LodeHQSubscribe →

Rust project goals update, C++26 reflection, etcd 3.7.0 beta

Dev · 2026-05-26

Languages & Frameworks
New to the web platform in April2 MIN

April's stable browser releases (Chrome 147, Firefox 150) bring the contrast-color() CSS function to Baseline, ariaNotify(), element-scoped view transitions, and the CSS border-shape property. Betas preview container queries, lazy-loaded media, and more.

Time is a construct but it can still break your software30 MIN

Stack Overflow podcast discusses the long-awaited Temporal proposal, a TC39 standard replacing JavaScript's notoriously broken Date object. Host Ryan Donovan and Boa creator Jason Williams explore why date/time handling is so hard and how Temporal finally fixes it after nine years in development.

Think Like the JavaScript Engine1 MIN

A new 5-hour freeCodeCamp course by Sumit Saha teaches developers to think like the JavaScript engine. It covers scopes, closures, execution context, prototypes, event propagation, and performance optimization, going beyond syntax to explain how the engine actually compiles and runs code.

Rust project goals update, April 2026 (end of 2025H2)59 MIN

The 2025H2 Rust Project Goal period concluded with 41 goals pursued, 13 designated as flagship. Curated outcomes cover Pin ergonomics, faster compilation, higher-level Rust, dormant traits, and more. Many goals continue into the 2026 period.

Compile-Time Map and Compile-Time Mutable Variable with C++26 Reflection18 MIN

This article explores C++26 reflection features for compile-time maps and mutable variables. Building on the P2996R13 ticket counter, it demonstrates stateful metaprogramming using substitute and define_aggregate. These techniques enable key-value stores evaluated entirely at compile time.

Tools & Platforms
Announcing etcd 3.7.0-beta.02 MIN

etcd v3.7.0-beta.0 introduces RangeStream for streaming large result sets, removes legacy v2store components, and marks v3.4 EOL. This beta includes bbolt v1.5.0 and raft v3.7.0.

Kubernetes v1.36: ハル (Haru) released30 MIN

Kubernetes v1.36 (Haru) ships 70 enhancements: user namespaces, declarative validation, and PSI metrics reach GA, alongside 25 new beta and 25 alpha features. This release strengthens security, validation workflows, and observability for the container orchestration platform.

Launching S3 Files: making S3 buckets accessible as file systems8 MIN

AWS launched S3 Files, making S3 buckets accessible as high-performance file systems via NFS v4.1+ on EC2, ECS, EKS, and Lambda. It eliminates the tradeoff between S3's cost/durability and file system interactivity, achieving ~1ms latencies with automatic synchronization.

Introducing the pkg.go.dev API5 MIN

The official pkg.go.dev API launches, giving developers programmatic access to Go module metadata, package info, versions, and vulnerability data. Built with a stateless GET-only architecture, it eliminates the need for web scraping in tooling and IDE integrations. The API supports AI-assisted coding by providing high-fidelity context about the Go ecosystem.

How we replaced Ingress-NGINX at Stack Overflow9 MIN

Stack Overflow details their migration from Ingress-NGINX to a Gateway API-based solution after the project's retirement announcement. They evaluated Traefik, Istio, and NGINX Gateway Fabric, using Claude to analyze existing ingress patterns. The post covers their criteria, testing process, and architectural decisions for new Kubernetes traffic routing.

Google Cloud deletes Australian trading fund's infra3 MIN

Google Cloud accidentally deleted UniSuper's entire subscription, wiping both primary and replicated data. The $124B Australian fund avoided total data loss only thanks to a third-party backup. GCP CEO Thomas Kurian publicly took the blame in a rare admission.

Datasette Agent - extensible AI assistant for Datasette3 MIN

Simon Willison announced the first release of Datasette Agent, an extensible AI assistant for Datasette built on his LLM library. It provides a conversational interface for querying data and generating charts via plugins. The demo runs on Gemini 3.1 Flash-Lite and supports local models too.

