How to build an LLM evaluation from real failures
Build the eval from failures you have actually read, rather than from a rubric you wrote in advance. Group those failures, put one domain expert in charge of th…
Englishdev.to Testing
最新资讯
言栖资讯收录 178457 篇 AI、开发与建站文章，按发布时间排序，支持搜索与按语言筛选。
OMID browserOptional describes the JavaScript environment.
The name browserOptional is easy to misread as permission to skip verification JavaScript. It describes the script's execution environment. Confusing those mean…
Englishdev.to Testing
My app passed 1,100 tests. Six minutes on a real phone found five bugs.
I'm building a budgeting app called Budget Lover. It is a personal project, written in Flutter with Firebase behind it, and it tries to answer one question: how…
Englishdev.to Testing
Building Pixel Control Center: High-Performance Android Mirroring on Desktop with Flutter & Scrcpy
If you have ever mirrored an Android handset to your PC or Linux workstation using scrcpy , you know it is the gold standard for low-latency screen casting and …
Englishdev.to OpenSource
How I built a local AI assistant that admits when it didn't actually do the thing
How I built a local AI assistant that admits when it didn't actually do the thing The sentence that started this project was five words long: "I've sent the ema…
Englishdev.to OpenSource
Load every server in the official MCP registry into pandas: 55,773 rows, three ways to get them
If you want to know what is actually in the official Model Context Protocol (MCP) registry, you can click through 55,000 pages, or you can load one file. I keep…
Englishdev.to OpenSource
Let your AI assistant draft customer replies (with a human in the loop): elozi-mcp
I build Elozi, a customer-service workspace. I wanted AI assistants to help with support without giving them free rein, so I published elozi-mcp (MIT, v0.1.1). …
Englishdev.to OpenSource
CI 红了先别改脚本：先看 job 是断言失败、还是根本没拿到 runner
（整理、非新实测。以下失败记录均为 2026-10-01 以后 Lite 仓 test-install 的真实 run 结论，时间与结论以 GitHub Actions 页面为准；价格与链接见文末。） 先说结论 同样一个红叉，至少分三种死法，修法完全不同： 断言失败：脚本或安装步骤真的跑挂了，改代码才有用。 runne…
Englishdev.to OpenSource
How to Review AI-Generated Code in 2026: The 5 Decisions I Keep Human
Picture a team whose agents now read every line of every PR. Which parts of review should people still do by hand? My answer is five questions, none of them abo…
Englishdev.to Programming
Designing Resilient Circuit Breakers with Fallback Strategies: Mitigating Cascading API Failures
Designing Resilient Circuit Breakers with Fallback Strategies: Mitigating Cascading API Failures 1. System Requirements and Verification Architecture In distrib…
Englishdev.to Programming
Password Storage: What to Use in 2026
Password storage is one of the few areas in software with a genuine right answer, published, maintained and free. It is also one of the most consistently got wr…
Englishdev.to Programming
How to Mask Sensitive Data in SQL: Make a Safe Database Copy
Data masking replaces sensitive values, like names, emails, and card numbers, with hidden versions. The result is a database copy you can share with developers …
Englishdev.to Programming
Never Let a Display Toggle Write to the Database
A support ticket arrives. Two users are looking at the same record and reading different numbers. Nothing in the logs. No exception, no failed request, no alert…
Englishdev.to Programming
A Complete Free TypeScript Curriculum — 107 Chapters, From Basics to Compiler Internals
I've been looking for a TypeScript resource that doesn't stop halfway. Most either cover the basics and quit, or assume you already know what a conditional type…
Englishdev.to TypeScript
Never Call an LLM From Your ERP: AI Middleware Architecture
Your ERP just called an LLM synchronously. The model took 31 seconds. The ERP gave up at 30. Now 1,200 orders are stuck between "paid" and "shipped", and nobody…
Englishdev.to TypeScript
Network Timeout Causes: External Logs vs Embedded Checks for Failure Alerts
A rollback can erase the evidence needed to identify network timeout causes. For a logistics notification service, prefer external logs and metrics over checks …
Englishdev.to TypeScript
Failed Email API Error Logging and Retry Strategies (with a Compliance Evidence Boundary)
For failed email API deliveries, error logging is an evidence problem before it is a retry problem. For a developer-tools support queue, record the request, the…
Englishdev.to TypeScript
Nuxt CLI v4 Is Great for AI Coding Agents
Have you ever had a dev server running, and then asked your coding agent to test something in that app? It starts its own server on port 3001. Then it restarts …
Englishdev.to Vue
Building a weather new tab extension for Firefox: wttr.in, world clocks, no backend
Every new tab is a blank canvas. I wanted mine to tell me the weather and the time in the cities I care about — without opening another app or asking a backend …
Englishdev.to JavaScript
Settings and the Keybindings editor
JSLab feature notes: one docs page, explained. Today: Settings and Keybindings and the command palette . Settings: one window, one search field ⌘, (or JSLab → S…
Englishdev.to JavaScript