#token probability
833 related articles
·7 min
Claude Sonnet 5 Released: The Most Capable Agent Model at Mid-Tier Pricing
Anthropic releases Claude Sonnet 5, its most agentic Sonnet model yet. With planning, browser/terminal tool use, and autonomous execution, it brings flagship Agent capabilities to mid-tier pricing.
Read more →
·7 min
Claude Sonnet 5 Released: The Most Powerful Agent-Capable Mid-Tier AI Model
Anthropic releases Claude Sonnet 5, its most agentic mid-tier model with planning, browser/terminal tool use, and autonomous execution—bringing flagship Agent capabilities at significantly lower cost.
Read more →
·5 min
Building RAG from Scratch: Creating a Domain Expert System with Local Small Models
Step-by-step guide to building a complete RAG pipeline with Ollama + LangChain + FAISS + Qwen 1.5B. Run document retrieval and intelligent Q&A locally without a GPU.
Read more →
·6 min
Gas Town: A Multi-Agent Workspace Manager Built in Go
Gas Town is an open-source multi-agent workspace manager built in Go with 16,000+ GitHub Stars. This article analyzes its architecture, Go language advantages, and typical multi-agent collaboration scenarios.
Read more →
·6 min
Codex Beginner's Guide: Four Product Forms Explained, Plus Comparison with Claude Code and Cursor
In-depth analysis of OpenAI Codex's four usage forms, comparing Codex, Claude Code, and Cursor across price, stability, and frontend/backend fit to help developers choose the right AI programming tool.
Read more →
·3 min
Taste-Skill: The Open-Source AI Taste Engine That Fights Generic AI-Generated Content
Taste-Skill is a viral open-source JavaScript project with 56K+ GitHub Stars. It uses prompt engineering and anti-slop rules to help AI generate higher-quality, more distinctive content.
Read more →
·3 min
Claude Code in Practice: Writing Maintainable AI Code with the Harness Methodology
Learn how to write controllable, maintainable AI code using the Harness methodology with Claude Code. Covers SDD, Agent orchestration, and enterprise-grade AI programming practices.
Read more →
The Anti-Slop Fiction Contest: How to …
·2 min
The Anti-Slop Fiction Contest: How to Make AI Write Like It Isn't AI
A fiction contest challenging AI writing 'slop' sparks industry reflection. Learn what AI slop is, why it happens, and how prompt engineering can help AI write with genuine soul.
Read more →
·6 min
What Is Vibe Coding? A Complete Breakdown of the LLM + Agent + Workflow Three-Layer Architecture
Vibe Coding, coined by ex-Tesla AI Director Karpathy, redefines AI programming. This article breaks down the LLM + Agent + Workflow three-layer architecture.
Read more →
·4 min
Adding a 'Repair Layer' to Open-Source Models: How DeepSeek Can Outperform Claude Opus
Independent developer Ahmad Awais found that open-source LLM failures stem from Tool Calling bugs, not model capability. A deterministic repair layer + repair hints can make DeepSeek outperform Claude Opus.
Read more →
·3 min
Light-Skills: Open-Source Research AI with 28 Skill Modules Covering the Full Workflow
Light-Skills is an MIT-licensed open-source research AI with 28 interconnected Skills, 9 knowledge bases, 317 knowledge cards, and 49 scripts covering the full research workflow — with a hard rule against fabricating citations or data.
Read more →
·4 min
MCP Practical Tutorial: From Server Development to Multi-Client Integration
A complete MCP practical guide using the official Python SDK — covering environment setup, FastMCP server development, Inspector debugging, and multi-client integration with Cursor and Cline.
Read more →
·3 min
Needle: Fine-Tuning a 26M Ultralight Tool-Calling Model Locally — Full Walkthrough
Needle is a 26M-parameter tool-calling model. Learn how to replace Gemini with Ollama for local training data generation and fine-tune on a single GPU, achieving 96.7% F1 — ideal for edge AI deployment.
Read more →
·3 min
Fixing the Three Big Weaknesses of LLMs: Knowledge, Computation, and Tool Use
LLMs struggle with knowledge accuracy, math, and external capabilities. This article explains three prompt engineering solutions — generated knowledge prompting, PAL/PoT, and tool use — to systematically strengthen LLM reliability and build better AI Agents.
Read more →
·4 min
ChatGPT Plus Subscription Guide: Third-Party Top-Up Explained, Risks Analyzed, and Compliant Alternatives
Struggling to subscribe to ChatGPT Plus in China? This guide explains how third-party top-up services work, the real security and ban risks involved, and safer alternatives including official subscriptions, API access, and Chinese LLMs.
Read more →
·4 min
Tencent Marvis Hands-On: How Multi-Agent AI Takes Over Your Computer
Deep dive into Tencent's Marvis AI agent: multi-agent architecture, intelligent file organization, document analysis, cross-device sync, and local privacy mode explained.
Read more →
·8 min
The Codex Orange Book: An Open-Source Full-Stack Practical Guide for AI Programming Beginners
The Codex Orange Book is an open-source unofficial guide with 2.5K+ GitHub Stars, covering installation, plugins, MCP extensions, and hands-on projects to help beginners master AI programming tools.
Read more →
·6 min
Getting Started with LangChain: A Deep Dive into Model and Agent Core Concepts
Deep dive into LangChain's core Model and Agent concepts, covering unified model interfaces, agent tool calling, middleware mechanisms, and key principles for building LLM applications.
Read more →
·7 min
AI Agent Memory Layer Comparison: Mem0, Zep, Letta, and Cloudflare — Four Major Solutions Evaluated
In-depth comparison of four AI agent memory layer solutions: Mem0's extract-retrieve approach, Zep's temporal knowledge graphs, Letta's self-editing memory, and Cloudflare Durable Objects as infrastructure primitives.
Read more →
·7 min
Vibecoding in Practice: How to Guide AI to Precisely Borrow from Open-Source Projects Without Going Off Track
In AI-assisted coding, having AI borrow from open-source projects often leads to wholesale copying. This article uses the GapGun audio tool as a case study to show how to set borrowing boundaries and guide AI precisely.
Read more →