AI 资讯

The new Siri AI is conversational, omnipresent, and actually helpful.

\shadcn/ui and Material UI optimise for opposite priorities. Choose shadcn/ui to own your component code, ship a near-zero runtime, and control every pixel; choose Material UI (MUI) for breadth — 90+ components and a paid data grid — behind Google's Material Design. shadcn/ui has ~116,000 GitHub stars and ships copy-paste components; MUI has ~98,000 stars and ~7.3M weekly npm downloads. Both are MIT-licensed and free for commercial use. This guide covers the parts you only learn by shipping both: how each behaves in the Next.js App Router, the runtime cost, real theming and dark-mode code, forms, data tables, and migration mechanics. What's the real difference between shadcn/ui and Material UI? The difference is ownership, and it decides everything downstream. MUI is an npm dependency ( @mui/material ) you install and import from node_modules — you never touch the source. shadcn/ui is a copy-paste registry: you run a CLI, the component lands in your repo, and it is now your code. shadcn/ui is unstyled, built on Radix UI primitives and Tailwind CSS. MUI ships Material Design and an Emotion (CSS-in-JS) runtime. With MUI you install and import: bash npm install @mui/material @emotion/react @emotion/styled cta.tsx import Button from " @mui/material/Button " ; export function Cta () { return < Button variant = "contained" > Get started </ Button >; } With shadcn/ui the CLI copies the source into your project and you import from your own path — there is no library to upgrade or override: bash npx shadcn@latest add button cta.tsx import { Button } from " @/components/ui/button " ; export function Cta () { return < Button > Get started </ Button >; } That ownership changes how you customise. shadcn's button.tsx lives in your repo and uses class-variance-authority (cva) for variants — you add one directly: components/ui/button.tsx // components/ui/button.tsx — this file is yours const buttonVariants = cva ( " inline-flex items-center justify-center rounded-md ... " , { varia

How to Build Systems That Actually Know Your Data (Not Hallucinate About It) Introduction: The Story Begins 👦 Nephew: Uncle, I keep hearing "RAG this, RAG that" in tech interviews. When I ask what it means, people throw around words like "Retrieval-Augmented Generation" and I just nod like I understand. But honestly? I'm lost. 👨‍🦳 Uncle: (laughing) That's the best honest question I've heard all week. Let me ask you something first. If I gave you a question right now - "What year did India win the World Cup?" - how would you answer? 👦 Nephew: Well... I'd pull up Google, search for it, read the answer, then tell you. 👨‍🦳 Uncle: Exactly. You don't answer from memory alone. You go fetch the information first, then answer based on what you found . That's RAG in real life. And that simple idea - fetch first, answer after - fixes almost every problem we face with AI today. 👦 Nephew: But uncle, AI can remember things from its training. Why does it need to fetch? 👨‍🦳 Uncle: Ah! That's where we land in trouble. Come, sit... SECTION 1: RAG FUNDAMENTALS - The Core Concept The Problem We're Actually Solving 👨‍🦳 Uncle: Imagine you're hiring for a tech company. You receive 500 resumes for a Senior React Developer role. Now tell me - how would you actually process them? 👦 Nephew: I'd... probably make a spreadsheet? List all the candidates with key skills? 👨‍🦳 Uncle: Right. But here's the catch - you can't read all 500 resumes deeply. So what do you really do? 👦 Nephew: Skim for keywords like "React", "JavaScript", "5 years"? 👨‍🦳 Uncle: Exactly. You skim and hope you don't miss anyone good. Now, here's the problem: what if a candidate wrote "React.js" instead of "React"? Your eyes might still catch it. But a dumb computer doing exact string matching? It says "no match". What if someone wrote "Built real-time user interfaces with the React framework"? The candidate clearly knows React, but the word "React" appears nowhere in that sentence. The computer misses them. This is exactly wh

