Search the archive
Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Results
9,381 matchesABSURD WEBSITE
ABSURD.website is an art project that launches a new, intentionally absurd product or service each month.
GAUNTLET
Gauntlet is a new programming language that transpiles to Go, aiming to improve upon Go's design while maintaining full compatibility with the Go ecosystem.
HEALTH AS A DEV
Maintaining good health as a software developer is important, as snacking on junk food during debugging and working late nights lead to decreased productivity and overall well-being.
👉 Read the guide.
Sponsored
IMPLEMENTING MULTI-TENANCY INTO A NEXT.JS SUPABASE APP
If you're building something team-based—think workspaces, orgs, or shared dashboards—you've probably bumped into the “how do I handle multi-tenancy?” question.
PROGRESSIVE JSON
Most apps send JSON all at once, forcing the client to wait until everything is ready, but what if we streamed it progressively, like how progressive JPEGs load from blurry to clear? A "Progressive JSON" approach is explored here that sends data breadth-first with placeholders, letting the client process what's ready and treat missing parts as Promises. This is similar to how React Server Components stream UI and use Suspense to control visual loading states.
VIRTUAL THREADS ATE MY MEMORY: A WEB CRAWLER'S TALE OF SPEED VS. MEMORY
Switching a web crawler from platform threads to Java Virtual Threads improves speed, but also causes an OutOfMemoryError due to a lack of built-in backpressure. Virtual Threads allows downloads to race ahead, but without limits, the system quickly floods memory with unprocessed tasks. The fix is to introduce explicit concurrency control (like semaphores) to balance speed with memory.
CONCURRENT WEB CRAWLING IN RUBY WITH ASYNC
This article shows how to transform a slow, synchronous Ruby web crawler into a fast concurrent one using the Async library. Crawling time is reduced from 3 seconds to half a second in the example by implementing non-blocking I/O with Fibers, connection reuse, and rate limiting.
WHAT'S WORKING FOR YC COMPANIES SINCE THE AI BOOM
WHEN FINE-TUNING ACTUALLY MAKES SENSE: A DEVELOPER'S GUIDE
Fine-tuning helps solve specific, high-impact problems that prompting alone can't, like inconsistent JSON output, bloated prompts, or models that ignore custom logic or tone. It's especially valuable for improving task-specific quality, reducing cost and latency, and making smaller models smarter through training on targeted examples.
Sponsored
KNOCK: INFRASTRUCTURE TO SEND PRODUCT MESSAGING
Building, managing, and scaling cross-channel notifications gets complicated fast. Knock [10] is a single API for managing messages across email, in-app, Slack, push, and SMS. Plus, a visual workflow builder and template editor anyone can use. All with observability, like you built it in-house. Send 10k notifications per month for free [10].
MCP SERVER IN BASH
`mcp-server-bash-sdk` is a lightweight implementation of a Model Context Protocol (MCP) server in Bash. It provides a zero-overhead alternative to other MCP server implementations by using Bash and `jq` for JSON processing.
CAP
Cap is a fast, lightweight, open-source CAPTCHA alternative that uses SHA-256 proof-of-work instead of tracking or fingerprinting. It has a tiny JavaScript widget, zero-dependency server, optional CLI, Docker-based standalone mode, and middleware for frameworks like Express, Hono, and Elysia.
ROOT SHELL ON CREDIT CARD TERMINAL
This dev reverse-engineered a Worldline Yomani XR credit card terminal and found it had surprisingly weak security despite hardware tamper protections. He discovered a root shell accessible via a debug port, even from outside the device without triggering tamper detection. However, the vulnerable Linux system appears isolated from the secure processor handling sensitive card data, which limits the potential damage.
HOW GENERATIVE ENGINE OPTIMIZATION (GEO) REWRITES THE RULES OF SEARCH
Generative Engine Optimization (GEO) is a new paradigm shifting from traditional SEO's focus on page rank to language models and how content is referenced in AI-generated answers. This new approach prioritizes well-organized and meaningful content that language models can easily extract and reproduce, requiring brands to optimize for model relevance rather than just search engine rankings.
ESLINT V9.0.0: A RETROSPECTIVE
ESLint v9.0.0 was released in April 2024 with a new configuration system, but its launch faced initial negative feedback from users due to numerous breaking changes, such as the combination of the new configuration system and rule API changes. While the ESLint team provided documentation, users found it overwhelming and preferred automated tooling for upgrades. The lessons learned include prioritizing smaller, more frequent releases, focusing on tooling over documentation, and improving error message context.
WHY ARE 2025/05/28 AND 2025-05-28 DIFFERENT DAYS IN JAVASCRIPT?
JavaScript interprets dates with the `/` separator as local time and dates with the `-` separator as UTC, leading to different date calculations due to browser parsing inconsistencies and historical standards.
