conditional Lagrangian Optimal Transport
Hyperparameter trajectory inference framework using conditional Lagrangian optimal transport to enable post-deployment hyperparameter adjustments without retraining.
3/4/2026
CodecFlow: Efficient Bandwidth Extension via Conditional Flow Matching in Neural Codec Latent Space
Speech bandwidth extension method using conditional flow matching in neural codec latent space for improved clarity and intelligibility.
3/4/2026
See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles
Evaluates multimodal GUI agents' ability to identify and execute toggle controls, revealing reliability bottlenecks in ubiquitous GUI interaction.
3/4/2026
Spilled Energy in Large Language Models
Reinterprets LLM softmax as energy-based model to track 'energy spills' during decoding, correlating them with factual errors and biases.
3/4/2026
Model Already Knows the Best Noise: Bayesian Active Noise Selection via Attention in Video Diffusion Model
Proposes ANSE method using Bayesian active noise selection with attention mechanisms to improve video diffusion quality by selecting optimal initial noise seeds.
3/4/2026
Navigating with Annealing Guidance Scale in Diffusion Space
Research on classifier-free guidance scale annealing in diffusion models to improve image quality and prompt alignment convergence during sampling.
3/4/2026
WebDevJudge: Evaluating (M)LLMs as Critiques for Web Development Quality
WebDevJudge benchmark evaluates LLMs-as-judges for web development quality assessment, testing reliability on open-ended tasks with dynamic environments.
3/4/2026
RxnNano:Training Compact LLMs for Chemical Reaction and Retrosynthesis Prediction via Hierarchical Curriculum Learning
RxnNano trains compact LLMs for chemical reaction prediction using hierarchical curriculum learning, emphasizing chemical intuition over parameter scaling.
3/4/2026
ATPO: Adaptive Tree Policy Optimization for Multi-Turn Medical Dialogue
ATPO uses hierarchical reinforcement learning to optimize LLM behavior for multi-turn medical dialogues with incomplete information.
3/4/2026
Is Retraining-Free Enough? The Necessity of Router Calibration for Efficient MoE Compression
Analysis of MoE compression methods identifies router-expert mismatch as key degradation factor; proposes calibration approach for efficient model deployment.
3/4/2026
Self-Play Only Evolves When Self-Synthetic Pipeline Ensures Learnable Information Gain
Research on self-play loops in LLMs showing sustainable self-evolution requires learnable information gain, not just more synthetic data generation.
3/4/2026
NExT-Guard: Training-Free Streaming Safeguard without Token-Level Labels
NExT-Guard provides training-free safeguarding for streaming LLM deployments without requiring token-level annotations or supervision.
3/4/2026
Forecasting as Rendering: A 2D Gaussian Splatting Framework for Time Series Forecasting
Novel 2D Gaussian Splatting approach for time series forecasting that reshapes 1D sequences to preserve chronological continuity.
3/4/2026
MedFeat: Model-Aware and Explainability-Driven Feature Engineering with LLMs for Clinical Tabular Prediction
MedFeat integrates LLM domain knowledge into feature engineering for clinical tabular prediction, balancing model characteristics with feature importance signals.
3/4/2026
MedCalc-Bench Doesn't Measure What You Think: A Benchmark Audit and the Case for Open-Book Evaluation
Audit of MedCalc-Bench clinical calculator benchmark reveals implementation issues and proposes open-book evaluation methodology for more accurate LLM assessment.
3/4/2026
Characterizing and Predicting Wildfire Evacuation Behavior: A Dual-Stage ML Approach
ML research paper using correspondence analysis, clustering, and classification to model wildfire evacuation behavior from survey data.
3/4/2026
Subspace Geometry Governs Catastrophic Forgetting in Low-Rank Adaptation
Geometric theory of catastrophic forgetting in LoRA through gradient subspace interactions, deriving quantitative forgetting formula.
3/4/2026
Scaling Reward Modeling without Human Supervision
Unsupervised reward modeling scaling via preference learning on web document prefixes/suffixes, reducing human annotation costs.
3/4/2026
Efficient Sparse Selective-Update RNNs for Long-Range Sequence Modeling
Efficient RNN architecture with selective state updates for long-range sequence modeling, reducing unnecessary computation on static inputs.
3/4/2026
Neural Paging: Learning Context Management Policies for Turing-Complete Agents
Neural Paging architecture enabling Turing-complete agents by learning hierarchical context window management policies.