AI coding agents are expensive — not because models cost too much per token, but because they send too many of them. An SRE debugging session with a raw agent: 65,694 tokens in. With Headroom in the middle: 5,118. Same bug found. Headroom is a new open-source context compression layer that intercepts everything your agent reads — tool outputs, log dumps, RAG chunks, files, conversation history — and compresses it before the LLM ever sees it. It's local, reversible, and available as a drop-in proxy, a library, or an MCP server. The numbers that matter Savings on real agent workloads: Code search (100 results): 17,765 → 1,408 tokens (92% reduction) SRE incident debugging: 65,694 → 5,118 tokens (92%) GitHub issue triage: 54,174 → 14,761 tokens (73%) Codebase exploration: 78,502 → 41,254 tokens (47%) Accuracy on standard benchmarks (GSM8K, TruthfulQA, SQuAD v2, BFCL) is preserved — some scores actually improve slightly, likely because the model sees cleaner signal. What's doing the compression Under the hood, Headroom routes content through a stack of specialised compressors: SmartCrusher — JSON, nested objects, arrays of dicts CodeCompressor — AST-aware for Python, JS, Go, Rust, Java, C++ Kompress-base — a custom HuggingFace model trained on agentic traces, for prose and mixed content CacheAligner — stabilises prompt prefixes so Anthropic/OpenAI KV caches actually hit It also does CCR (reversible compression) — originals are cached locally and the LLM can retrieve them on demand if it needs them. Nothing is destroyed. Why the proxy mode matters The most interesting deployment path: headroom proxy --port 8787 , then point your existing tool at localhost. Zero code changes. Works with any language. Or even simpler: headroom wrap claude wraps Claude Code, routes its traffic through Headroom automatically. One command, savings start immediately. Same for Codex, Cursor, Aider, Copilot CLI. "Library — compress(messages) in Python or TypeScript, inline in any app. Proxy — hea

Hello Dev Community! 👋 It is officially Day 50 — a massive half-century milestone on my daily, unbroken streak toward mastering full-stack MERN engineering! Reaching Day 50 feels absolutely incredible. Yesterday, I mapped out dynamic path parameters. Today, I wired the input engine by building a complete asset workflow: Capturing Host "Add New Product" data payloads and committing them to local file storage pipelines! Following Prashant Sir's backend sequence , today was all about bridging the gap between host client forms and backend architecture using the Model-View-Controller framework. 🧠 Key Learnings From Day 50 (Product Ingestion & Storage) Processing data mutations sent from input forms requires tight coordination between parsing middlewares and file serialization engines. Here is how I structured the logic today: 1. Intercepting Form Submissions ( POST /host/add-product ) Set up a clean route mapping inside hostRouter.js to process dynamic data blocks sent by the host. The endpoint parses input parameters securely via backend streams. 2. Utilizing Class Instances for Storage Instead of directly pushing raw unstructured dictionaries into file records, I initialized a new object instance using my Day 48 structural class framework ( new houseList(...) ). This forces incoming data attributes—like name, price, location, and images—to match my exact system layout blueprint. 3. Asynchronous File Serialization Invoked the instance method .save() , which runs a non-blocking background task: it reads the active database layout array inside homesdata.json , appends the newly formulated object safely, and flushes the stringified update back onto the hard drive array using Node's fs operations. javascript // A conceptual look at how my controller hands data over to the model layer today const Product = require("../model/home"); exports.postAddProduct = (req, res) => { const { title, price, location, rating, imageUrl } = req.body; // Instantiating the core class data mold