PUBLISHED FOR THE FIRST TIME: THE PRINCETON INTERCAL COMPILER'S SOURCE CODE
The original source code for the influential esoteric programming language INTERCAL-72, created by Princeton undergraduates in 1972 and known for its intentionally difficult and parodic design, has been rediscovered, transcribed, and made publicly available.
CODEX CLI IS GOING NATIVE
OpenAI is rewriting Codex CLI in Rust to improve cross-platform stability, security, performance, and extensibility.
WHY DEEPSEEK IS CHEAP AT SCALE BUT EXPENSIVE TO RUN LOCALLY
DeepSeek-V3's architecture, like other Mixture-of-Experts models, benefits from large batch sizes to efficiently utilize GPUs, making it cost-effective at scale.
HOW NASA IS USING GRAPH TECHNOLOGY AND LLMS TO BUILD A PEOPLE KNOWLEDGE GRAPH
NASA is using graphs and LLMs to create a People Knowledge Graph. This graph-powered system is built on Memgraph and runs on AWS. It connects employees, projects, and skills to facilitate subject matter expert discovery, project similarity analysis, and real-time organizational insights. NASA ingests data from various sources, extracts skills using LLMs, and uses a RAG-based chatbot interface for querying the graph in natural language.
WHY REACT ERROR BOUNDARIES AREN'T JUST TRY/CATCH FOR COMPONENTS
React error boundaries are not simply `try/catch` blocks for components, but a declarative way to handle errors within the UI. Error boundaries, implemented as class components, catch errors during rendering, lifecycle methods, and constructors, displaying fallback UIs instead of crashing the app. However, they don't catch errors in event handlers or asynchronous code, which means that methods like `react-error-boundary` are needed to surface these errors.
HUMAN CODERS ARE STILL BETTER THAN LLMS
This dev had an experience fixing a complex bug in Redis using Vector Sets and compared their problem-solving abilities to those of Gemini 2.5 Pro. While the LLM provided some helpful suggestions, the dev ended up coming up with a more creative and efficient solution involving a fixed accumulator and murmur-128 hashing to detect corrupted links. Human creativity and out-of-the-box thinking are still better than LLMs for complex problem-solving.
MY WEBSITE IS UGLY BECAUSE I MADE IT
The "ugliness" of this dev's website is what makes it uniquely theirs. They designed the site with chaotic elements and unconventional CSS tricks to achieve a "used bookstore" feel, prioritizing personal expression over conventional aesthetics.
REDWOODSDK
RedwoodSDK is a React framework for building server-side web applications on Cloudflare. It has a standards-based router with middleware and interruptors for controlling requests and responses. It also provides built-in integration with Cloudflare services like D1, R2, and Workers, along with local emulation through Miniflare.
FORESIGHTJS
ForesightJS is a JavaScript library that improves website performance by predicting user intent through mouse movement analysis. It does data prefetching before clicks or hovers to reduce unnecessary prefetch requests and improve perceived loading speed.
ANNOUNCING ANGULAR V20
Angular v20 has been released. This release stabilizes reactivity APIs like effect and toSignal, improves debugging with Angular DevTools and Chrome DevTools integration, and improves the developer experience with style guide updates and language service support. New experimental APIs, such as resource streaming and httpResource, are introduced for managing asynchronous state.
WHY DID MICROSOFT-BACKED $1.3BN BUILDER.AI COLLAPSE? ACCUSED OF USING INDIAN CODERS FOR 'AI' WORK
Builder.ai was a startup valued at $1.3 billion and backed by Microsoft. It collapsed in May after being accused of misleading investors by overstating its AI capabilities and relying primarily on Indian coders behind the scenes. The company's business model, which promised AI-driven app development for non-engineers, was revealed to be largely a marketing gimmick, leading to inflated revenue projections and unsustainable expansion.
LARGE LANGUAGE MODELS CAN RUN TOOLS IN YOUR TERMINAL WITH LLM 0.26
LLM 0.26 introduces support for tools, allowing LLMs to access external functionalities via the LLM CLI tool and Python library. These tools are accessible through plugins or ad-hoc Python functions. This addition allows for the creation of "agents" by enabling tools in a loop.
GETTING A CEASE AND DESIST FROM WAFFLE HOUSE
During a hurricane in 2024, this dev reverse-engineered the Waffle House website to create an unofficial "Waffle House Index" map tracking restaurant closures as an indicator of disaster severity. The website unexpectedly gained traction, catching the attention of Waffle House's corporate team and even political commentators. Waffle House requested the author to cease and desist due to trademark infringement.
MAKING C AND PYTHON TALK TO EACH OTHER
An intro to calling Python functions from within a C program that covers setup, initialization, memory management, and argument passing.
MEMVID
Memvid is a Python library that improves AI memory by encoding text into videos for efficient storage and lightning-fast semantic search.
