AI Agents Blog
Deep dives and practical articles about AI agents — agent frameworks, orchestration patterns, MCP integrations, and autonomous workflows. Curated from our main blog: every card links to the full article. 840 articles and growing.
Pie AI Hits Major Milestone with Pie 1.0 Launch and Over $100M in Funding
Pie AI, the innovative startup from ex-DeepMind experts, celebrates a huge win: launching Pie 1.0, their powerful AI agent platform, backed by a whopping $106M seed round. Discover how it's transforming enterprise workflows.
Model See, Model Do: Revolutionizing AI with Imitation Learning from Screen Videos
Discover how vision-language models like GPT-4V can master complex tasks by simply watching expert AI agents on screen. This breakthrough in imitation learning boosts performance on benchmarks like OSWorld without needing massive datasets.
Supercharge Multi-Agent AI: Tool Namespaces to Stop Bots from Clashing!
Discover how tool namespaces empower multiple AI agents to collaborate seamlessly without interference. Scale your bot armies like never before with LangGraph's game-changing approach!
Drones Unleashed: LLMs Powering Autonomous Aerial Swarms in Real-World Missions
Discover how large language models are transforming drones into intelligent, self-governing teams for disaster response, navigation, and swarm operations. From Microsoft to EPFL, see the cutting-edge projects revolutionizing robotics!
Autonomous Weapons See Rising Support in Ukraine: Key Insights from Latest Polls and AI Developments
Support for AI-controlled killer drones surges to 52% in Ukraine amid ongoing conflict. Explore this shift, new AI benchmarks from Meta, and updates on models like Grok in The Batch Issue #198.
Meta's Llama 3.1 405B Shatters Benchmarks: Highlights from The Batch Issue #88
Dive into Meta's game-changing Llama 3.1 405B model that tops major AI benchmarks, OpenAI's new Realtime API for voice, and Google DeepMind's agent innovations. Essential reads for AI enthusiasts.
Crafting an Autonomous Multi-Agent System for Data Pipelines and Infrastructure Strategies with Efficient Lightweight Qwen Models
Discover how to build a smart, self-managing multi-agent system using compact Qwen2.5 models to optimize data workflows and infrastructure decisions. Perfect for developers seeking efficient AI-driven intelligence without heavy resources.
Microsoft Launches Agent Lightning: Empowering RL Training for LLM-Based AI Agents
Discover Agent Lightning, Microsoft's innovative framework that makes reinforcement learning (RL) training for large language models (LLMs) in AI agents faster, cheaper, and more stable. Perfect for developers building advanced agents.
Constructing Ethically Robust Autonomous Agents: Value-Driven Reasoning and Adaptive Self-Correction with Open-Source Large Language Models
Explore a comprehensive framework for developing autonomous agents that embed ethical values into their reasoning processes and employ self-correction for reliable decision-making, leveraging powerful open-source LLMs like Llama and Mistral.
MiniMax Unveils Open-Source MiniMax-M2: A 3B Powerhouse for Coding and Agentic Tasks at 1/8th Claude 3.5 Sonnet Cost and 2x Speed
MiniMax just open-sourced MiniMax-M2, a compact 3B model crushing coding benchmarks and agentic workflows while costing a fraction of Claude 3.5 Sonnet and running twice as fast.
Building an Agentic Decision Tree RAG System: Intelligent Query Routing, Self-Checking, and Iterative Refinement Tutorial
Discover how to create a sophisticated RAG system using decision trees for smart query handling, automatic error correction, and continuous improvement. This guide uses LangGraph and LlamaIndex for robust AI-driven retrieval.
Create a Powerful Computer-Use AI Agent: Reasoning, Planning, and Virtual Execution with Local LLMs
Discover how to construct a robust AI agent that autonomously thinks, strategizes, and performs browser actions using entirely local AI models—no cloud dependency required.
Google vs. OpenAI vs. Anthropic: Unpacking the Fierce Race in Agentic AI Development
Discover how Google, OpenAI, and Anthropic are pushing the boundaries of agentic AI—autonomous systems that think, plan, and act. This breakdown compares their latest tools, benchmarks, and what's next in the arms race.
Salesforce AI Research Launches WALT: Web Agents That Autonomously Learn Reusable Tools from Any Website
Discover WALT, Salesforce AI's groundbreaking framework that empowers LLM agents to automatically extract and reuse tools from websites, revolutionizing web automation without manual API engineering.
UltraCUA: The Groundbreaking Foundation Model Uniting GUI and API Agents for Smarter Computer Use
Discover UltraCUA, the revolutionary foundation model that seamlessly blends general-purpose GUI agents with specialized API agents, unlocking unprecedented automation in real-world computing tasks!
OpenAI Unveils ChatGPT Atlas: A Chromium-Powered Browser with Embedded AI Agent for Smarter Browsing
OpenAI's ChatGPT Atlas redefines web browsing by integrating GPT-4o directly into a Chromium-based browser, enabling AI-driven task automation like form filling and bookings without leaving the page.
Creating a Collaborative Multi-Agent AI Framework with AutoGen, LangChain, and Hugging Face: Practical Guide to Agentic Workflows
Learn to build an intelligent system where AI agents team up to research, code, execute, and review tasks autonomously using AutoGen, LangChain tools, and Hugging Face models for real-world agentic AI applications.
Discovering LangChain's DeepAgents Library: In-Depth Guide and Hands-On Example of Multi-Layer Agent Systems
Dive into LangChain's DeepAgents library, a powerful tool for creating sophisticated research agents with hierarchical structures, memory, and tool integration. Explore a complete practical example to see it in action.
Mastering Context Engineering: A Comprehensive Guide to Empowering AI Agents
Discover how context engineering elevates AI agents beyond basic prompts, enabling smarter decision-making and task execution in complex environments. This guide covers principles, techniques, and tools for optimal performance.
Building Adaptive AI Systems: Hands-On Guide to Model Context Protocol (MCP) for Seamless Real-Time Tool and Resource Integration
Discover how the Model Context Protocol (MCP) revolutionizes AI development by enabling dynamic, real-time integration of tools and resources. This guide walks you through practical implementations with code examples to supercharge your AI apps.
W4S: Training Weak Meta-Agents to Master Agentic Workflows with Powerful LLMs – A Game-Changing RL Approach
Discover W4S, a clever reinforcement learning method that lets a lightweight meta-agent craft complex agentic workflows, leveraging beefy LLMs for execution without breaking the bank.
Kong's Volcano: Revolutionizing AI Agent Development with TypeScript MCP SDK for LLM Reasoning and Real-World Actions
Discover how Kong's new Volcano SDK empowers developers to build robust, production-ready AI agents in TypeScript, seamlessly blending LLM reasoning with practical actions via MCP integration.
Rogue by Qualifire AI: The Ultimate End-to-End Framework for Testing and Benchmarking AI Agents
Discover Rogue, Qualifire AI's groundbreaking open-source framework that revolutionizes AI agent evaluation with comprehensive end-to-end testing across planning, tools, and reasoning. Perfect for developers building reliable agentic systems!
Building Your Own AI-Driven Crypto Agent: Hybrid Encryption, Digital Signatures, and Smart Security Myths Busted
Discover how to create a powerful AI agent that handles hybrid encryption, digital signatures, and adaptive security using LangChain and Ollama. Bust common myths and get hands-on code to secure your data like never before.