Feeling overwhelmed by the constant flood of new AI features, MCP servers, and agentic platforms? In a world full of tech noise, it's easy to get exhausted trying to keep up. I just watched an incredible video by Burke Holland where he strips away the hype and shares 10 highly practical, concrete strategies to make AI coding tools actually work for your daily workflow. If you want to stop overcomplicating your setup and start getting better production results, here is the ultimate breakdown. The 10 AI Coding Tips (TL;DR Summary) Huge shoutout and credit to Burke Holland for these insights: 1) Use Visual Studio Code to maximize your environment with powerful themes, extensions, and inline terminal chats. 2) Always turn on YOLO / "allow all" mode so your AI agent can execute commands seamlessly without breaking your flow with constant permission prompts. 3) Never run agents on your own machine , choosing instead to isolate them via remote SSH or dev containers so YOLO mode is completely safe. 4) Prototype and mock everything upfront to map out UI design languages and logic before implementing code. 5) Always plan and grill by leveraging interactive planning modes to answer critical edge-case questions before generating file. 6) Rubber duck your plans across different AI model families (like combining Claude and GPT) to cross-verify solutions and expose blind spots. 7) Utilize autopilot and sub-agents to delegate parallel tasks and route smaller, faster models where appropriate. 8) Use built-in browser tools to visually review live previews and directly prompt structural or stylistic adjustments. 9) Run iterative multi-model reviews on autopilot to catch hidden bugs and refine code quality until reaching a clear point of diminishing returns. 10) Learn from your session history using tools like Chronicle to analyze your prompting habits and continually optimize how you interact with the agent. 📚 Recommended Reading If you are looking to dive deeper into perfecting your

OpenAI, Anthropic, and Gemini each report token usage differently, and it stops being trivia the moment you track LLM cost. I build Spanlens, an open-source LLM observability tool that sits in front of all three as a proxy and records every call with its model, latency, tokens, and cost. To do the cost part I read the token usage back out of every response, including the streaming ones. I assumed the three providers would report usage in roughly the same way. They send the same kind of data, after all: input tokens, output tokens, maybe a cached count. How different could it be. Pretty different, it turns out. Here is the whole thing in one table, then each gotcha in detail with the real parser code from the repo. Provider Where usage lives (streaming) Cache accounting Field names OpenAI final chunk, needs stream_options: { include_usage: true } prompt_tokens includes cache prompt_tokens / completion_tokens Anthropic split across message_start + message_delta input_tokens excludes cache, so add it input_tokens / output_tokens Gemini usageMetadata , two stream formats not applicable promptTokenCount / candidatesTokenCount Gotcha 1: the usage numbers live in different places in the stream For a non-streaming call this is boring. Every provider hands you a usage object on the response body and you read it. Streaming is where it gets weird, because the token counts are not in the content chunks. They show up somewhere else, and "somewhere else" is different for each provider. OpenAI puts the usage in a final chunk, after all the content, right before [DONE] . You only get it if you ask for it with stream_options: { include_usage: true } . Miss that flag and you stream the whole response and end up with no usage at all. export function parseOpenAIStreamChunk ( line : string ): Partial < ParsedUsage > | null { if ( ! line . startsWith ( ' data: ' )) return null const data = line . slice ( 6 ). trim () if ( data === ' [DONE] ' ) return null const json = JSON . parse ( data

Working with AI Means Thinking More, Not Less Yes, this text is long. Yes, it repeats itself in places. I did not clean that up. A text that sounded too smooth while arguing that AI forces you to think more, not less, would be at least slightly dishonest. This is not fast food for quick consumption. And yes, don’t worry: you won’t hear anything especially new here. That is part of the problem too. There is a popular and very seductive story about AI in software development. Now that the machine can write code, the human gets to think less. You just point it in the right direction, and the model will quickly and cheaply do a significant part of the work on its own. In that picture, AI is primarily an accelerator for code production, and human thinking gradually shifts from necessity to optional extra. I keep feeling more and more strongly that this description is dangerously wrong. A more accurate formula for my own experience right now is this: I’m the tech lead, the AI is the entire team in one body . And if you take that metaphor seriously, the conclusion is the exact opposite of the mainstream narrative. Working with AI is not a way to think less. It is a mode in which you need to think more, not less . Not because the AI is bad. But because it is too good at one very treacherous thing: it confidently and smoothly fills in what was left unsaid. I’m the tech lead, the AI is the team At first this metaphor felt like a neat formulation. Now it feels like a literal description of what is going on. If you treat AI as a very fast and very capable executor, a lot of things become clearer immediately. It really can wipe out months of routine work. It can spin up prototypes quickly, take over test scaffolding, try out alternatives, make local edits, help break a task into parts, and sometimes even suggest a decent direction. On the surface, this really does look like a silver bullet. Especially if the human knows the stack and can read code. The pace becomes so extreme th

Plus: Gay bars in San Francisco using face scanners, France quits Palantir, Apple plans to change its private email and more.

