AgriPath: A Systematic Exploration of Architectural Trade-offs for Crop Disease Classification
AgriPath systematically compares CNNs, vision-language models, and generative VLMs for crop disease classification across diverse conditions.
AgriPath systematically compares CNNs, vision-language models, and generative VLMs for crop disease classification across diverse conditions.
VisionClaw presents wearable AI agent on Meta Ray-Ban glasses with egocentric perception and speech-driven task execution via OpenClaw agents.
CodecFlow optimizes streaming video analytics for vision-language models by exploiting temporal/spatial redundancy end-to-end, reducing multimodal inference costs.
AIPOCH Medical Skill Auditor evaluates medical research agent capabilities with veto gates for quality control.
App Store submissions surged 30% in 2024, driven by AI coding tools enabling app development.
CVD compression algorithm achieves 99.9% compression on AI art images, offers lossless and lossy modes.
Implementation enabling Rust's std::thread to run on GPUs, advancing GPU-native software development with familiar language abstractions.
MemForge system enables persistent long-term agent memory using PostgreSQL and multi-tier storage, scores 92% on LongMemEval.
Viatoris provides cryptographic audit trails and signed receipts for AI agent actions in enterprise compliance scenarios.
Claude Code plugin for research synthesis with typed claims and conflict detection. LLM application with zero dependencies.
Image CDN designed for autonomous AI agent access without human interaction. Developer tool specifically for AI agents.
2500 vision language model benchmarks and evaluations dataset. Machine learning research evaluation resource.
HN discussion comparing LLM-inferred vs deterministic approaches for codebase understanding in AI agents. Architecture question for agent tools.
Formal reasoning engine for LLM code analysis and structural understanding. LLM application combining symbolic reasoning with code analysis.
AI-powered code review personas based on open-source maintainers' PR review patterns. LLM application for code review.
CyberAgent case study on ChatGPT Enterprise adoption with 93% monthly active usage rates.
Case study: CyberAgent adopts ChatGPT Enterprise and Codex for team productivity and decision-making across businesses.
Infer is a lightweight bash-based agent harness for running local models. Unix-pipe friendly agent framework emphasizing simplicity.
Composer 2.0 uses MCP to connect AI coding agents (Claude, Cursor, Copilot) and auto-generates architecture diagrams from codebases.
Prefab is a generative UI framework for Python built on MCP protocol. Enables Python servers to ship interactive UIs into conversations.
Local SEO analysis agent using Qwen3-8B LLM to scrape URLs and generate PDF reports. LLM application with document generation.
Canvas interface enabling multiple AI agents to collaborate as a design team for creating marketing and design assets.
Open source local-first observability platform for autonomous AI agents. Tracks costs, API calls, and agent actions without cloud dependency.
CLI tool using AI to analyze multi-service logs and explain production incidents and failure propagation.
Open-source protocol providing unified interface to switch between LLM providers while maintaining local encrypted conversation history.
Coz is an open source profiler for C/C++/Rust using causal profiling to measure optimization potential and throughput impact. Developer tool.
Educational guide explaining AI tokens and context window limitations across Claude, GPT-4o, Cursor, and GitHub Copilot.
Open source multi-agent job search tool using AI agents to evaluate job listings and generate personalized CVs. CLI-based job search automation.
Open-source multi-agent orchestration harness in Go with dashboard, chat, and kanban board for coordinating AI teams.
Context-aware moderation tool for Twitch using ML to understand stream context, chat mood, and streamer behavior for intelligent moderation.
Benchmark comparing Google's Gemma 4 E4B model against Gemma family across 8 enterprise task suites on Apple Silicon.
Developer tool providing lightweight dev environments using Dockerfile and Justfile without Node.js or VS Code plugin overhead.
Open-source CLI mapping codebases into persistent architectural graphs enabling AI agents to retain system knowledge across sessions.
CLI tool providing control layer for AI coding assistants to restrict which files and scopes can be modified during code generation.
Using agentic AI to resurrect and continue development of a 1992 text-based multiplayer game (MUD), leveraging modern language models for NPC and world generation.
Linggen: Open-source Rust-based AI coding agent with P2P remote access via WebRTC, plan mode for code execution approval, and support for multiple LLM providers.
arXiv research paper on how LLM fine-tuning activates verbatim recall of copyrighted book content.
Local data lake IDE for AI-powered analytics and data engineering with SQL/Python support, natural language querying, and zero cloud requirements.
CongaLine open-source tool for self-hosted isolated AI agent fleet using Docker containers with focus on security and team deployment without shared instances.
Developer used 15 AI agents with Claude to design adaptive wearable footwear, documenting where agents succeeded and failed in product design workflow.
AMD AI director reports Claude Code performance degradation since February update, raising concerns about reliability for complex tasks.
Exploration of Spiking Neural Networks as transformer alternatives for brain-inspired AI in C/C# without external libraries.
iOS app providing private health insights from Apple Health using user-supplied Claude, OpenAI, or Gemini API keys for data analysis.
Gimble is an open-source CLI tool for debugging physical systems by capturing logs and telemetry with AI-grounded analysis.
Documentation of system failure and forum moderation issues with Cursor AI agent tool on high-end workstation.
Analysis of copyright and licensing implications when LLM-generated code is integrated into open source projects, raising concerns about value loss.
Analysis identifying specific constraints under which LLMs minimize hallucination: extended thinking mode, bounded context, text-only input.
Modular AI agent system with 6 specialized agents that builds searchable knowledge graphs from RSS, papers, GitHub. Configuration-driven, built on Claude Code.
Druids is an open-source library for structuring and running multi-agent coding workflows with abstracted VM infrastructure and agent communication.
Essay comparing corporate AI adoption mandates to China's Great Leap Forward, critiquing unrealistic AI transformation expectations.