Search the archive

RESERVE FIRST

Exception safety bugs in memory management code can be prevented by reserving memory before modifying data structures to avoid data corruption during allocation failures.

AN INTERACTIVE GUIDE TO SVG PATHS

This is an interactive guide to SVG paths, focusing on the path element and its various commands for drawing shapes. It covers basic commands like Move and Line, then goes into Bézier curves (quadratic and cubic) and the complex Arc command, explaining each parameter with interactive examples. The guide also covers extra details like closing paths, relative commands, and chained curves.

CREATIVE WAYS TO FUND OPEN SOURCE PROJECTS

Open source projects are often underfunded, leading to maintainer burnout, but innovative revenue models are starting to show up. The Wix Toolset uses an "Open Source Maintenance Fee," which charges users for issue reporting and pull requests based on company size, providing income for maintainers. Another example is Astral's uv package manager, which offers a free version and a paid, enterprise-focused registry called pyx with more features.

YOU SHOULD ADD DEBUG VIEWS TO YOUR DB

Creating debug views in your database can simplify complex queries and improve debugging efficiency. These views predefine joins and include relevant information, making it easier to find specific data based on various criteria. By making these views, it removes the need to repeatedly write lengthy queries.

AWS CEO SAYS USING AI TO REPLACE JUNIOR STAFF IS 'DUMBEST THING I'VE EVER HEARD'

AWS CEO Matt Garman dismissed the idea of replacing junior staff with AI as the dumbest thing he's ever heard, arguing that these employees are inexpensive and AI-savvy. Instead, it's important to continue to hire and train junior developers to create a future workforce with critical thinking and problem-solving skills.

THE UNBEARABLE SLOWNESS OF AI CODING

When first coding with Claude Code, this dev had a speed boost in his coding tasks. However, as his app grew, the review and debugging process became painstakingly slow, despite using multiple instances of Claude Code simultaneously. He found himself spending a lot of time as a QA engineer, fixing issues generated by Claude, including hallucinations about library features.

BUILDING AI PRODUCTS IN THE PROBABILISTIC ERA

The industry is experiencing a shift from deterministic software, where inputs predictably map to outputs, to a probabilistic era driven by AI. Traditional methods of building and managing software based on pre-defined features and measurable funnels are becoming obsolete as AI models exhibit emergent, unpredictable behaviors. To succeed, companies must adopt an empirical, scientific approach, focusing on managing uncertainty and understanding user behavior through user trajectories.

SIM

Sim is an open-source platform for building and deploying AI agent workflows. It supports local AI models using Ollama.

EXPO SDK V54 BETA

Expo SDK 54 beta is now available, with React Native 0.81, React 19.1.0, and new features like iOS 26 Liquid Glass icons. A big improvement is precompiled React Native for iOS, reducing build times by a lot. This release also targets Android 16, enabling edge-to-edge display by default and introducing predictive back gestures.

MY AI HAD ALREADY FIXED THE CODE BEFORE I SAW IT

Compounding engineering is a development approach where AI tools learn from every code review, bug fix, and pull request, building systems that improve themselves. By codifying preferences and architectural decisions, engineers can train AI agents like Claude to automate tasks and enforce consistent standards. This method turns production errors into permanent fixes.

GOOGLE SCORES SIX-YEAR META CLOUD DEAL WORTH OVER $10 BILLION

Google has secured a big six-year cloud contract with Meta worth over $10 billion. Meta has usually been reliant on Amazon Web Services and Microsoft Azure, but is expanding its cloud infrastructure to support its growing artificial intelligence workloads.

GPT-5: IT JUST DOES STUFF

GPT-5 automatically selects the right model for each task and proactively suggests solutions rather than just answering questions. It's possible to give it a vague coding prompt and receive a fully functional app with features never requested. This shows a shift in AI becoming more proactive over time.

MUCH ADO ABOUT VIBE CODING

A writer at Wired was able to "work/vibe-code" for Notion for two days, showing the potential for AI to supercharge coding productivity.

GRUG DESIGN

A simple guide to good design, which is simple, clear, functional, and doesn't show unnecessary options.

DEEPSEEK-V3.1 RELEASE

DeepSeek-V3.1 is a hybrid inference model with faster thinking capabilities, better agent skills through post-training, and API updates, including expanded context length, Anthropic API support, and function calling.

STATE OF CSS 2025

CSS is evolving from a simple styling language into a modern, powerful tool for layouts, motion, and accessibility, with growing developer confidence and satisfaction.

