TLDR Web Dev

TLDR WEB DEV 2025-02-14

5 sections14 entrieslinks blockcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01ARTICLES & TUTORIALS

REPLACE YOUR JAVASCRIPT ANIMATION LIBRARY WITH VIEW TRANSITIONS

The new View Transitions API in Chrome and Safari makes web animations easier. For example, a navigation animation can be created using only HTML, CSS, and a few lines of JavaScript. The process involves using `startViewTransition` to smoothly transition between DOM states, along with CSS to style the transitions and `view-transition-name` to target specific elements for animation.

builder.io15 min read

HOW WE HACKED A SOFTWARE SUPPLY CHAIN FOR $50K

Two bug bounty hunters collaborated to find a critical vulnerability in a company's recently acquired subsidiary. They discovered a GitHub Actions token embedded within a Docker image, allowing them to potentially compromise the entire development-to-production lifecycle. By analyzing the Docker image layers using tools like `dive` and `dlayer`, they recovered a private npm token, letting them inject malicious code into a private package used across the company's systems.

landh.tech19 min read

02OPINIONS & ADVICE

AI IS STIFLING TECH ADOPTION

AI models are stifling the adoption of new technologies due to outdated training data and inherent biases. The "AI knowledge gap” is caused by the time lag between a technology's release and its inclusion in AI training data, which is discouraging developers from using newer tools. Also, AI models often have a preference for certain technologies (like React and Tailwind), which influences developers' choices.

vale.rocks10 min read

ASK HN: FORMER EMPLOYEES' RSUS AT RISK AFTER STARTUP'S IPO (HACKER NEWS THREAD)

Former employees of a recently IPO'd startup are seeking advice on their RSU settlement process, which is different from typical practices. The company requires them to prepay taxes in cash before receiving shares, which means they might lose their shares if the payment isn't made by a strict deadline. This is different from the sell-to-cover option offered to current employees, which is unfair.

news.ycombinator.com

03LAUNCHES & TOOLS

Sponsored

HOW TO INSPECT REACT SERVER COMPONENT ACTIVITY WITH NEXT.JS

If you're used to debugging and profiling code with in-browser dev tools, React Server Components (RSCs) are going to ruin your day. Learn how to use OpenTelemetry with Next.js to trace the execution of RSCs [8] on the server side — and add observability and visualization with Dash0 [9].

dash0.comSponsor

LAUNCHING INTEROP 2025

The Interop 2025 project improves web platform interoperability across different browsers. Interop 2024 achieved a 95% interoperability score, and 2025 will address 19 focus areas, including improvements to existing features and new ones like Storage Access and CSS Zoom.

hacks.mozilla.org10 min read

DATA FORMULATOR

Data Formulator is a Microsoft Research app using AI to make data visualization easier. It lets users create visualizations iteratively with natural language, transforming data and generating charts based on user instructions and prompts.

github.comGitHub repository

EMITTERY

Emittery is a simple asynchronous event emitter for Node.js and browsers. It has a clean API with methods for subscribing to events (`on`, `once`), emitting events (`emit`, `emitSerial`), and managing subscriptions. Emittery uses symbols for event names to avoid collisions and has many debugging options.

github.comGitHub repository

04MISCELLANEOUS

HOW I BUILT AN AI COMPANY TO SAVE MY OPEN SOURCE PROJECT

A developer's open-source planning project, OptaPlanner, almost died after Red Hat's acquisition by IBM. As a result, he created Timefold AI, a company built around a commercial offering (open core business model) extending OptaPlanner, with the help of his open-source reputation.

timefold.ai19 min read

EMERGING PATTERNS IN BUILDING GENAI PRODUCTS

When building production-ready generative AI products, evals are used to assess model performance and embeddings are used to represent large datasets efficiently. RAG is used to add context and ground LLM responses. LLMs are still not hallucination-free nor are they always efficient in retrieving context using RAG. Some solutions for inefficient retrieval are discussed, like hybrid retrievers, query rewriting, and rerankers.

martinfowler.com30 min read