Best of The New Stack — November 2025

1. 1
   Article

   The New Stack·26w

   The React Component Pyramid Scheme: An Over-Engineering Crisis

   The pursuit of reusable React components has created over-engineered codebases filled with complex abstractions and prop-heavy components that are harder to maintain than simpler alternatives. The DRY principle, when applied dogmatically, leads to premature abstraction and components like Button.jsx that try to handle every use case with dozens of props and conditionals. Intentional code duplication is often more pragmatic than forced reuse, as it preserves clarity and allows code to evolve independently. The solution is adopting a post-reuse mindset that values simple, self-contained, context-aware components over universal abstractions, recognizing that the best component is one you can understand instantly rather than one you can use everywhere.

   74

   4
2. 2
   Article

   The New Stack·27w

   Why the Frontend Should Run AI Models Locally With ONNX

   Running AI models locally in the browser using ONNX Runtime Web offers significant advantages over cloud-based approaches. Local execution eliminates privacy concerns by keeping sensitive data on-device, enables offline functionality, and provides instant feedback loops. ONNX acts as a universal format for ML models, allowing models trained in PyTorch or TensorFlow to run anywhere via JavaScript. Angular's Signals feature (v16+) provides the performance isolation needed for heavy inference operations. The approach enables mixing local models for low-latency tasks with cloud calls for complex reasoning, while maintaining transparency about data handling.

   56

   5
3. 3
   Article

   The New Stack·24w

   2025’s ‘Advent of Code’ Event Chooses Tradition Over AI

   Advent of Code 2025 introduces significant changes after 10 years: reducing puzzles from 25 to 12, removing global leaderboards, and officially banning AI use. Creator Eric Wastl made these adjustments to reduce stress and time commitment while maintaining the event's spirit. Survey data shows 62% of participants use zero AI, with sentiment against AI in the competition increasing. Python remains the most popular language (40%), followed by Rust (16%), with VS Code as the dominant editor and Linux gaining ground over Windows.

   34

   2
4. 4
   Article

   The New Stack·28w

   JavaScript Utility Library Lodash Changing Governance Model

   Lodash, the widely-used JavaScript utility library with 2.57 billion weekly npm downloads, is transitioning from single-maintainer governance to a Technical Steering Committee model under the OpenJS Foundation. Next.js introduces file system caching for development and open sources its evals for AI agents. TanStack Start launches a $140,000 hackathon running through November 17. Deno Deploy receives a complete rebuild with integrated CI/CD, Postgres support, and automatic observability. Vercel releases the Workflow Development Kit in public beta, a TypeScript framework for building durable applications using declarative directives, alongside Agent Investigations for intelligent monitoring.

   26

   1
5. 5
   Article

   The New Stack·28w

   Moving From C++ to Rust? Clickhouse Has Some Advice

   ClickHouse, a 1.5 million line C++ analytics database, shares lessons from their incremental migration to Rust. Rather than a full rewrite, the team integrated Rust modules for specific features like BLAKE3 hashing and Delta Lake support. Key challenges included managing reproducible builds, handling 672 additional Rust dependencies, writing C++/Rust interop wrappers, and dealing with excessive panic behavior in Rust libraries. While Rust offers memory and thread safety advantages, combining it with C++ introduces complexity around build systems, sanitization, and cross-compilation. The team now accepts Rust contributions but has decided against a complete rewrite.

   25
6. 6
   Article

   The New Stack·28w

   OpenAI Co-Founder: AI Agents Are Still 10 Years Away

   OpenAI co-founder Andrej Karpathy predicts AI agents are still a decade away from replacing human workers, despite recent progress with large language models. He argues the industry is over-hyping current capabilities, citing issues like lack of multimodal functionality, continual learning, and the significant demo-to-product gap. Karpathy draws from his experience leading Tesla's self-driving efforts to illustrate how difficult it is to move from working demos to production-ready systems. He's now focusing on AI education through Eureka Labs, releasing projects like nanochat to help developers understand LLM implementation from the ground up.

   19

   1
7. 7
   Article

   The New Stack·28w

   Best Linux Distros for Development

   A curated guide to five Linux distributions optimized for software development: Debian for stability and vast software repositories, Fedora for cutting-edge toolchains and technologies, Pop!_OS for GPU support and the fast COSMIC desktop, openSUSE for rolling releases with testing frameworks, and Linux Mint for ease of use. Each distribution offers unique advantages like built-in development tools, container runtime support, and different approaches to package management and system stability.

   13
8. 8
   Article

   The New Stack·25w

   Is AI Creating a New Code Review Bottleneck for Senior Engineers?

   AI coding tools can rapidly generate scaffolding and basic patterns, but often leave developers struggling with the remaining 70% of work—integration, security, edge cases, and debugging. Google's Addy Osmani discusses how AI-generated code creates a code review bottleneck for senior engineers, as trust in AI output declines despite rising adoption. He recommends understanding generated code thoroughly, investing in context engineering, maintaining strong test coverage, and using AI as a learning tool rather than a crutch. Real productivity gains appear to be less than 2x, with greenfield projects seeing better results than legacy codebases with technical debt.

   10