Not available in English yet
AI Agent 怎么记住你 — 中国工程师把记忆拆成 4 层,落地成本藏在这
Related Reading
More on #MCP
Enterprise AI Q&A Reality: The LLM Is the Engine, PDF Cleaning Is the Chassis
This week's enterprise knowledge base Agent tutorial caught our eye: the real bottleneck isn't the LLM — it's turning thousands of PDFs into a citable
Roo Code Pushes a Local 8B From 2s to 10s — The Wrapper Layer Is the Bottleneck
A developer benchmarked the same Llama 3.1 8B: 2s in CLI, 10s in VSCode's Roo Code plugin. The 5-10x gap is 100% in the wrapper layer, not the model.
Cloudflare rewrites the CLI for AI agents: 48% of calls are no longer human
Cloudflare launches cf CLI designed for AI agents, as 48% of Wrangler calls already come from agents — the "readers" of dev tools are changing.
OpenRig: virtual teams of Claude and Codex—but is grouping really better?
OpenRig 0.5.9 bundles Claude Code, Codex into persistent virtual dev teams—signaling AI's shift to swarm intelligence, with four open issues.
Tag the Bot in IM, It Picks Up the Job — A Chinese Open-Source Agent Experiment
Chinese open-source DSH ships dsh-waker: @ a bot in IM to dispatch and follow up with AI digital workers. Early Agent-via-chat experiment worth a tech
AI Agents Keep Failing in Enterprise — Not Model IQ, But the Company Itself
Why AI Agents keep failing in enterprise: not the model, it's context. We unpack Context Engineering and what it means for managing company knowledge.