Search the archive

AMAZON EC2 M9G INSTANCES

Amazon EC2 M9g instances offer the best price performance for general-purpose workloads.

TELL US WHAT YOU REALLY, REALLY… DO NOT WANT TO SPEND TIME WORKING ON

A survey by Stack Overflow explored developer frustrations amidst increasing AI adoption, finding that 84% currently use or plan to use AI, mainly for coding, searching, and debugging. However, developer problems aren't mainly from coding tasks themselves or AI's shortcomings, but rather from a lack of adequate and accessible documentation. Tasks like learning new codebases, deployments, and handling support tickets become frustrating due to fragmented and bad documentation efforts.

WHAT I LEARNED BUILDING A VECTOR DATABASE ON OBJECT STORAGE

Building something is a great way to learn. This developer rebuilt Turbopuffer, a vector database built entirely on object storage, to understand the tradeoffs made in its development. The post follows along as they figure out how to build a vector database from first principles.

automation that works.

INCOMPLETE LIST OF MISTAKES IN THE DESIGN OF CSS

VOXCSS

VoxCSS is a CSS voxel engine that renders customizable, interactive 3D grids of HTML cuboids in the DOM.

WIREBROWSER

Wirebrowser is a debugging, interception, and memory-inspection toolkit powered by the Chrome DevTools Protocol that unifies network manipulation, API testing, automation scripting, and deep JavaScript memory inspection into one interface.

THE (SUCCESSFUL) END OF THE KERNEL RUST EXPERIMENT

The Rust experiment in the Linux kernel has concluded, as Rust is now considered a core and permanent part of the kernel.

WHY FROZEN TEST FIXTURES ARE A PROBLEM ON LARGE PROJECTS AND HOW TO AVOID THEM

Test fixtures often become "frozen" in large software projects because modifications risk breaking thousands of unrelated tests that make implicit assumptions about the data. The recommended solution is to write highly focused tests that assert only the specific property they intend to verify, rather than comparing entire data structures.

BACKGROUND CODING AGENTS: PREDICTABLE RESULTS THROUGH STRONG FEEDBACK LOOPS

Spotify's background coding agents use verification loops to stay on track. Independent verifiers, like Maven, build systems and tests that run automatically and give incremental feedback. An LLM judge also vetoes PRs when agents get too creative and drift from their prompts.

OWN A GRAPH

If you're working in tech, you should "own a graph" as a critical unit of responsibility. This practice helps focus on and track progress for important, multi-quarter problems, making sure your work is directed towards meaningful outcomes. Graphs are the most powerful tool for concisely communicating impact.

IF YOU'RE GOING TO VIBE CODE, WHY NOT DO IT IN C?

that actually match your needs. Get the guide

INTRODUCING: DEVSTRAL 2 AND MISTRAL VIBE CLI

Mistral AI has launched Devstral 2 and Devstral Small 2, its next-generation open-source agentic coding models, along with the Mistral Vibe CLI for end-to-end code automation. Devstral 2 (123B) is a state-of-the-art model establishing new benchmarks for open-weight code agents. The Mistral Vibe CLI provides an open-source command-line interface that uses Devstral models to autonomously explore, modify, and execute changes across entire codebases.

DURABLE STREAMS

Durable Streams is an open protocol that provides real-time sync to client applications. It supports catch-up reads and live tailing. The protocol works anywhere HTTP works, and it is built on standard HTTP with no custom protocols. Durable Streams is built for production scale. It complements backend streaming systems by reliably streaming data to client applications.

STREAMDOWN

Streamdown is a drop-in replacement for `react-markdown` specifically engineered to handle the unique challenges of streaming Markdown content from AI models. It fixes issues like incomplete or unterminated Markdown blocks. Streamdown supports GitHub Flavored Markdown, LaTeX math rendering, Mermaid diagrams, and Shiki code highlighting.

FATE

Fate is a modern data client for React and tRPC that simplifies data fetching and state management. It combines features like view composition, normalized caching, data masking, and Async React capabilities with tRPC's type safety.

MAKING MACOS BEARABLE

Stock macOS can be inefficient and disruptive. To regain a frictionless workflow, this dev re-engineered macOS to reduce visual searching and manual interaction. This was done by using Aerospace for deterministic window management, Tmux for streamlined development, modal editing for text manipulation, and tools like Vimium and Homerow for keyboard-driven navigation across the browser and OS.

WRITING OUR OWN CHEAT ENGINE: INTRODUCTION

This post introduces the goal of writing a small Cheat Engine clone in Rust to understand how the original tool functions by replicating its tutorial steps. The first step focuses on learning to enumerate active processes, opening them, and retrieving their module names programmatically. This is done through Windows API functions like `EnumProcesses`, `OpenProcess`, `EnumProcessModules`, and `GetModuleBaseNameA` while managing Rust's `unsafe` code.

2 YEARS WITH SHAPE-UP, AND WHY WE SWITCHED BACK

After two years, this dev's company switched from the Shape-Up software development methodology because its rigid 6-week cycles and project size mismatch no longer suited their needs for flexibility and a clearer product direction.

DJANGO: WHAT'S NEW IN 6.0

Django 6.0 introduces a host of new features, including template partials, a built-in tasks framework, Content Security Policy (CSP) support, and a modernized email API.

