3/18/2026
Show HN: Trickle – See PyTorch tensor shapes inline in VSCode as you code
VSCode extension displaying PyTorch tensor shapes inline at runtime without manual type annotations, aiding ML model development.
3/18/2026
Show HN: Motif - Analyze your Cursor and Claude Code chat history
Tool analyzing Cursor and Claude Code chat histories, extracting and reporting on AI-assisted coding conversations as work credentials.
3/18/2026
Agentic review of Linux Kernel code changes
Application of AI agents for automated code review of Linux kernel patches.
3/18/2026
Show HN: Packet28 – Context Layer for AI coding agents
Open-source MCP + hook system that optimizes context management for coding agents by reducing token budget waste from repetitive tool output.
3/18/2026
Ask HN: How do you distinguish prompt issues from model failures in LLMs today
Discussion on debugging strategies to distinguish between prompt engineering issues and underlying model failures in LLM applications.
3/18/2026
Ask HN: How do you use AI agents in your day to day life?
Discussion thread asking users about their workflows and use cases with AI agents.
3/18/2026
Google Engineers Launch "Sashiko" for Agentic AI Code Review of the Linux Kernel
Google engineers develop Sashiko, agentic AI system for Linux kernel code review. AI agent for software engineering.
3/18/2026
Show HN: Hanoi-CLI – simulate and optimize pod placement in Kubernetes
CLI tool for analyzing Kubernetes pod placement and suggesting optimization. DevOps tool, not AI-related.
3/18/2026
Top AI coding tools make mistakes one in four times, study shows
Study shows top AI coding tools make mistakes ~25% of the time. LLM evaluation and reliability assessment.
3/18/2026
Show HN: TengineAI – Infrastructure for running AI tools in production
Production infrastructure for running and securing AI agent tools with permissions and observability. Developer tool for safe tool execution in AI systems.
3/18/2026
We optimized Dropbox Dash's relevance judge with DSPy
Dropbox optimized relevance judging for search using DSPy framework. LLM application with systematic approach to ranking and evaluation at scale.
3/18/2026
Frihet MCP Server – 35 tools to let AI agents run your business operations
MCP server with 35 business tools for AI agents via natural language. AI agent tooling for business operations automation.
3/18/2026
LLMs predict my coffee: Why not benchmark with physical experiments?
Benchmark testing LLM capabilities on physical experiment prediction. LLM evaluation on real-world physics problems.
3/18/2026
Exogram Protocol RFC: Server Side IAM for AI Agents
RFC for server-side IAM protocol governing autonomous AI agents. Infrastructure specification for production agent safety and governance.
3/18/2026
Show HN: BendClaw – Distributed AgentOS Written in Rust
Open-source distributed AgentOS in Rust enabling multiple agents with shared knowledge layer. Production AI agent infrastructure with distributed compute.
3/18/2026
Show HN: Save Claude tokens with semantic search powered by SQLite and Ollama
Lumen: local semantic code search engine for AI agents using Ollama, SQLite, and embedding models. Reduces token costs up to 39%.
3/18/2026
Show HN: Yapper – free, open-source alternative to Superwhisper for macOS
Yapper: open-source macOS dictation app using Whisper for transcription and optional LLM post-processing.
3/18/2026
Permit MCP Gateway, a new trust and enforcement layer for MCP
Permit MCP Gateway adds identity, authorization, and runtime controls to Model Context Protocol for AI agents.
3/18/2026
Show HN: Xybrid – run LLM and speech locally in your app (no back end, Rust)
Xybrid: Rust library for embedding LLM and speech pipelines locally in applications without server backend. Supports GGUF, ONNX, CoreML with Flutter and Swift integration.
3/18/2026
Daemons that clean up the mess agents leave behind
Discussion on daemon processes to manage cleanup from AI agent operations. Limited technical detail in title.
3/18/2026
Ask HN: How do you handle payments for AI agents?
Ask HN: Community discussion on payment and billing management for AI agents calling APIs. No technical solution provided.
3/18/2026
Show HN: SentrySearch – Semantic search over dashcam footage
Semantic search CLI for dashcam footage using Gemini Embedding 2 for cross-modal video-to-text retrieval.
3/18/2026
Show HN: AgentPay – Let AI agents pay for APIs autonomously
AgentPay: Payment infrastructure for AI agents enabling autonomous API calls with credit limits, spending controls, and transaction tracking.
3/18/2026
Qianfan-OCR – 4B open-source VLM replacing multi-stage OCR pipelines
Qianfan-OCR: 4B-parameter open-source vision-language model for document understanding and image-to-Markdown conversion.