You have ISTQB Foundation. ISTQB Advanced. Certified ScrumMaster. A cloud cert. A security testing cert. Maybe a Python for Testers badge from a platform. And you still cannot write a test that finds a real bug. I interviewed someone like you last quarter. The resume was a wall of acronyms. The conversation was a wall of theory. "I follow the V-model." "I use equivalence partitioning." "I believe in shift-left." Then I asked: "Show me one test you wrote that caught something the developer missed." Silence. Not because they were nervous. Because they had never written a test that found a bug. They had written tests that passed. They had written tests that covered requirements. They had never written a test that broke something. That is the difference between a certification holder and a tester. Certifications test your memory. Bugs test your thinking. Let me show you what I mean. The Certification Trap Certifications are not useless. They give you vocabulary. They give you structure. They give you something to put on LinkedIn so recruiters stop asking if you know what a test case is. But they do not teach you how to find bugs. Here is why. Every certification exam tests known knowledge. You study a syllabus. You memorize definitions. You answer multiple-choice questions about boundary value analysis. You pass. Then you sit in front of an application. The application does not have a syllabus. It does not have a boundary value analysis section in the documentation. It has a login form that sometimes lets you in with a password that is clearly wrong, but only on Tuesdays, and only if the server clock is behind by exactly four minutes. No certification prepares you for that. The tester with 10 certifications treats testing like a checklist. They write test cases from requirements. They execute them. They mark pass or fail. They report coverage metrics. The tester who finds bugs treats testing like an investigation. They start with a hypothesis. They try to prove the appl

Hello, I'm Maneshwar. I'm building git-lrc, a Micro AI code reviewer that runs on every commit. It is...

Architecture that lives only in people's heads doesn't survive agents. For most of my career, the real rules of a codebase weren't written down. People knew them. Senior engineers knew which layers could talk to which. They knew which dependencies were forbidden, which schemas were effectively frozen, and which shortcuts would create problems six months later. New engineers learned those rules the traditional way: break one, get caught in review, get the explanation, and eventually remember not to do it again. It wasn't perfect, but it mostly worked. What I didn't fully appreciate until I started working heavily with coding agents is how dependent that model is on tribal knowledge. Humans accumulate context over time. Agents don't. They don't remember the migration that went sideways three years ago. They weren't around when the team spent weeks untangling a dependency cycle. They don't know why a particular boundary exists. They only know what they can see. Which means if a rule isn't written down, from the agent's perspective, the rule doesn't exist. I've seen agents wire inner layers directly to outer layers. I've seen them introduce dependencies we intentionally avoided and extend contracts everyone on the team considered settled. The code often worked, which was the dangerous part. The problem wasn't correctness. The problem was architectural drift. That's when something clicked for me. Architecture can't remain folklore once agents start writing code. It has to become law. Not a convention. Not a suggestion. Not something a reviewer remembers at 6 PM on a Friday. A law. Written down, explicit, and enforceable. That's what I mean by a constitution. A Constitution Is Not Documentation The first mistake I made was treating the constitution like another documentation file. It isn't. Documentation explains how the system works today. A constitution defines what the system is allowed to become. Those sound similar, but they serve very different purposes. Package nam