THE NEXT ERA OF REACT HAS ARRIVED: HERE'S WHAT YOU NEED TO KNOW

React's new “Async React” combines five primitives to handle coordination automatically: `useTransition` tracks async work, ‘useOptimistic` makes UI instant (checkboxes toggle before server responds), Suspense handles loading, use() replaces `useEffect` for data fetching, and `useDeferredValue` keeps inputs responsive.

DOWNDETECTOR AND THE REAL COST OF NO UPSTREAM DEPENDENCIES

Downdetector is a service built to monitor outages, but it experienced downtime itself during a Cloudflare outage.

SO YOU WANT TO SPEAK AT SOFTWARE CONFERENCES?

The journey to becoming a software conference speaker begins by developing and refining unique talks at local user groups to practice, then progressing to community conferences to be seen and build a professional network.

developers down. The latest Docusign for Developers release

simplifies agreement processes, surfaces insights, and helps developers deploy faster.

REACT ROUTER'S TAKE ON REACT SERVER COMPONENTS

React Router is introducing experimental support for React Server Components (RSC), which improves performance by allowing UI to be rendered on the server. Developers can now return pre-rendered UI directly from loaders or make entire routes server components, reducing client-side hydration and payload size. The integration also supports server functions for component-level data mutations and client components for interactive elements.

93% FASTER NEXT.JS IN (YOUR) KUBERNETES

Watt made Next.js 93% faster in K8s (182ms → 11.6ms) by ditching PM2's ~30% IPC tax and letting the Linux kernel distribute connections directly via SO_REUSEPORT. Traditional approaches aren't great because PM2 adds coordination overhead, and single-CPU pods can't self-heal when their event loop gets blocked.

HOW EXCHANGES TURN ORDER BOOKS INTO DISTRIBUTED LOGS

Order books are distributed logs of market events. This architecture guarantees fairness through deterministic ordering. Modern exchanges behave like ultra-low-latency log processors. The sequencer defines the timeline, the matching engine reduces that timeline into a book, and replication keeps the log durable without slowing the fast path.

HAS THE COST OF BUILDING SOFTWARE JUST DROPPED 90%?

Agentic coding with AI is poised to drop the cost of building software by up to 90%. This reduction is due to AI agents automating traditionally labor-intensive tasks like testing, API development, and infrastructure setup, reducing implementation time and project overhead. This shift is expected to increase latent demand for software.

DEPRECATIONS VIA WARNINGS DON'T WORK FOR PYTHON LIBRARIES

urllib3 v2.6.0 contained removals for several APIs that have been known to be problematic since 2019 and have been deprecated since 2022. The deprecations were marked in the documentation and changelog, but the team still received feedback that the removal was unexpected and was breaking dependent libraries. This incident makes it clear that DeprecationWarning in its current state doesn't work for deprecating APIs, at least for Python libraries.

Start with a free YugabyteDB account

VIBESDK

Cloudflare VibeSDK is an open-source, full-stack AI web application generator that allows users to create and deploy applications using natural language prompts. It has AI code generation with error correction, live previews in sandboxed containers, interactive chat, and one-click deployment to Cloudflare Workers for Platforms.

MERMAID TO MOTION

Fanfa is a tool that turns Mermaid into a cinematic SVG animation. Users can choose from several themes, edit arrow types, and control the animation speed. They can then export the diagram as a PNG, SVG, or PDF document. A live, editable example is available on the site.

THE HIGHEST QUALITY CODEBASE

This developer ran an experiment where they looped an agent over a codebase for 36 hours to see what it did. The experiment resulted in more code to be maintained, most of it largely useless. The agent added tons of tests, but removed the tests that mattered the most. The resulting app still worked, but with a few new bugs.

A BRIEF HISTORY OF SPECIFIERS AND PROTOCOLS

Node.js introduced the node: protocol in 2021. Deno intro dude https imports in 2018, and then the npm: protocol in 2022. Deno then introduced package.json support in 2023 and the jsr: protocol in 2024. Also in 2024, Deno started moving away from HTTP imports. The current landscape is a messy state of affairs.

GOAL SETTING FOR PRODUCTIVE ENGINEERS

Engineers often struggle to achieve long-term career goals due to chaotic work environments. To combat this, this article proposes two goal-setting approaches: top-down goals for clear outcomes and bottom-up "tiny experiments" for ambiguous situations.

compromise to quality. Check out TestNitro by QualityLogic

IBM TO ACQUIRE CONFLUENT

IBM is acquiring Confluent in an all-cash deal for $31.00 per share to unify enterprise data and scale AI across organizations by using Confluent's real-time data streaming platform.

SCALA 3 SLOWED US DOWN?

Libraries can behave very differently between Scala versions, especially when they rely on meta-programming.

MICROSERVICES SHOULD FORM A POLYTREE

To avoid the common complexities and failures of microservices, dependency structures should form a "polytree,” a directed acyclic graph whose underlying undirected graph is a tree, to guarantee clear ownership, predictable behavior, and easier system evolution.

* Unlimited parallel test runs

* 24-hour maintenance and on-demand test creation * Zero flakes, guaranteed

Schedule a demo to learn more

TANSTACK'S OPEN. AI. SDK

TanStack AI launched this week in alpha. It features a clean architecture, deep type safety, tight ecosystem integration, and few assumptions about user platforms. While it is less production-ready than Vercel's AI SDK, TanStack AI is the more interesting bet for the future due to its architectural decisions. This post looks at both TanStack AI and Vercel AI SDK and compares their current features.