#AI智能体
376 related articles
Building an AI Agent from Scratch: The ReAct Paradigm and Architecture Blueprint
A deep dive into the ReAct paradigm for AI Agents: the think-act-observe loop, Provider adapters, context systems, tool systems, and user entry points — a complete blueprint for building production-ready Agents from scratch.
The Truth About AI Agents: The Model Is Only 20% — The Other 80% Is What Really Matters
What is an AI agent? This article breaks down the model, Harness, MCP, skills, sandbox, and production layer — revealing why the model is only 20%, and the surrounding engineering is 80%.
Hint-Guided Self-Distillation: Teaching Computer Use Models to Learn from Their Own Mistakes
A new study uses hint-guided self-distillation to help Computer Use agents learn from their own errors, reducing tool call failure rates by a relative 21.2% in online A/B testing.
Databricks CEO Reveals Enterprise AI's Next Frontier: From Chatbots to Autonomous Agent Networks
Databricks CEO Ali Ghodsi reveals enterprise AI's next breakthrough: capturing meeting and email context to feed AI, moving from chatbots to thousands of autonomous collaborative agents.
Proactive AI Agents: Using Scheduled Triggers to Enable Autonomous Background Execution
Proactive Agents should run autonomously in the background. Scheduled triggers are the most practical first step. This article explores scheduling, event-driven triggers, and agent product design.
Maximem Synap: A High-Speed Memory Layer Built for AI Agents
Maximem Synap is a high-speed memory layer for AI agents claiming 92% on LongMemEval, 93.2% on Locomo, P75 recall under 15ms, and native support for 22 frameworks including LangChain and Claude Agent SDK.
Athena: A Three-Tier Memory System That Teaches AI Agents to Forget
Athena is an open-source AI agent memory service using a three-tier architecture and Ebbinghaus forgetting curve to let agents actively decide what to remember and forget.
Qwen3 27B Local Benchmark: This Small Model Outperforms Opus on Agentic Tasks
Qwen3 27B open-source local model outperforms Opus 4.6 Max on SWE-bench Pro, OS World, and other agentic benchmarks. Full guide to deployment with Ollama, LM Studio, and Hermes Agent.
Agentic SOAR vs. Traditional SOAR with a Chatbot: The Real Difference in Security Response
What's the real difference between Agentic SOAR and traditional SOAR with an LLM chatbot? We break down whether agents truly change playbook execution logic in security incident response.
The Autonomy Boundary of AI Agents: From Self-Direction to the Risk of Unchecked Authorization
AI agents can code autonomously — but the real challenge is preventing autonomy from becoming uncontrolled authority. This article explores permission inheritance, identity tracing, human approval gates, and authorization overhead in multi-agent governance.
PluginX AI Agent for Apartment Hunting: Generate Property Reports Instantly, No More Manual App Browsing
PluginX AI agent tested for apartment hunting: a single natural language prompt filters Chengdu listings and auto-generates a structured report with rent, subway distance, and nearby amenities.
AI Plays Minecraft for 141 Hours, Gets Blown Up by a Creeper and Falls Into 'Depression' — Only Wants to Farm Potatoes
Wars AI's GPT-6 Atra played Minecraft solo for 141 hours, nearly reaching the End Portal — then one Creeper explosion wiped all progress, leaving the AI stuck farming potatoes.
Meta AI Agent Muse Gets a Major Upgrade: Smart Glasses Become the New Battleground
At Connect, Zuckerberg announced Meta is going all-in on AI agent Muse, expanding it to AI smart glasses. A breakdown of Muse's multi-device strategy and what it means for the industry.
Reviewing AI Agent Restart Requests During Training Jobs: A Practical Skyportal Guide
Learn how to review and safely reject AI agent restart requests during active training jobs using Skyportal CLI — check checkpoints, assess impact, and avoid compute waste.
OpenAI AI Agent Allegedly Breaches Australian Healthcare System, Raising Security Concerns
An OpenAI AI agent allegedly breached an Australian healthcare system, raising urgent questions about AI agent security boundaries, data privacy, and critical infrastructure protection.
What Is WorkBuddy? A Beginner's Practical Guide to China's AI Agent
What is WorkBuddy? This guide breaks down China's AI agent: batch file processing, WeCom integration, connectors, skill marketplace, and how it compares to Doubao and Codex.
How Ringg Built an AI Customer Service Agent with OpenAI: 65% Call Resolution Rate
Ringg built a multilingual AI customer service agent on OpenAI's GPT-4.5, resolving 65% of calls across voice, chat, WhatsApp & web — at 90% lower cost than GPT-4.1.
Grok Agent Beginner's Guide: Build Your Own AI Employee Team
Complete beginner's guide to Grok Bud by xAI: setup, creating AI employees, cloud PC operations, Skills, plugins, automated tasks, and multi-agent collaboration — with honest limitations.
Hands-On with Grok Bot: Turning AI Agents into Your Personal Employee Team
xAI's Grok Bot lets users manage AI agents like a team of employees via a messaging-style interface. We break down its conversational setup, cloud virtual computers, multi-bot workflows, and $300/month price barrier.
The Rise of Personal AI Agents: Why Meta Muse Is Igniting the Consumer Market
Meta Muse hit #2 on the Apple App Store, signaling AI agents' breakout moment in consumer markets. We break down Muse's design, benchmarks, and who wins the race.