Security Advisory for Cargo (CVE-2026-5223)1 MIN

CVE-2026-5223: Cargo incorrectly handled symlinks inside crate tarballs from third-party registries, allowing malicious crates to override another crate's source code. The fix ships in Rust 1.96.0 on May 28th, 2026. Crates.io users are unaffected.

AI-Assisted Development
All the news from the Google I/O 2026 Developer keynote4 MIN

Google I/O 2026 unveiled Gemini 3.5 and Antigravity 2.0, an agent-first development platform with CLI and SDK controls. New Android CLI tools, Android Bench for LLM evaluation, and agent-powered migration tools were announced. The keynote marked a shift from assistive AI to autonomous agents navigating workflows independently.

Gemini 3.5 Flash: more expensive, but Google plan to use it for everything2 MIN

Google shipped Gemini 3.5 Flash at I/O, skipping preview and going straight to GA. The model costs 3x its predecessor but is being rolled out across Google's consumer products. A clear sign AI labs are testing price tolerance.

Bliki: Vibe Coding3 MIN

Martin Fowler explores "vibe coding", building software by prompting an LLM without ever looking at the generated code. Coined by Andrej Karpathy, it enables non-programmers to build apps but introduces maintainability, correctness, and security risks. Fowler distinguishes it from agentic programming, where developers still review the code.

Engineering Practice
Coding agents are giving everyone decision fatigue9 MIN

Coding agents are shifting engineers' work from writing code to reviewing AI output and crafting prompts, intensifying cognitive load. The bottleneck has moved to judgment and code review, making workdays denser rather than easier. The post explores how to reconfigure development to lighten this new burden.

The Pulse: Forward deployed engineering heats up again9 MIN

Forward Deployed Engineer (FDE) demand is surging at Google, OpenAI, and Anthropic. Google has streamlined hiring to as few as two interviews, while both OpenAI and Anthropic are creating standalone FDE consulting companies backed by billions in outside investment.

What ClickUp's mass layoff tells us about the future of work2 MIN

ClickUp laid off 22% of staff while deploying 3,000 internal AI agents, with CEO Zeb Evans calling it an AI embrace, not cost-cutting. Remaining employees will direct AI agents and could earn million-dollar salaries. The move signals a radical shift in how tech companies view AI-driven workforce restructuring.

Interviews Aren't About You (Sorry)3 MIN

Interviews aren't about your résumé, they're about solving the hiring manager's problem. The post urges candidates to ask what prompted the opening, then connect their experience to that specific pain. Be a detective, not a performer.

Other
From latency to instant: Modernizing GitHub Issues navigation performance14 MIN

GitHub Issues rewrote navigation to feel instant using client-side caching with IndexedDB, smart prefetching, and a service worker. Their approach shifts work to the client, rendering from local data first, then revalidating in the background. The post details tradeoffs and patterns applicable to any data-heavy web app.

What is Code10 MIN

Unmesh Joshi argues code serves two purposes: instructing machines and communicating intent to future humans. As LLMs commoditize the first, the conceptual model embedded in code, its vocabulary and design, becomes the enduring value that developers must cultivate.

Bliki: Architecture Decision Record4 MIN

Martin Fowler's updated bliki entry explains how ADRs capture single decisions with context and rationale, stored alongside code using inverted-pyramid writing. Accepted records are never modified, only superseded with a link. This fresh take reinforces ADRs as a lightweight tool for clarifying team thinking and preserving architectural intent.

Principles of Mechanical Sympathy9 MIN

Martin Fowler's article distills mechanical sympathy into practical principles like predictable memory access, cache-line awareness, and the single-writer rule. These techniques, borrowed from racing and popularized by Martin Thompson's LMAX architecture, help developers write software that leverages modern hardware. The principles apply broadly, from AI inference to distributed data platforms.

Get Dev in your inbox, every issue.
Subscribe free
Privacy · Terms · About · Contact
© 2026 LodeHQ