3/18/2026
Nvidia NemoClaw
Nvidia NemoClaw open-source stack for running OpenClaw autonomous assistants safely with sandbox orchestration.
3/18/2026
Show HN: Agent Trust – Cryptographic identity and reputation for AI agents
Agent Trust: System providing cryptographic identity and reputation mechanisms for AI agents. Minimal details in snippet.
3/18/2026
Show HN: Habits, an open-source builder and runtime for automations and apps
Habits: Open-source graph-based system for building automations, AI agents, and apps from unified model. Exports to Docker, binaries, desktop and mobile apps.
3/18/2026
Why AI agents can reduce productivity
Discussion on potential productivity drawbacks of AI agents. Limited substantive technical content.
3/18/2026
Show HN: Lukan – An open-source agentic workstation in a single Rust binary
Lukan: Open-source MIT-licensed agentic workstation as single Rust binary with zero dependencies. Includes secure remote access, file management, and integrated AI agents.
3/18/2026
Show HN: A 30-minute course to get up to speed on how AI agents work
30-minute interactive course teaching AI agent fundamentals through 60 lines of Python, browser-based.
3/18/2026
Show HN: PlanckClaw an AI agent in 6832 bytes of x86-64 assembly
PlanckClaw: minimal AI agent implementation in 6.8KB of x86-64 assembly for tool-calling workflows.
3/18/2026
Show HN: Permit MCP Gateway – Fine-Grained Authorization and IGA for MCP
Permit MCP Gateway: authorization proxy for Model Context Protocol servers with fine-grained access control.
3/18/2026
Show HN: On-device meeting transcription for your Mac
On-device meeting transcription app for macOS using Apple Core Audio APIs and local processing.
3/18/2026
570k Lines of LLM Code Compiled Fine. It Was 20,171x Slower Than SQLite
Benchmark showing LLM-written Rust SQLite reimplementation compiled correctly but ran 20,171x slower than original.
3/18/2026
Ask HN: How is your company managing internal AI agents?
Discussion thread on operational management of custom internal AI agents in companies.
3/18/2026
Is there an AI garage startup path?
Analysis of startup viability in AI foundation models given computational and capital barriers.
3/18/2026
Show HN: Atria – terminal UI for managing multiple coding agents
Terminal UI multiplexer for managing multiple coding agent sessions with status visibility and quick controls.
3/18/2026
Build Memory-Aware Agents
Article about building agents with memory capabilities. Title only, minimal content provided.
3/18/2026
Mining your team's PR reviews into automated code review rules
Article about extracting code review patterns from PR history to create automated review rules.
3/18/2026
Build AI Agents for Elevation, Not Replacement
Opinion piece on designing AI agents for human capability enhancement rather than replacement.
3/18/2026
Show HN: SHTMLs – HTML pastebin where the AI uploads its own output
SHTMLs: HTML pastebin for sharing Claude-generated outputs with programmatic API for AI agents.
3/18/2026
Show HN: Deploybase CLI – Search GPU and LLM pricing from your terminal
CLI tool for searching and comparing GPU and LLM pricing across 100+ cloud providers with live data.
3/18/2026
Show HN: Vibe – a language for humans and AI to reason about programs together
Vibe language for Scheme-to-LLVM compilation with AI assistance for co-reasoning about programs.
3/18/2026
Show HN: Ranvier – Prefix-aware routing for LLM inference
Ranvier enables prefix-aware routing optimization for LLM inference to improve performance.
3/18/2026
Detecting When Your AI Agent Dies
Title only, no content provided. Likely about monitoring AI agent failures or health checks.
3/18/2026
Why Agent UIs Lose Messages on Refresh
Technical explanation of UI state management issues when AI agent chat interfaces refresh and lose message history.
3/18/2026
Running 9 AI agents and realizing we needed a control plane
Experience report on operating multiple AI agents and identifying the need for a centralized control plane.
3/18/2026
Show HN: Ibkr-CLI – A local-first Interactive Brokers CLI built for AI agents
Ibkr-cli is a CLI tool for Interactive Brokers designed for AI agents to execute trading actions programmatically without the legacy desktop interface.
3/18/2026
Veto: Permission policy engine and LLM firewall for AI coding agents
Veto is a permission policy engine and LLM firewall designed to control actions in AI coding agents.
3/18/2026
Heaps do lie: debugging a memory leak in vLLM
Technical debugging guide for memory leak in vLLM, a widely-used LLM inference optimization library.