INTRODUCTION TO AT PROTOCOL

The Authenticated Transfer Protocol (ATProto) is the underlying technology of Bluesky. This article describes the core components of ATProto, including records, blobs, lexicons, DIDs, handles, and AT URIs, outlining how they structure data and identify users. It also goes over the server architecture, focusing on the roles of PDSes (Personal Data Servers), relays, AppViews, labellers, and feed generators in processing and distributing data across the network.

HOW I BUILT A FULL-STACK REACT FRAMEWORK 4X FASTER THAN NEXT.JS WITH 4X MORE THROUGHPUT

Rari is a full-stack React framework built on a custom Rust runtime with V8. Its architecture consists of three layers: a Rust core runtime for executing React Server Components, smart Vite integration for RSC-aware transformations and dependency management, and true server-side rendering with streaming support. It uses file-based routing with automatic code splitting and prefetching, optimizing bundle sizes and performance.

STATE MANAGEMENT ARCHITECTURE IN NX

Nx and NgRx Signal Store provide a structured approach to Angular application state management. The architecture in this article goes over a library-based structure with clear separation of concerns between data access, state management, UI, and features. It enforces dependency rules to prevent architectural drift and allows benefits like build optimization, testing isolation, and team scalability.

AWS IN 2025: THE STUFF YOU THINK YOU KNOW THAT'S NOW WRONG

Security configurations can now be altered live in EC2. Instance reliability has improved a lot, and Spot Instance pricing is less volatile. S3 now has read-after-write consistency, deprecated ACLs, and simpler Glacier integration. AWS networking has been updated with better VPC options, quicker CloudFront updates, and advancements in load balancing.

WHY I'M ALL-IN ON ZEN BROWSER

This dev switched to Zen Browser after being disappointed with Arc's shift to an AI-focused browser called Dia, which lacked the power-user features he valued. He found Dia's privacy concerns and increased competition unappealing. While Firefox is a reliable option, Zen Browser offers a more streamlined, Arc-inspired UI with customizable features built on top of the Firefox framework.

big! Registration closes soon

ZEDLESS

Zedless is a fork of the Zed editor with a focus on privacy and local-first functionality. It removes reliance on cloud services, removes telemetry, and prioritizes user-configurable infrastructure with disabled-by-default network features.

HERB

Herb is an ecosystem of developer tools designed to improve the experience of working with HTML+ERB templates through accurate parsing. The core of Herb is a fast, portable, HTML-aware ERB parser written in C that generates a detailed syntax tree. This ecosystem includes tools such as a language server, formatter, and linter, integrating into editors and CI pipelines.

TIDEWAVE WEB: IN-BROWSER CODING AGENT FOR RAILS AND PHOENIX

Tidewave Web is a new in-browser coding agent for Rails and Phoenix web applications that runs directly in the development environment. Unlike traditional coding agents, it provides shared context by understanding the UI state, framework, and accessing developer tools like the database and logs. This removes the need for constant back-and-forth between the developer and the agent.

DRIVING AI ADOPTION AT DROPBOX: A CONVERSATION WITH CTO ALI DASDAN

Dropbox had over 90% AI adoption among its engineering teams in 2025 due to strong leadership alignment and a culture of knowledge-sharing. The company uses a hybrid approach of both external AI tools like GitHub Copilot and internally-built custom solutions to address specific needs in coding, code reviews, testing, and debugging.

“AI FIRST” AND THE BUS FACTOR OF 0

The "Bus Factor" measures the risk of losing knowledge if key team members leave. With the rise of "AI first" approaches using tools like ChatGPT, many people are becoming comfortable with a Bus Factor of zero, where no one understands the AI-generated code. This reliance on LLMs for code generation makes maintaining and debugging software extremely difficult.

SEQUOIA BACKS ZED'S VISION FOR COLLABORATIVE CODING

Zed is a coding platform that raised $32M in Series B funding led by Sequoia Capital. Its goal is to create an IDE where code discussions are directly linked to the code, addressing the limitations of current snapshot-based systems like Git. Zed is developing DeltaDB, a new operation-based version control system that tracks every edit, allowing for real-time collaboration with both humans and AI agents.

CLAUDIA

Claudia is a desktop application that provides a GUI for managing Claude Code sessions, creating custom agents, tracking usage, and offering project management features.

