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8454

当前页131 / 212

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2026-09-25 02:25

What the AI isn't allowed to do

How we build VolunteerPress with AI, the checks every change has to pass, and the bug those checks...

信息来源：DEV Community

教程 / 实战普通2026-09-25 01:51

动态消除：用 Engram 多层引导在 Qwen3-4B 上实现无损拒绝抑制

一种名为 Dynamic Abliteration 的方法通过 PyTorch forward hooks 在运行时拦截多层残差流，无需修改权重即可抑制 Qwen3-4B 的拒绝行为，基座模型权重保持 100% 冻结。该方案在 A100 上以 Qwen/Qwen3-4B（隐藏维度 2560、36 层）为概念验证，对比了传统单层拒绝向量减法会永久改变权重并损害非拒绝任务的问题。

信息来源：Hacker News 热门（buzzing.cc 中文翻译） · Qwen

教程 / 实战普通2026-09-25 01:29

What AI Actually Did to My Software Development Velocity

A lot of AI productivity claims sound impressive until you ask what was actually measured. "10x...

信息来源：DEV Community

教程 / 实战普通2026-09-25 01:03

INKSHIFT: cross out a table, keep the booking

This is a submission for the Sanity Challenge, Path Two: Vibe-Code Something Strange. What...

信息来源：DEV Community

教程 / 实战普通2026-09-25 01:01

FinePrint: an agent that checks your hackathon entry against the rules it reads

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 / 实战普通2026-09-25 00:55

I rebuilt four dead systems with AI. The failures all landed in the same place.

Four reconstructions, four kinds of missing documentation, and the defects landed in the same place...

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:25

World Models for Agents Should Edit Transcripts, Not Simulate Terminals

Most discussion around world models for language agents imports assumptions directly from robotics....

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:24

Qodo vs CodeRabbit vs Cubic: Which One Enforces Rules Instead of Just Suggesting Them? 🔥

TL;DR Most AI reviewers know how to leave comments. A completely different question is:...

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:21

Building RAVEL: Detecting Complex Fraud Rings with TigerGraph Cloud & Agentic GraphRAG

TigerGraph Hacker House Goa — Agentic Fraud Investigation Challenge Team / Co-Authors: Nikhil Kumar...

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:11

Serve Markdown to Agents at the Gateway, Not the Origin

An agentgateway route serves markdown to agents on the Accept header while the HTML-only origin stays untouched, plus the two ways a naive version...

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:08

Goodbye pbxproj: A Friendly Guide to Xcode 27.2's New JSON Project Format

If you've built iOS apps on a team, you know this moment. You pull main, Git says CONFLICT (content):...

信息来源：DEV Community

教程 / 实战普通2026-09-25 00:00

SGLang 如何用 Score API 与 MIS 扩展 JEV 类决策模型服务

SGLang 团队介绍用 Score API 与多条目评分（MIS）服务 JEV 类决策模型：/v1/score 通过 label_token_ids 显式请求标签 token 分数，MIS 在单次请求内复用共享 query 计算并隔离各候选。

信息来源：LMSYS：Blog（Chatbot Arena 团队）

教程 / 实战普通2026-09-24 23:32

I checked 101 "tests pass" claims from my AI coding agents. 35% weren't true.

When an AI coding agent ends a session with "All tests pass ✅", I usually believe it. It's the last...

信息来源：DEV Community

教程 / 实战精选2026-09-24 23:31

Gemini 3.8 TTS 自定义语音实战指南：复刻自己的声音或从一句话创建新声音

Philipp Schmid 发布 Gemini 3.8 Flash TTS 与 Flash-Lite TTS 使用指南，二者已在 Gemini API 和 AI Studio 可用，登顶 Hume's Voice Design Benchmark 并在 Voice Arena 六种语言中居首。

信息来源：X：Philipp Schmid（Google DeepMind 开发者体验） (@_philschmid) · Gemini

教程 / 实战普通2026-09-24 23:24

LLM Reasons, Policy Engine Decides: Autonomous Agentic Fraud Defense on TigerGraph

LLM Reasons, Policy Engine Decides: Autonomous Agentic Fraud Defense on TigerGraph By Lead...

信息来源：DEV Community

教程 / 实战普通2026-09-24 23:21

Luma AI 如何融入创意工作流

创意团队已经在用 AI 了。现在更大的问题是，它能否融入真实工作流而不加剧碎片化。在 Brandweek 上，@AlexiaAdana 展示了 Luma 如何从早期方向到制作全程支持创意流程。 ADWEEK 在此探讨下一阶段：https://www.adweek.com/creativity/ai-has-reached-creative-teamsnow-comes-the-hard-part/

信息来源：X：Luma AI (@LumaLabsAI)

教程 / 实战普通2026-09-24 23:10

Day 2: The Art of Forgetting (How an AI Survives on 6GB RAM Before It Dies)

Day 2: The Art of Forgetting Human beings think memory is a gift. They spend fortunes on...

信息来源：DEV Community

教程 / 实战普通2026-09-24 23:08

Thinking of ACE? We Can Do It with Fewer Tokens

TL;DR ACE and ALTK-Evolve both turn agent trajectories into reusable lessons without...

