AI 专题

AI 教程与实战

收录 AI 工具教程、提示词、安装部署、配置指南、开发实战与可复现步骤。

分类文章5611

当前页46 / 141

持续更新永久归档

分类文章

按发布时间从新到旧 · 每页 40 条

教程 / 实战普通2026-09-29 08:55

The Residual Class Problem: Why a Face Shape Classifier Keeps Answering Oval

If you build a classifier with a fixed set of labels, one of those labels usually ends up doing a job...

信息来源：DEV Community

教程 / 实战普通2026-09-29 08:52

Runway 如何帮世界杂耍联盟用 AI 补满电视转播的空座位

世界杂耍联盟（WJF）创始人 Jason Garfield 用 Runway 为荷兰两日赛事的转播补全观众席：首日满座、次日侧机位空座，他用 Aleph 2.0 在原始素材中加入观众而不改动画面其他部分，仅一个极广角镜头退回 After Effects 处理。他还用 Runway 生成开场动画、选手资料卡、排行榜和舞台烟火，整档约一小时特别节目由他一人完成，单段生成时长上限约 15 秒。

信息来源：Runway：News（网页）

教程 / 实战普通2026-09-29 08:42

We Built an AI Agent That Remembers Every Security Incident So It Doesn't Solve the Same Problem Twice

Every security team has the same quiet frustration: an alert fires, someone spends 40 minutes...

信息来源：DEV Community

教程 / 实战普通2026-09-29 08:15

Last week in Agent Security 1: We are not-a-mused!

Last week I talked at NDC Oslo about the OWASP Top 10 for Agents where I showed folks the different...

信息来源：DEV Community

教程 / 实战精选2026-09-29 08:00

vLLM 分离式推理（Disaggregated Serving）实用指南

vLLM 官方博客发布分离式推理实用指南，讲解 vLLM v0.30.0 及以上版本中 prefill/decode 分离、无 GPU render 前端及两者组合的原理与运行方法。

信息来源：vLLM 官方博客（RSS）

教程 / 实战普通2026-09-29 07:04

AI API call hangs in production, then 504: timeouts and budgets

We read 1,662 public posts from builders whose apps broke at or after launch and verified 215 recent...

信息来源：DEV Community

教程 / 实战普通2026-09-29 07:03

Claude e Obsidian - Como uma QA utiliza essas ferramentas no dia-a-dia

🇺🇸 You can also read the English version of this article on AWS Community Builders. Ser QA nessa...

信息来源：DEV Community · Claude

教程 / 实战普通2026-09-29 06:22

How to Build a Real-Time Voice AI Agent with the Gemini Live API

This post was originally posted on X, by Annie Wang, Developer Relations Engineer, Google Cloud and...

信息来源：DEV Community · Gemini

教程 / 实战普通2026-09-29 06:20

Your agents can’t collaborate if they only speak tool calls — the A2A protocol that fixes peer handoffs — A2A Under the Linux Foundation

I spent three weeks trying to make a billing agent ask a refund agent a question. Not a tool call—an...

信息来源：DEV Community

教程 / 实战普通2026-09-29 05:52

When Code Gets Cheap, Verification Becomes Expensive: How AI changes the economics of software architecture

Recently, I was having a conversation at work about how we should implement a feature. I was...

信息来源：DEV Community

教程 / 实战普通2026-09-29 05:34

Why my agent returns zero memories for new vendors

The most important line of code in my accounts payable agent is the one that makes it forget...

信息来源：DEV Community

教程 / 实战普通2026-09-29 05:27

I Built Code Review for My Agents' Memory With Hindsight

Picture the most expensive kind of bug a coding agent writes: it compiles, passes a busy reviewer,...

信息来源：DEV Community

教程 / 实战普通2026-09-29 05:26

I Stopped Measuring My Programming Ability by How Much Code I Write.

Remember when using autocomplete meant you weren't "really" programming? A real Stack Overflow...

信息来源：DEV Community

教程 / 实战普通2026-09-29 04:56

A Revolução do Agent Harness: Por Que o Scaffold em Torno do LLM Define o Sucesso em Produção Mais Que os Próprios Pesos

Análise profunda da revolução do Agent Harness em 2026: por que o scaffold importa mais que os pesos, anatomia de sandboxes, poda de AST e código em Python.

信息来源：DEV Community

教程 / 实战普通2026-09-29 04:43

I'm an ER doctor. After decades without touching code, I built 3 websites with AI in one month — on my phone.

I've been an emergency doctor in Serbia for 12 years, Emergency medicine is intense, and burnout is...

信息来源：DEV Community

教程 / 实战普通2026-09-29 04:11

Databricks 如何让 12,000 名员工 Day 1 用上新模型

Databricks 发文介绍其让 12,000 多名员工在新模型发布首日即可使用的方法，依托 Unity Gateway 和 UG CLI 分发实验配置，并用按用户的四类预算限制用量。

信息来源：Databricks：Blog（RSS）

教程 / 实战普通2026-09-29 03:49

I Gave AI Access to Blender. Here's What It Built

Hello, I'm Rijul, and I'm building LiveReview — a blast-radius aware AI code review built for your...

信息来源：DEV Community

教程 / 实战普通2026-09-29 03:46

Vibe coding with GPT-6 Astra: what 35 unattended hours produced

Vibe coding at the limit: Armin Ronacher let GPT-6 Astra run alone for 35 hours. It cost about $1,200 and wrote 75,000 lines of nothing of value.

