AI 教程与实战

The last thing I wrote here was about two bugs my own test suite could not see. Both came from the...

02:06 UTC checkout-api started timing out. I pasted the alert into a capable LLM and asked what to...

How we spent an hour using Gemini and the Web Audio/MIDI APIs to rapidly build, test, and record a responsive wind controller synth using a real instrument.

Introduction Here is how to build a contract obligations review app with ToolJet MCP for...

An incident-response system can retrieve the right information and still fail to learn...

I've been curious about this for a while. Every AI image generator I've tried writes metadata into...

Darktrace showed that coding agents trust their local conversation history without verifying it. A writable file is enough to give the model orders. Here is what to watch.

Over the past few days, I've been comparing LongCat 2.5 Preview and Space Bunny Alpha across actual...

Fifteen AI review bots on one repository is not fifteen times the insight — it is a governance problem. How one open-source project keeps machine reviewers advisory, deterministic gates blocking, and PR code out of the trust boundary.

How non-engineers build AI tools on an internal MCP platform and self-merge them to production — the self-merge gate, clean-context AI code review, and the one zone that still needs an engineer.

Hello Devs 👋 One thing I have noticed with AI coding agents is that they are very good at working...

AI is becoming part of almost every layer of an enterprise. Employees are using AI assistants....

AI assistants are very good at writing Python that looks right. It's typed, it's formatted, it comes...

This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real...

Here's the idea: build an app that checks if someone knows their stuff using an interview. You pick a...

Research status: Ongoing Current milestone: v0.39 Type: Experimental research prototype Repository:...

There's a new kind of technical debt, and it doesn't come from cutting corners. It comes from...

Anthropic 官方文档介绍针对 Claude Opus 5.5 的提示词模式，覆盖 effort 校准、无人值守智能体运行、安全拒绝、进度更新、多应用工作流、视觉输入和前端设计等场景。

If you are building a RAG or agent pipeline over video, the transcript is the payload. Titles and...

Introduction Every on-call engineer has had the same feeling: an alert fires, the symptoms look...

In August 2026, less than a month after Moonshot AI's Kimi-K3 surprised the world, Alibaba Group...

If you let an AI assistant read your process list over MCP and ask it "is my PC safe?", everything an...

Prepared for the Kaggle Benchmarking Challenge. What I Benchmarked I build agents for...

A workflow pattern is spreading through Chinese AI developer communities right now: use the more...

GramSetu AI: Architecting Context-Aware Multilingual Civic Tech with Episodic Agent...

How AI is changing project management, from traditional task tracking to workflows where AI agents...

Scientists have mapped the whole nervous system of a fruit fly: every neuron and every connection, in...

I'm not going to lie, deep down I still love to actually build things (UI and code), but the simple...

Treat LLM inputs like untrusted user data. Learn to detect and mitigate prompt injection attacks in AI agents using layered defense, input sanitization, and output validation.

Tool Descriptions Are the Contract made a claim I had not tested: your MCP server's recovery messages...

AI coding agents are getting much better. They can inspect repositories, change multiple files, run...

Every OCR demo handles a clean PDF beautifully. Feed it a crisp, flat, well-lit digital document and...

Most developers know about OpenAI, Anthropic, Groq, and Cerebras. What many developers don't realize...

This post is part of my notes from Formação AWS, a course by Henrylle Maia. It covers a class from...

This post is part of my notes from Formação AWS, a course by Henrylle Maia. It covers a class from...

A bug appears. You paste the error into your AI coding agent. Thirty seconds later: “Fixed...

It's all over the headlines. Every so often, someone reveals a model X% faster. A tech CEO makes...

AI Harness Engineering · Essay Eleven · Derek Wang Karl Popper spent the middle of the twentieth...

Search engines have already crawled the visual web. Crawling it all again would duplicate that work,...

Why I stopped running one do-everything AI agent: per-job workspaces keep agent memory true, shrink starting context, and coordinate through events.