LEARNING C3
This dev documents their experience learning the C3 programming language, highlighting its features, strengths, and weaknesses, and comparing it to languages like C, Zig, and Rust while building a basic calculator to showcase its capabilities.
NOVA
Nova is a JavaScript and WebAssembly engine written in Rust that lets Rust projects execute JavaScript code.
PORTING TERRARIA AND CELESTE TO WEBASSEMBLY
Terraria and Celeste were successfully ported to WebAssembly (WASM), allowing them to run in web browsers. The process involved decompiling the C# code, configuring it for WASM, and resolving compatibility issues with the FNA engine, audio libraries, and threading. To load mods, they made the game binary load at runtime instead of compiling it with the project and used the Everest mod loader.
SLOW QUERY DEBUGGING WHAT BROKE OUR SAAS APP
A slow database query calculating user credit balances took down Placid.app when high-volume API requests triggered thousands of concurrent executions of an expensive query that should have been cached. Multiple workers tried to calculate the same user's balance simultaneously and the load ballooned. This caused a cascade of failures, timeouts, and blocked queries that eventually froze the system.
THE BEAUTY OF TANSTACK ROUTER
TanStack Router is a great routing solution for React apps due to its focus type safety and developer experience. The router has features like type-safe routing parameters, validated search params, and fine-grained subscriptions to prevent unnecessary re-renders. It supports both code-based and file-based routing and has integrated support for React Suspense.
HTAP IS DEAD
Database architectures have evolved from the unified systems of the 1970s to the specialized OLTP and OLAP systems that emerged due to conflicting workload demands. HTAP (Hybrid Transactional/Analytical Processing) became popular in the 2010s to bridge the gap between transactional and analytical systems. However, HTAP as a single database failed to gain widespread adoption due to the dominance of cloud data warehouses, faster hardware negating the need for distributed OLTP and organizational silos.
AS A DEVELOPER, MY MOST IMPORTANT TOOLS ARE A PEN AND A NOTEBOOK
A pen and notebook are better for problem-solving and creative thinking than a computer. Stepping away from the computer allows for more thoughtful design, bug fixing, and understanding of code flow. Writing and sketching turn abstract ideas into tangible artifacts, revealing gaps in knowledge.
Sponsored
LEARN HOW TO QUICKLY MONETIZE YOUR SAAS WITH SUBSCRIPTIONS POWERED BY CLERK BILLING
Subscription plans boost SaaS revenue but are hard to build from scratch. Clerk Billing simplifies it—offering a polished UI and built-in tools to manage plans and gate premium features without custom logic. This article [10] shows how Clerk Billing works and how to implement it in your app.
ZUSTAND EXPO DEVTOOLS
Zustand Expo DevTools is a DevTools plugin that integrates Zustand state debugging into the Expo DevTools environment. It has real-time state inspection, action tracking, and time-travel debugging for Zustand stores within Expo projects.
FIREDBG
FireDBG is a time travel visual debugger for Rust programs. It has a VS Code extension.
DEFUDDLE
Defuddle is a JavaScript library that extracts the main content from web pages by removing clutter, standardizing HTML, and providing a clean output suitable for HTML-to-Markdown conversion.
HOW CODE REVIEWS ARE CHANGING WITH AI
While AI coding tools get most attention, AI code reviews are having more immediate impact by automating routine checks like coding conventions and test coverage. AI will become better at reviewing existing code before mastering autonomous creation of complex systems. The key shift is moving from "approve first" to "possibly revert later" workflows.
GO CRYPTOGRAPHY SECURITY AUDIT
The core Go team contracted an external security firm to audit Go's core cryptography packages, resulting in only one low-severity finding. The results validate Go's strong cryptographic security practices and the team also explains other security enhancements coming to Go.
LLM CODEGEN GO BRRR – PARALLELIZATION WITH GIT WORKTREES AND TMUX
AI coding agents can be parallelized using Git worktrees to improve throughput and control over a codebase. uzi is a CLI tool designed to automate the orchestration of multiple AI agent worktrees. It makes tasks like prompting, executing commands, and managing agents easier.
A THOUGHT ON JAVASCRIPT "PROOF OF WORK" ANTI-SCRAPER SYSTEMS
JavaScript-based "proof of work" systems, used to deter LLM scrapers, pose a dilemma because while scrapers could dedicate CPU time to solve these challenges, doing so exposes them to potential exploitation by malicious actors injecting other JavaScript for purposes like cryptocurrency mining or CPU resource draining.
OSV
OSV is a high-performance CSV parser for Ruby supporting both hash-based and array-based row formats with various configuration options.
BEWARE OF FAST-MATH
Using "fast-math" compiler flags can speed up math computations at the cost of accuracy and potentially introduce subtle and hard-to-debug errors by violating IEEE 754 floating-point standards.