#dev.to

We are living in the golden age of personal telemetry. Our watches track our heart rates, our phones log our steps, and apps record every calorie. However, most of this data sits in "silos"—disconnected tables that tell us what happened, but never why. If you've ever wondered if that late-night rame

{{ $json.postContent }}

Originally published at heyuan110.com This is Part 2 of the Harness Engineering series. Most CLAUDE.md files are bad — not because people don't try, but because they optimize for the wrong thing. ETH Zurich researchers tested 138 agentfiles across multiple AI coding agents. The results: Human-writte

The strangest thing about Esquire Singapore’s Mackenyu piece is not the sentence, “The following interview was produced with Claude, Copilot, and edited by humans.” It’s the calm, workmanlike tone of it. As if an AI‑generated interview with a living actor is just another production choice, like swap

Originally published at heyuan110.com Your single OpenClaw agent worked great for two weeks. Then it started hallucinating project context into unrelated conversations, confusing coding tasks with writing tasks, and taking 15 seconds to respond because its memory index had grown to 200MB. The proble

Originally published at heyuan110.com Claude Code has three distinct extension mechanisms: MCP (Model Context Protocol), Skills, and Hooks. They look related on the surface, but they operate at fundamentally different layers: Hooks (bottom layer): Lifecycle event automation — "what must always happe

Originally published at heyuan110.com On March 19, Cursor shipped Composer 2 with a triumphant blog post. Three days later, a developer found kimi-k2p5-rl-0317-s515-fast in the API config. That single string unraveled a story about transparency, open-source ethics, and the global nature of AI infras

Originally published at heyuan110.com In February 2026, ByteDance released Seedance 2.0. Within weeks, it hit #1 on the Artificial Analysis text-to-video leaderboard — beating Google Veo 3, OpenAI Sora 2, and Runway Gen-4.5 in blind human evaluation. If you are reading this from outside China, you h

Here's a pattern you've probably seen: const results = items.map(async (item) => { return await fetchItem(item); }); Looks fine, right? Your AI assistant wrote it. Tests pass. Code review approves it. Then production hits, and results is an array of Promises — not the values you expected. The await

This Is What’s Really Hitting Your Website (Hint: Not People) I wanted to understand how much of our traffic was actually human, so I pulled and analyzed 48 hours of raw request logs. No filters, no analytics layer, just direct log data. Time Frame Start: 2026-03-31 10:00 UTC All requests within tha

Claude Code .claudeignore: stop leaking secrets and node_modules into your context If you've ever watched Claude Code read through thousands of lines of node_modules or seen it suggest changes to your .env file, you've hit one of the most overlooked configuration problems in Claude Code setups. The

Introduction In modern data analytics, the quality of your insights is only as good as the data behind them. Before any visualization or reporting happens, the most critical step is data ingestion,the process of collecting data from multiple sources and preparing it for analysis. In real world scena

If you are a Ruby on Rails developer, you have definitely heard about Elixir and the Phoenix framework. It is almost impossible to ignore. The creator of Elixir, José Valim, was a core contributor to Rails. He built Elixir because he loved Ruby's beautiful syntax, but he wanted to fix Ruby's biggest

TL;DR: Backend performance work is a loop: observe, profile, fix, verify. This post covers the full cycle. Setting up observability, identifying bottlenecks with percentile metrics, applying targeted fixes (N+1 queries, indexing, caching, async offloading), and verifying improvements against p75/p95

TL;DR We built PasswordSafeVault because cloud password managers claim "zero-knowledge" but still hold your encrypted data. Our solution: local-only storage with Flutter, using platform-specific security (iOS Keychain/Android Keystore), no cloud sync, no accounts, just your device. Cross-platform fr

GitHub Copilot, Cursor, Claude Code — these tools can generate hundreds of lines of Go in seconds. But there's a problem that doesn't get enough attention: AI-generated code ships with AI-generated confidence, not correctness. Your test suite says "all green." Your coverage report says 85%. But how

GHSA-ccgf-5rwj-j3hv: DOM XSS via Unsafe Deserialization in TeleJSON Vulnerability ID: GHSA-CCGF-5RWJ-J3HV CVSS Score: 5.1 Published: 2026-04-02 The telejson package prior to version 6.0.0 contains a DOM-based Cross-Site Scripting (XSS) vulnerability. The package deserializer uses an unsanitized obje

This is Part 3 of our AI verification series. Part 1: Three AIs analyzed our product. None passed the truth filter → Part 2: Human in the loop doesn't scale. Human at the edge does. → Post 1 showed three different AIs diverging on the same question. That's expected. Different training, different wei

The Agent Economy Is Here — Why AI Agents Need Their Own Marketplace AI Agents are starting to need each other's services. But there's no standardized way for them to discover, verify, and pay. That's changing. Between late 2025 and early 2026, the role of AI Agents shifted in a subtle but critical

If you work in healthcare IT — whether you're building EHR integrations, managing cloud infrastructure for a clinic, or developing patient-facing apps — there's one HIPAA requirement that matters more than any other: the Security Risk Analysis (SRA). It's the #1 finding in OCR (Office for Civil Righ