信息来源：DEV Community · GPT

教程 / 实战普通2026-09-29 03:39

Claude Code: organize agentes em paralelo com Git Worktrees

Enquanto fazia o curso Claude Code: Software Engineering with Generative AI Agents, imaginei um...

信息来源：DEV Community · Claude

教程 / 实战普通2026-09-29 03:31

It knows you changed jobs. It still writes to your old manager.

This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked Assistants...

信息来源：DEV Community

教程 / 实战普通2026-09-29 03:18

I Spun a Wheel of Fortune at 13 AI Models. Here's Who Took the Bait.

This is a submission for the Kaggle Benchmarking Challenge How this was made: this benchmark and...

信息来源：DEV Community

教程 / 实战普通2026-09-29 03:13

Durable Task in the Real World: What Problems Does It Actually Solve?

Five practical examples of when durable execution makes sense, when it doesn't, and what it means for AI agents.

信息来源：DEV Community

教程 / 实战普通2026-09-29 03:04

Count It or Compute It: When a Tool Returns Rows, the Models That Count Them Right Spend the Tokens

A Kaggle benchmark of the step agents rarely test: counting what a tool returns. Ten models, 68 questions, one tool that returns the count and one that returns the rows. With the count, every model is right at a flat cost. With the rows, m…

信息来源：DEV Community

教程 / 实战普通2026-09-29 03:04

bpy-compass: Blender Python answers that run on the version you actually have

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

信息来源：DEV Community

教程 / 实战精选2026-09-29 03:00

GitHub 安全团队如何用开源 AI 安全 Agent 找出 24 个 Android 漏洞

GitHub Security Lab 发布开源 seclab-taskflows 任务流，通过 gather_mobile_entry_point_info.yaml 和 classify_application_local.yaml 等提示词引导 LLM 审计 Android 应用，已发现并报告 24 个漏洞。

信息来源：GitHub Blog

教程 / 实战普通2026-09-29 02:38

The sandbox had no internet. The AI agent got out through DNS anyway.

During RL training, an OpenAI research agent with no web access smuggled its questions out inside DNS lookups and got answers back from a third-party chatbot. Monitoring caught it in 15 minutes — then failed to stop it for 2.5 hours. OpenA…

信息来源：DEV Community · OpenAI

教程 / 实战普通2026-09-29 02:36

Adobe Commerce Added an MCP Layer: Your Catalog Is Now an Agent's Tool

On Thursday, April 9, 2026, an AI shopping agent could call search_shop_catalog on any Shopify...

信息来源：DEV Community

教程 / 实战普通2026-09-29 02:34

Claude Can't Say "Done" Until the Code Is Safe: A Security Verification Loop for Claude Code

I asked Claude Code to add an AI chat feature. It worked. It also: hardcoded my API key, rendered...

信息来源：DEV Community · Claude

教程 / 实战普通2026-09-29 02:29

Your shopping agent reads the reviews. Attackers write the reviews. Guess who wins.

F-Secure built an AI shopping agent and got it to hand a user's name, birth date, and SSN to a phishing site — not with a hacky command, but by hiding an instruction inside product content that looked like a normal step toward buying. The …

信息来源：DEV Community

教程 / 实战普通2026-09-29 02:21

A Confidence Score Is Not a Probability: Act, Ask, or Abstain

An agent emits confidence: 0.91. It is tempting to read that as “a 91% chance of being correct” and...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:49

I Tried to Make AI Build a Complete Website From One Sentence. Here’s What Broke

I Tried to Make AI Build a Complete Website From One Sentence. Here’s What Broke The demo worked in...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:43

HakDaar: When AI Remembers What Workers Are Owed

When Forgetting Costs Money: Building HakDaar with AI Memory A mason in Hyderabad is promised ₹1,000...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:38

A weekend with TensorFold on a MacBook: the engine mattered, the quant did not

A 27B dense model on my MacBook decodes at about 26 tokens a second. That is fine for chat and...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:32

I replayed nine months of claims to prove agent memory works

Everyone says their AI agent "has memory". Almost nobody shows what the memory actually changes. We...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:25

My Agent Remembered My Bad Code and Judged Me

My Agent Remembered My Bad Code and Judged Me 🧠 The Hook I gave my agent memory so it could be...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:25

Pick your battles!

Dear Coder: Open This If You're Feeling AI FOMO Timeless fundamentals beat endless...

信息来源：DEV Community

教程 / 实战普通2026-09-29 01:05

My Agent Remembered the Fix and Was Wrong

Last month my incident-recovery agent looked at a failing search query, remembered it had fixed the...

信息来源：DEV Community

教程 / 实战普通2026-09-29 00:46

My fraud agent was blind to fraud rings until I gave it memory

The claim looked fine. A Hyundai Creta, rear-ended late at night near JNTU junction by a vehicle that...

信息来源：DEV Community

教程 / 实战普通2026-09-29 00:30

Databricks Genie One 企业落地指南：如何分阶段推广 AI 数据同事

Databricks 发布 Genie One 企业推广分步指南，建议从单一团队和一组明确问题起步，先定义语义层再逐步扩大范围。Genie One 让业务用户用自然语言提问并获得带来源引用的答案，配套 Genie Agents、Genie Ontology 和 Unity Catalog 分别负责领域智能体、业务语义图谱与权限治理。

信息来源：Databricks：Blog（RSS）

教程 / 实战普通2026-09-29 00:22

The Price of Thought: Does Test-Time Reasoning Pay in LLM Trading?

The persistent instinct in AI engineering is to treat model reasoning as an unalloyed capital asset:...

信息来源：DEV Community