信息来源：DEV Community

教程 / 实战普通2026-09-24 23:03

Endura Therapeutics CEO 谈 Foundries 与 Navigators 如何降低科学成本

Endura Therapeutics CEO Adrian Sanborn 在 Latent Space 发表客座文章，提出生物科技公司应对 AI 的两种路径：用新技术把实验数据产出提速一个数量级的 Foundries，以及把 LLM 融入日常运营提升决策的 Navigators。

信息来源：Latent Space（RSS）

教程 / 实战普通2026-09-24 23:00

Shield AI、Waabi 与通用汽车将在 TechCrunch Disrupt 2026 探讨"失败不可接受"时的 AI 系统构建

TechCrunch Disrupt 2026 将于 10 月 13–15 日在旧金山 Moscone West 举办，Shield AI CTO Nathan Michael、Waabi 创始人兼 CEO Raquel Urtasun 与通用汽车机器人战略总监 Mikell Taylor 将在 Real World AI Stage 讨论高风险场景下的 AI 系统构建。

信息来源：TechCrunch：AI（RSS）

教程 / 实战普通2026-09-24 22:49

You Can’t Buy “The Best Local AI PC” in 2026

In 2026, personal AI hardware has split into two paths: CUDA compute vs massive Unified Memory. Here is how to choose which constraint to remove first.

信息来源：DEV Community

教程 / 实战普通2026-09-24 22:45

Microsoft Agent 365 Supports n8n Agents in Microsoft 365 With Governance

Microsoft Agent 365 supports third-party agents built with n8n, creating a route for n8n automations...

信息来源：DEV Community

教程 / 实战普通2026-09-24 21:58

Your AI Code Reviewer Needs a Test Suite Too

🤖 The Problem You plugged an LLM into your PR pipeline. It leaves comments about SQL...

信息来源：DEV Community

教程 / 实战普通2026-09-24 21:50

GraphSentinel: a fraud investigator that knows when to stop and ask

Most fraud demos end with a big red "FRAUD 0.94". Real fraud desks don't work like that. An analyst...

信息来源：DEV Community

教程 / 实战普通2026-09-24 21:18

100% vuln detection wasn't enough: measuring whether AI respects the patch

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

信息来源：DEV Community

教程 / 实战普通2026-09-24 21:00

Shapes are the type system of ML, and nobody checks them for you

Originally published at cchinchilla.dev. Part 3 of From code to weights, a 12-part series on ML...

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:47

Your agent framework stops at the demo. This open-source stack ships to production.

A technical deep dive into apowerb, the open-source AI agent runtime by [thaink²](https://thaink2.com/opensource) database-materialized agents, RAG, Text-to-SQL, event-driven execution, and how it compares to LangGraph, CrewAI and n8n.

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:31

Meta 承认 Muse 深受 OpenClaw 启发但为从零构建

Meta MSL 产品负责人 Nat Friedman 在 X 上回应称，Muse 从零构建，但在产品层面深受开源项目 OpenClaw 启发，目标是打造类似产品并扩展到数十亿用户。此前有用户发现 Muse 与 OpenClaw 使用同名配置文件，其中 SOUL.md 内容几乎完全一致；Muse 近期已登顶美国 App Store 免费榜，早期增长速度超过 ChatGPT。

信息来源：IT之家（RSS） · ChatGPT

教程 / 实战普通2026-09-24 20:30

From Fraud Alert to Governed Action: Building an Agentic Fraud Investigation System with TigerGraph

How we built an evidence-driven, uncertainty-aware fraud investigation agent using TigerGraph, GraphRAG, deterministic planning, policy governance, and Next Best Action.

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:22

7 Agent Eval Mistakes That Cost Me Weeks (And the One-Line Fixes That Ended Them)

You've been there. You run the eval, the number comes back, and something about it doesn't sit...

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:15

Stop Slapping "AI" Onto Legacy Code: AI Feature vs. AI-native Architecture

Every tech stack today is legally required to have an "AI-powered" badge. Go-to project management...

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:07

Build an Invoice Verification App With ToolJet MCP

Introduction Learn how to build an invoice verification app with ToolJet MCP for an...

信息来源：DEV Community

教程 / 实战普通2026-09-24 20:00

A Vendor Deleted a Shutdown Date and Nothing Told Anyone. So I Counted the Next 90 Days.

On 2026-07-28, Google's deprecation page listed October 16, 2026 as the shutdown date...

信息来源：DEV Community

教程 / 实战普通2026-09-24 19:55

用 google-cloud-developer 插件让智能体按 Google Cloud 最佳实践完成部署实测

Remigiusz Samborski 用 Antigravity CLI 实测 Google 的 google-cloud-developer 插件，一条命令安装五个技能和 Developer Knowledge MCP 服务器，支撑 Claude Code、Codex CLI 等智能体。

信息来源：Google AI：DEV 作者专属（RSS） · Claude

教程 / 实战普通2026-09-24 19:45

Two weeks of serving Markdown to agents, straight from the nginx logs

Since early September, every public page of our website has had a Markdown twin, reachable by header...

信息来源：DEV Community

教程 / 实战普通2026-09-24 19:06

Devlog: I Built a 3D Library in Three.js Without a Level Editor — So I Made My Own

Updated 3D Environment for Library Reading DEV Articles A lot of you...

信息来源：DEV Community

教程 / 实战普通2026-09-24 18:56

I rebuilt my "export PDF text for AI" feature three times in one day. Here's what each version got wrong.

In July 2025, researchers found manuscripts on arXiv with instructions such as "give a positive...

信息来源：DEV Community

教程 / 实战普通2026-09-24 18:43

Sanity's Knowledge Base stopped 6 of 7 poisoned pages. The 7th fooled Claude Opus 5.

This is a submission for the Sanity Challenge: Ship an agent that queries real content. Every help...

信息来源：DEV Community · Claude

教程 / 实战普通2026-09-24 18:28

Confused Deputy: The Old Bug That AI Agents Keep Reintroducing

Back in 1988, a compiler running on a commercial timesharing system had permission to write to a...

信息来源：DEV Community

教程 / 实战普通2026-09-24 18:11

I Made a VS Code Extension to Copy Your Repo to Your Clipboard as Clean Markdown Context for Your Chatbot

The UI: An example output: # Project...

信息来源：DEV Community