Hello Dev Community! 👋 It is officially Day 48 of my unbroken full-stack engineering journey! Yesterday, I refactored my modular core patterns into MVC architecture. Today, I linked up a major functional extension inside the /model layer by introducing JavaScript Classes (OOP) to coordinate my local file operations and storage data patterns! Instead of writing loose object definitions, I stepped up my enterprise game by structuring a reusable class footprint that encapsulates data parameters and handles non-blocking file-system persistence asynchronously. 🧠 Key Learnings From Day 48 (OOP Modeling & File Systems) As clearly shown in my development workspace layout within "Screenshot (116).png" , modeling data with dedicated classes shifts your core structural logic from simple scripts into highly scalable engines: 1. The Model Data Blueprinter ( constructor ) I used the standard ES6 class framework inside home.js to structure an explicit data mold ( houseList ) with attributes tracking: houseName , price , location , rating , and photoUrl . This ensures every entry traveling through our server follows an identical structure. 2. Streamlining Async Persistence ( save() ) Rather than relying on globally declared floating arrays, my .save() blueprint method triggers an internal lookup to read existing data stacks asynchronously before safely using fs.writeFile() to serialize and flash mutated JSON rows into a local data asset ( homesdata.json ). 3. Static Decoupled Fetchers ( static fetchAll() ) I mastered using static methods. Since reading a data grid requires pulling records without creating an instance of a single house first, making fetchAll(callback) static allows our controllers to tap the hard disk records straight from the class reference layout: javascript // A conceptual look at my file-reading design today static fetchAll(callback) { const filePath = path.join(rootDir, 'data', 'homesdata.json'); fs.readFile(filePath, (err, data) => { if (err) { callback(JSON.

Dalla libreria embedded che ha invaso ogni dispositivo a un’implementazione moderna con concorrenza, async I/O e vector search: cosa cambia davvero per chi sviluppa app. Nel frontend e nel full‑stack capita spesso di parlare di database come servizi: Postgres gestito, cluster, repliche, connessioni, pooling, credenziali e una lunga lista di “cose che possono rompersi”. Ma esiste un’altra filosofia, più vicina all’idea di “dipendenza” che di “infrastruttura”: un motore SQL che vive dentro l’applicazione. Questa è la ragione per cui SQLite è ovunque. È una libreria, non un server. Legge e scrive su un singolo file su disco. Riduce drasticamente configurazione, porte, processi separati e complessità operativa. Ed è proprio questa semplicità a renderla una delle fondamenta silenziose dell’informatica moderna: la usi in browser, smartphone, desktop app, tool CLI, IoT… spesso senza nemmeno accorgertene. Ora immagina di riscrivere tutto da capo, in Rust, cercando di essere compatibile al 100% e allo stesso tempo più “moderna”. Sembra un’idea folle per definizione—finché non inizi a guardare ai limiti pratici che oggi emergono in molte applicazioni. Perché toccare SQLite, se funziona così bene? SQLite non è “il problema”. Anzi: è considerata estremamente robusta perché è conservativa, minimalista, e custodita con un rigore quasi maniacale. Il punto è un altro: il suo modello di sviluppo e manutenzione è atipico rispetto a quello che molti intendono per open source collaborativo . Il codice è disponibile e utilizzabile liberamente, ma l’evoluzione è guidata da pochissime persone e—di fatto—non segue la dinamica classica delle contribution esterne. Questa scelta ha un effetto collaterale positivo: riduce il rischio di regressioni introdotte da cambiamenti non coerenti con la visione del progetto. Ma ha anche un costo: se la tua azienda o il tuo prodotto hanno esigenze nuove (concorrenza più spinta, I/O non bloccante, funzionalità specifiche), “aspettare che arrivi upstream” n

Using Claude Fable 5 or Mythos 5 on Amazon Bedrock requires opting into provider_data_share, sending prompts and outputs to Anthropic for 30-day retention with human review. Previous Bedrock models kept inference data inside the AWS boundary. Three days after launch, Anthropic asked AWS to revoke access to both models citing US export control compliance. By Steef-Jan Wiggers

AWS recently introduced Amazon Cognito multi-region replication, which automatically replicates user identities and user pool configurations from a primary region to a secondary one. This enables applications to continue authenticating users from a replica region during outages, without requiring custom replication and failover mechanisms. By Renato Losio

We are building a VM that helps you simulate realistic production conditions, model latencies, different interleaving, user requests, and find bugs. Every non-deterministic property is turned into a knob you or a coding agent can control. We have helped teams perfectly reproduce support incidents and found bugs in some of the world's most well tested software (including a database) I you would like to try it out, I am giving away $100 in free credits to first 100 signups.