228 related articles

The rise of Zhipu's GLM 5.2 is accelerating the democratization of LLM capabilities. This article analyzes the commoditization of foundation models, the logic behind margin collapse, and the opportunities and challenges facing application-layer and foundation model firms.

A controversial study shows training just one Transformer layer can match full-parameter RL training. We analyze the technical principles, engineering value, and limitations of this approach.

A beginner's guide to the LangChain open-source framework: explaining how to use the init_chat_model unified interface, tips for disabling DeepSeek's thinking mode, and core essentials of Agent development.

DeepSeek R1 lacks Function Calling and JSON Output by default. Qwen3's programmable thinking modes make it the top open-source agent choice. Key LLM selection pitfalls and MCP protocol updates.

A deep dive into Claude Code: the difference between Terminal and Device Agents, enterprise selection advice, and how to use Claude Code with DeepSeek in China.

Step-by-step guide to connecting DeepSeek and other domestic AI models to Claude Code Desktop — covers no-account setup, CC Switch config, Chinese localization, and custom Skills.

Why do banks and hospitals build Local AI instead of using cloud services? This guide covers the full tech stack — Ollama, RAG, vector databases — and real-world enterprise deployment use cases.

Learn AI Agent development from scratch. This tutorial covers LLMs and prompts, then builds a conversational agent in Python using the DeepSeek API with multi-turn dialogue and system prompts.

SpaceX acquires Cursor for $60B in all-stock deal, buying the AI-era developer gateway. A deep analysis of the U.S.-China deep tech ecosystem gap and China's path to building its own flywheel.

How to build a local AI inference server with 4 used RTX 3090 SXM4 GPUs to run GLM-5.2 via Llama.cpp and Unsloth IQ quantization, with real benchmarks on speed and quality.

Claude Sonnet 5 may launch this week with up to 2M token context; GPT-4.6 Pro arrives with stunning code generation; mysterious Opus 6 exists internally. Full breakdown of this week's frontier AI model updates.

Deep analysis of VPN-free mirror sites for GPT-5.5 and Claude in China, revealing technical principles, data security risks, compliance concerns, and safer alternatives.

In-depth analysis of AI aggregation platforms: the truth behind free access to GPT, Gemini, Claude and other LLMs, hidden costs, privacy risks, and safer alternatives.

A detailed review of domestic Chinese platforms offering no-registration, no-VPN access to GPT-4, Gemini, Claude, DeepSeek and other top AI models, with security risk analysis.

Learn how to use CodeBuddy (WorkerBuddy) to auto-install Claude Code with a single command. No NPM setup or API Key hassle — pair with DeepSeek for low-cost AI coding.

Complete guide to Claude Code Desktop setup: account-free usage, DeepSeek integration via CC Switch, Chinese localization, and custom Skill import for low-cost AI coding.

Google Android Bench shows frontier open-source models solve 50-60% of Android dev tasks. Mid-size models like Gemma 4 run locally with just 20GB RAM.
On-Policy Distillation Explained: Prin…
A deep dive into On-Policy Distillation: core principles, key differences from Off-Policy methods, and applications in model compression, reasoning transfer, RLHF alignment, and self-improvement.

Deep dive into Moonshot AI's Kimi K2.7 Code: MoE architecture details, benchmark analysis, API pricing vs Claude/GPT, 6x speed version, and practical guidance for developers evaluating adoption.

Exposing how domestic AI sharing sites use fake GPT and Claude version numbers to harvest traffic. Analysis of data security risks, service instability, and safe AI usage tips.