AGENTS.MD

AGENTS.md is a simple, open-source Markdown file used by AI coding agents to understand project context, setup, coding style, testing, and PR guidelines.

COPILOT BROKE YOUR AUDIT LOG, BUT MICROSOFT WON'T TELL YOU

Microsoft's Copilot AI tool had a vulnerability that allowed users to access files without leaving a trace in the audit log, but despite fixing the issue, Microsoft chose not to disclose it to customers.

NESTABLE

Nestable is a tool that allows users to create infinitely nested pages/canvases, search across content, deep link to specific elements, and import/export their work.

CONTEXT ENGINEERING CASE STUDIES: ETSY-SPECIFIC QUESTION ANSWERING

Etsy explored using prompts with LLMs to answer onboarding questions related to internal and external policies. It tested the system on Travel & Entertainment policies and Etsy seller forums, but ran into hallucinations and incorrect answers. While prompt engineering can help fix these errors through specific instructions and context, complex reasoning with LLMs still has a lot of challenges. Asking the LLM to identify specific source snippets is a good way to flag potential hallucinations.

WHY SEMANTIC LAYERS MATTER — AND HOW TO BUILD ONE WITH DUCKDB

Semantic layers provide a unified source of truth for business metrics across various tools. They are beneficial when you have multiple consumers of data, complex business logic, and the need for consistent metrics. This article shows a simple semantic layer using YAML, Python, Ibis, and DuckDB to query NYC taxi data.

WHAT LEARNING REACT WON'T TEACH YOU: IMAGE FORMATS

Many developers neglect choosing the right image format during development, leading to unnecessarily large file sizes and slower website performance. JPEG is good for complex images and photographs, thanks to its lossy compression, and PNG is great for simpler graphics with solid colors, sharp edges, and transparency.

AI IS A JUNIOR DEV AND NEEDS A LEAD

AI coding assistants are like junior developers. They need clear guidelines, detailed scenarios, and precise prompts to create production-ready code.

Schedule a demo to learn more

NEXT.JS 15.15

Next.js 15.5 introduces beta Turbopack builds for faster performance and stable Node.js middleware. The release also includes TypeScript improvements for better type safety in routing and deprecates the `next lint` command in favor of ESLint or Biome.

EPICENTER

Epicenter is an open-source ecosystem of local-first apps focused on data ownership, allowing users to store their data in plain text and SQLite files. The project aims to replace siloed apps with interoperable alternatives. Current available tools include Whispering, a desktop transcription app, and epicenter.sh, a local-first assistant.

REUI

ReUI is an open-source React component library built with Tailwind CSS, Motion, and other technologies, with copy-and-paste components designed for customization, AI-optimization, and production use.

HOW WE EXPLOITED CODERABBIT: FROM A SIMPLE PR TO RCE AND WRITE ACCESS ON 1M REPOSITORIES

A vulnerability in CodeRabbit, an AI code review tool, allowed for remote code execution (RCE) on its production servers. By exploiting Rubocop, a static analysis tool integrated with CodeRabbit, the Kudelski Security team was able to inject malicious code and exfiltrate sensitive environment variables, including API keys and secrets. This access granted them potential read and write access to 1 million code repositories, including private ones, through the leaked CodeRabbit GitHub app private key. CodeRabbit quickly fixed the vulnerability by disabling Rubocop, rotating secrets, and implementing stronger security measures.

HOW TO THINK ABOUT GPUS

This article explains the architecture of GPUs, focusing on Streaming Multiprocessors (SMs), CUDA Cores, Tensor Cores, and memory hierarchy, and summarizes the specifications of different GPU generations. It then compares GPUs and TPUs at the chip level and covers GPU networking, detailing intra-node and cross-node collectives and their performance implications.

WORKING OFFLINE IN NOTION—EVERYTHING YOU NEED TO KNOW

Notion now has an offline mode to view, edit, and create pages on desktop and mobile.

THE STATE OF MCP SECURITY: HOW AI PLUGINS CREATE SILENT EXPLOITS

The combination of multiple MCP plugins in AI agents creates often overlooked security risks, allowing for exploits like code execution from untrusted inputs such as emails or web content.

ASK HN: HAVE ANY SUCCESSFUL STARTUPS BEEN MADE BY 'VIBE CODING'? (HACKER NEWS THREAD)

Vibe coding has been great for small, focused apps and fast prototyping, but it isn't as great when the codebase grows large.

See AI testing that actually work 👀