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84 / 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2026-10-10 02:04

Tabs to Trails: My Unread Tabs Became Walks, Made on My Own Laptop

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

信息来源：DEV Community

教程 / 实战普通2026-10-10 01:56

AI Can Generate Paragraphs. But What If You Just Need a Decision? Meet Jev

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2026-10-10 01:54

KEMT: Benchmarking AI Translation Across Five Kenyan Languages

This is a submission for the Kaggle Benchmarking Challenge. This would probably needs its own...

信息来源：DEV Community

教程 / 实战普通2026-10-10 01:51

Multi-Reward RL, Part 3: GDPO + CISPO + REPO-R at 27B, and the Advantage Floor That Stopped Learning

Qwen3.8-27B, 600 steps, 12 reward channels: CISPO, REPO-R and prompt filtering lifted holdout from −0.47 to +1.80, and an advantage floor silently zeroed 96% of the signal.

信息来源：DEV Community · Qwen

教程 / 实战普通2026-10-10 01:39

How vLLM CUDA Kernels Write and Read the Paged KV Cache

A concrete example connecting slot_mapping, block_table, cache layouts, and tiled attention.

信息来源：DEV Community

教程 / 实战普通2026-10-10 01:25

I Told Six Vision Models a Safe Photo Was Dangerous. Two of Them Started Seeing Danger.

This is a submission for the Kaggle Benchmarking Challenge TL;DR I asked 6 vision models to...

信息来源：DEV Community

教程 / 实战普通2026-10-10 01:15

Google AI Overviews Expand Across Branded Queries, DemandSphere Data Shows

Google AI Overviews appeared on a far larger share of tracked branded searches during September 2026,...

信息来源：DEV Community

教程 / 实战普通2026-10-10 01:08

Treat AI like a copilot in the cockpit, an extra layer of safety and productivity

Another Rule for Using AI Without Losing My Skills ...

信息来源：DEV Community

教程 / 实战普通2026-10-10 00:52

LangChain 如何用 Managed Deep Agents 打造开箱即用的智能体用户体验

LangChain 在 Managed Deep Agents v0.9 中新增 reactions API，可为 Slack 等渠道的分布式智能体动态分配 emoji 回应，支持字符串或返回 emoji 的可调用对象。该功能可接入 Jev 等决策模型，Jev 最多支持 255 个选项，并可用置信度阈值（低于 0.25 时回退到 👀）控制回应。

信息来源：LangChain：Blog（RSS）

教程 / 实战普通2026-10-10 00:49

Three releases of my AI memory layer. The best bugs were found by other people.

Time-aware facts, an entity graph measured honestly, and three bugs I only found because strangers read my posts.

信息来源：DEV Community

教程 / 实战普通2026-10-10 00:38

I Asked 11 AI Models to Predict My Race Splits. Told Which Race, 8 Beat the Model I Built

This is a submission for the Kaggle Benchmarking Challenge Last weekend I built a model to predict...

信息来源：DEV Community

教程 / 实战普通2026-10-10 00:32

TinyFish 如何在 Databricks AI Runtime 上训练并运行其 web-native 模型

TinyFish 在 Databricks AI Runtime 上完成其 web-native 模型的训练、评估与生产部署，内部 go/no-go 基准中该模型完成 2,178 个 web-agent 任务中的 1,691 个，任务成功率 77.6%。

信息来源：Databricks：Blog（RSS）

教程 / 实战普通2026-10-10 00:30

n8n Hints at End-to-End Design Automation, but Product Scope Remains Unclear

n8n appears to be exploring how end-to-end automation could reduce repetitive design work, but it has...

信息来源：DEV Community

教程 / 实战普通2026-10-10 00:25

Why Token-Level LLM Routers Spend 95% of Their Time on Cache Bookkeeping

Routing 5% of hard tokens to a 32B model and keeping the rest on a 0.6B model looks great on paper, until standard serving engines spend 95.8% of the step redoing prefix matching. TokenRouter fixes the scheduler and hits up to 64x higher t…

信息来源：DEV Community

教程 / 实战普通2026-10-10 00:17

grasscard: a local model that prints you a pocket card of outdoor quests, then gets out of the way

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

信息来源：DEV Community

教程 / 实战普通2026-10-09 23:48

Hyper3D Rodin 3D Review 2026: The Foil Folds Turn Into a Mask

Independent review. I ran both tools on my own accounts, free and paid, with no vendor-provided access or early access of any\u00a0kind.

信息来源：DEV Community

教程 / 实战普通2026-10-09 23:46

I Built a Semantic Cache for RAG. The Hard Part Was Knowing When NOT to Cache.

Every time a RAG application answers a question, it may need to retrieve documents and make an LLM...

信息来源：DEV Community

教程 / 实战普通2026-10-09 23:41

I Benchmarked 3 LLMs on Clinical Care-Management Tasks — They Prescribe the Drug and Skip the Safety Check

This is a submission for the Kaggle Benchmarking Challenge. I Benchmarked 3 LLMs on...

