---
title: "AI | Blog | Phát Lâm"
url: https://phatdev.net/en/blog/c/ai
lang: en
updated: 2026-10-21
alternate:
  vi: https://phatdev.net/vi/blog/c/ai
markdown: https://phatdev.net/en/blog/c/ai.md
---

# AI

Technical notes drawn from real projects: frontend, backend, AI, security, SEO, media and mobile apps.

## Posts

- [What Is MCP? How AI Connects to Your Tools and Data, and the Risks to Know](https://phatdev.net/en/blog/mcp-explained.md) (AI | 2026-10-21 | 7 min read): An AI can write a great email but cannot see your calendar. MCP is the shared socket that connects AI to tools and data. Plain examples, permission and prompt injection risks, the 2026 state of the spec, and a working example.
- [Claude, Codex and Gemini: Which AI Coding Assistant for Which Job](https://phatdev.net/en/blog/claude-codex-gemini-for-coding.md) (AI | 2026-10-21 | 6 min read): Three AI coding assistants differ in how you use them, how you pay and what happens to your data. A plain-words comparison from official docs checked on 21 October 2026, including Gemini CLI moving to Antigravity CLI.
- [Common Problems When Using AI to Write Code, and How to Avoid Them](https://phatdev.net/en/blog/ai-coding-common-problems.md) (AI | 2026-10-21 | 6 min read): Six familiar traps: invented functions, scope creep, hitting a usage limit midway, forgetting earlier instructions, fake passing tests and leaked secrets. Each comes with signs, causes and prevention.
- [Letting an AI Agent Run Commands on Your Machine or Server: How to Stay Safe](https://phatdev.net/en/blog/ai-agent-safety-on-your-machine.md) (AI | 2026-10-21 | 6 min read): Seven steps for letting an AI agent run commands without risking your data: least privilege, sandbox, ask-first modes, no passwords, a separate git branch, backups and reading each command before approving.
- [AI Chatbots for Your Shop: What They Can and Cannot Do, and How Running Costs Work](https://phatdev.net/en/blog/ai-chatbot-for-shop.md) (AI | 2026-10-21 | 6 min read): A chatbot answering customers at midnight sounds great, but it only knows what you give it to read. This post covers what to hand over, what data to prepare, how costs work and how to limit wrong answers.
- [How to Ask an AI So It Gets It Right and Stops Making Things Up](https://phatdev.net/en/blog/prompt-engineering-for-developers.md) (AI | 2026-10-19 | 6 min read): The same AI gives very different results depending on how you ask. A five-part way to write the request, three layers against invention, and a working code check.
- [Using the Claude API in Your Product: Tool Calls, Live Text and Lower Bills](https://phatdev.net/en/blog/claude-api-tool-use-streaming-caching.md) (AI | 2026-10-19 | 7 min read): Let the AI call your functions, show text as it arrives instead of a blank screen, and stop paying repeatedly for the same text. Correct models for October 2026, with Python and TypeScript.
- [Handing a Big Job to Several AIs Without Breaking Things](https://phatdev.net/en/blog/agentic-workflow-orchestrator.md) (AI | 2026-10-19 | 6 min read): One AI working on a big job loses track halfway. Split it across several small AIs with one coordinator, check results with your own measurements, and always keep a way back.

RSS: https://phatdev.net/en/blog/rss.xml