Best of The New Stack — December 2025

  1. 1
    Article
    Avatar of newstackThe New Stack·23w

    ES2026 Solves JavaScript Headaches With Dates, Math and Modules

    ECMAScript 2026 is set to introduce significant improvements to JavaScript, including Math.sumPrecise for more accurate floating-point calculations, Uint8Array base64 encoding methods, and enhanced JSON parsing with source text access. The release will improve internationalization with Intl.Locale enhancements for handling regional date/time formats and weekend variations. Major additions include the Temporal API to replace JavaScript's problematic Date object with better time zone handling and calendar support, explicit resource management with using blocks for automatic cleanup, and import defer for optimizing module loading performance. Additional features include Error.isError for reliable error checking, Iterator.concat for chaining iterators, Array.fromAsync for async iterables, and Map/WeakMap upsert operations.

  2. 2
    Article
    Avatar of newstackThe New Stack·21w

    Web Components Are the Comeback Nobody Saw Coming

    Web components are experiencing a resurgence as developers tire of framework complexity and bloated bundles. Native browser APIs like custom elements and Shadow DOM offer framework-agnostic, lightweight alternatives that work across React, Vue, or vanilla JavaScript without modification. Major browser support is now stable, and libraries like Lit have improved developer experience. The interoperability advantage makes web components ideal for design systems and microfrontends, allowing teams to build reusable UI components once and deploy them anywhere. As performance budgets tighten and framework fatigue sets in, the native web platform capabilities are proving sufficient for most modern applications.

  3. 3
    Article
    Avatar of newstackThe New Stack·20w

    Trends That Defined JavaScript in 2025

    JavaScript in 2025 saw a shift toward performance optimization and web standards, with developers questioning React's dominance as modern browsers matured. Signals emerged as a key reactivity pattern across Angular, Vue, Solid, and Svelte. Compilers took on more optimization work, with React Compiler and Svelte 5's Runes automating performance improvements. VoidZero launched Vite+ as a unified Rust-based toolchain to address JavaScript's fragmentation. AI integration moved to the frontend through MCP servers and browser-based machine learning libraries, while new frameworks like Hono, Mastro, and Wasp addressed specific use cases from edge computing to multipage apps.

  4. 4
    Article
    Avatar of newstackThe New Stack·21w

    TanStack Adds Framework-Agnostic AI Toolkit

    TanStack released an alpha framework-agnostic AI toolkit supporting JavaScript/TypeScript, PHP, and Python with adapters for OpenAI, Anthropic, Gemini, and Ollama. Anthropic acquired Bun JavaScript runtime to power Claude Code while keeping it open source under MIT license. Vite 8 beta launched with Rolldown bundler, delivering 10-30x faster builds than Rollup. Infragistics open sourced its Ignite UI library with 50+ enterprise components for Angular, Blazor, React, and Web Components. Chrome published a guide to 22 new CSS and UI features from 2025, including customizable select elements and native carousel APIs. Svelte launched a new community website with dynamic content feeds and updated SvelteKit with streaming file uploads and CLI improvements.

  5. 5
    Article
    Avatar of newstackThe New Stack·21w

    Java Developers Get Multiple Paths To Building AI Agents

    Java is emerging as a strong contender to Python for enterprise AI development, with multiple frameworks now available for building AI agents. LangChain4j, Embabel, Koog, and Crew4J provide Java developers with tools for integrating LLMs and building agentic systems. While Python remains dominant for research and prototyping, Java's advantages in scalability, performance, concurrency, and enterprise integration position it well for production AI deployments. Industry analysts and framework creators predict Java could match or surpass Python for enterprise AI within 18 months to three years, particularly as the technology matures and production requirements become more critical.

See all The New Stack archives