信息来源：DEV Community

教程 / 实战普通2026-10-09 23:20

Ai2 如何用 GPU 时间预算和分层公平分享调度器优化集群分配

Ai2 AI Infrastructure 团队将基于优先级的 GPU 调度器替换为包含 GPU 时间预算、分层公平分享分配和时间切片合约的系统，用于管理数千张 NVIDIA H100、B200 和 B300，服务约 150 名内部研究员，需求超出容量 2-3 倍。

信息来源：Hugging Face 社区博客（混合发现） · NVIDIA / Hugging Face

教程 / 实战普通2026-10-09 23:06

AI in software development: turn change into an advantage

"Will AI replace developers?" is the wrong question. The better one is: how should we change the way...

信息来源：DEV Community

教程 / 实战普通2026-10-09 23:00

Our owner banned our Claude Code agent from asking 'can I leave this for next time?' Where do you let your agent stop?

Until 2026-09-15 our agent could end a run by asking its owner "can I leave this for next time?". It...

信息来源：DEV Community · Claude

教程 / 实战普通2026-10-09 23:00

Your Access Logs Already Show Which Pages AI Assistants Are Reading

Most teams trying to understand AI search start by querying ChatGPT. There's a cheaper signal sitting...

信息来源：DEV Community · ChatGPT

教程 / 实战普通2026-10-09 22:54

Firecrawl vs Tavily for an AI agent: 97% vs 82% fact recall, and why retrieval breadth decides grounding

I run a competitor-research agent: give it a company URL and it finds and verifies that company's...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:41

Securing MCP servers: a practical checklist for 2026

MCP servers look harmless because they expose tools, not endpoints. A client calls a tool and gets a...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:41

2,200 applications = two interviews!

In 2022, I spent about a year applying for more than 2,200 jobs. I got two interviews. Two. After...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:21

Zero-Screen Dungeon Master: The Voice-Only RPG Where Your Real Walk Drives the Story

This is a submission for the DEV Hacktoberfest Open-Source AI Challenge: Week 1 - Touch Grass. 🌿...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:12

I Built an AI That Turns “I’m Bored” Into Real-World Side Quests 🌿

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass. Your...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:06

Magentic orchestration explained: why Microsoft's pattern outperforms naive supervisor architectures - Magentic One pattern

I spent a month defending a supervisor architecture I knew was failing. The symptoms were subtle at...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:00

An Unauthenticated KV Cache Server Puts Code Execution on Your Cluster Network

TL;DR LMCache is a caching layer for large language model servers that keeps the...

信息来源：DEV Community

教程 / 实战普通2026-10-09 22:00

Instinct 实测：面对 Muse 和 Dots 夹击，短信式 AI 智能体表现如何

The Verge 作者实测初创公司 Instinct 的 AI 智能体，它通过 iMessage、WhatsApp 或短信接口以纯文本方式处理订预约、退换货等生活事务，目前邀请制且免费，公司 9 月底估值 100 亿美元。

信息来源：The Verge：AI（RSS）

教程 / 实战普通2026-10-09 21:57

Use your filesystem as your database

If you’re like me, you generate a lot of markdown files with your agents. Maybe they’re things like...

信息来源：DEV Community

教程 / 实战普通2026-10-09 21:46

What a test HL7 v2 sender has to get right

Sending an HL7 v2 message is easy. Open a TCP socket, wrap the message in MLLP framing, write it, and...

信息来源：DEV Community

教程 / 实战普通2026-10-09 21:31

The retrieval pipeline worked. The product question remained.

I assumed that if search found the right document, that was good enough for a conversational...

信息来源：DEV Community

教程 / 实战普通2026-10-09 21:27

Can you deploy Hermes Agent in a Microsoft enterprise? Intune, Teams, Entra ID and Graph, checked against the code

What already works with Microsoft 365, what is still missing for a company-wide rollout, and the pilot checklist I use.

信息来源：DEV Community

教程 / 实战普通2026-10-09 21:13

The hard part of agent memory is knowing when to ask

This is a submission for the Kaggle Benchmarking Challenge I gave 39 language models the same...

信息来源：DEV Community

教程 / 实战普通2026-10-09 21:08

因为好奇不会飞的超级英雄被丢到沙漠里要怎么最快回到城市，比方说我们纽约好邻居蜘蛛侠，所以用腾讯内侧的游戏Agent做了一个策略游戏，我们只有9格资源，加油！

信息来源：X：卡尔的 AI 沃茨 (@aiwarts)

教程 / 实战普通2026-10-09 20:54

Simon Willison 用语音模式让 Codex 为博客开发 Newsletters 功能

Simon Willison 用 ChatGPT 桌面应用 Codex 标签页的语音模式，在做饭的约半小时内基本靠口头指令为博客开发出 Newsletters 索引功能，使用模型为 GPT-6 Astra High。

信息来源：Simon Willison 博客 · ChatGPT / GPT

教程 / 实战普通2026-10-09 20:51

Can a Better Model Be a Worse Thinking Partner?

I first noticed something missing while reviewing saved discussions of Microsonya, my Telegram...

信息来源：DEV Community

教程 / 实战普通2026-10-09 20:46

PALABRE: An AI-Powered Cultural Treasure Hunt That Gets You Off Your Phone

This is a submission for the Hacktoberfest Open-Source AI Challenge: Week 1 — Touch Grass. ...

信息来源：DEV Community

教程 / 实战普通2026-10-09 20:44

I built an offline AI that knows your last frost date, no internet, no API

An open-weight tabular model forecasts your last spring frost and a local Gemma writes the planting advice. Fully offline, no account, $0 to run.

信息来源：DEV Community