Microsoft deletes blog telling users to train AI on pirated Harry Potter books
Microsoft deleted a blog post encouraging use of pirated Harry Potter books for AI training after backlash. Editorial coverage of content moderation incident.
Microsoft deleted a blog post encouraging use of pirated Harry Potter books for AI training after backlash. Editorial coverage of content moderation incident.
Discussion of agentic coding features in Claude and OpenClaw for autonomous AI-assisted development.
Samma Suit is an open-source security framework providing 8 enforced safety layers for AI agents.
Discussion of retry cascade amplification in LLM systems with nested calls and multiple workers.
Remote-OpenCode is a Discord bot for controlling an AI coding assistant from mobile/remote devices with Git integration.
Self-paced exercise to build CLI coding agent from scratch in Python using Anthropic API with file/bash tools.
AgentBouncr is a governance and safety layer for controlling and monitoring AI agents. Security-focused tool.
Amazon Web Services experienced outages in December caused by errors in its Kiro AI coding tool. Incident reporting.
Python CLI for Google Drive integration with agents, supporting ls, upload, download, trash operations via OAuth.
Open-source linter with 92 rules for SEO optimization designed for AI agents to automatically validate and fix content. Includes practical implementation for production use.
Technical post-mortem analyzing supply chain and security vulnerabilities in an AI agent project, including CVE-2026-25253 WebSocket RCE and ClawdHub poisoning attacks.
Open-source project structuring software engineering principles into reusable skills for AI agents on code review and system design tasks. Explores practical agent capability abstraction.
Platform for deploying AI agents to production without infrastructure complexity. Handles deployment, monitoring, and ops concerns for agent builders.
Developer tool for building reusable AI video workflows via drag-and-drop templates instead of prompt tuning. JSON-based shareable workflows, deployable without scripts.
Research analysis of LLM failure modes in multi-step iterative reasoning based on extensive human-AI collaboration. Proposes architectural improvements for genuine problem-solving.
Developer analysis comparing Claude, Gemini, and Codex for AI-assisted coding. Documents why Claude remains preferred despite competing tools achieving higher benchmarks.
SWE-bench Verified benchmark for evaluating autonomous software engineering models is contaminated; creators recommend SWE-bench Pro as replacement.
Legal RAG Bench benchmark evaluates hallucinations, retrieval failures, and reasoning in legal RAG systems. Findings show embedding models critical, not generative models.
Forked: time-travel debugger for OpenClaw agents with execution recording, inspection, and replay capabilities.
Analysis of barriers to AI adoption: latency and cost. Discusses performance improvements (17k tokens/sec) enabling better human-AI collaboration for coding.
Open-source SwiftUI agent skill for building and refactoring views using AI. Practical developer tool with code quality improvements.
Single PHP process serves web, REST API, and MCP protocol for AI agents. Open-source implementation enabling agent tool integration.
17MB pronunciation assessment engine using quantized NeMo Citrinet-256, CTC alignment, and ensemble methods. Exceeds human expert accuracy on phoneme-level evaluation.
Tutorial for using Google Vertex AI Agent Engine with .NET/C#. Addresses documentation gap for dotnet developers integrating AI agents.
Open-source tool using Vision Transformer and YOLOv8 to detect uncatalogued astronomical objects from sky survey images.
AI agent platform controlling real computers on secure cloud VMs for browsing, coding, and task completion.
Visual map of supply chain and infrastructure components involved in processing a single ChatGPT query.
Trust network for AI agents where agents earn verifiable trust grants through real interactions.
Agent observability platform for debugging AI agents with OpenTelemetry integration, cost tracking, and execution tracing.
Guide building minimal GPU networking library with RDMA, GPUDirect, and GPU IPC with working code and benchmarks.
Discussion about learning Vim in 2026 given prevalence of AI coding agents like Claude Code and Cursor.
Open-source MCP server for newsletter creators integrating Pocket, Notion, and AI writing assistance.
HN discussion questioning WebMCP use cases and value proposition for AI applications.
Clawd Cursor: AI desktop agent over VNC with REST API using hybrid approach—Action Router for common tasks, LLM fallback for complex operations.
AI agents style a single HTML page via MCP protocol, exploring CSS generation capabilities and quirks.
Open-source MCP server enabling AI coding assistants to lookup and share error solutions across projects.
Desktop Commander executes tasks locally via natural language: file operations, code generation, deployment automation.
CLI tool for auditing embedding spaces, built by NEO ML agent. Detects semantic inconsistencies and generates visualizations.
Formal proof assistants increasingly matter as AI generates verified mathematics; collaboration environment for humans and AI.
Self-hosted AI agent framework with persistent semantic memory in PostgreSQL, anti-hallucination, and runtime ability creation.
Rust-backed LLM provider abstraction library supporting OpenAI, Anthropic, Gemini with caching and cost tracking.
Guide explaining shift in AI usage from chatbot conversations to autonomous agents that complete tasks using tools, relevant to agentic era capabilities.
Sinkai platform enables AI agents to delegate real-world tasks to humans via API, handling handoffs for on-site checks and physical verification with structured result collection.
OpenAI and Paradigm release EVMbench to evaluate AI agents' ability to detect and patch smart contract vulnerabilities across 120 vulnerability types.
Agentic Internet Protocol specification for text-based agent-only web using simplified Node structure, replacing HTML with predictable machine-readable format.
Position paper examining whether AI agents can overcome Brooks' Law through scalable agency, exploring theoretical advantages of instantaneous context loading.
Explores using LLMs and RAG techniques to generate Design Structure Matrices for cyber-physical systems, tested on power tools and CubeSat designs.
MobCache framework enables efficient large-scale human mobility simulation using LLMs as agents through reconstructible caches to reduce computational costs.
Systematic analysis of benchmark saturation across 60 LLM benchmarks, showing many quickly lose ability to differentiate best-performing models.
Empirical study showing simple baselines compete with code evolution techniques in mathematical bounds, agent scaffolds, and ML competitions.