TLDR Index
TLDR Web Dev

TLDR WEB DEV 2023-08-02

5 sections12 entriesinline urlcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01ARTICLES & TUTORIALS

UNDERSTANDING REACT SERVER COMPONENTS

React Server Components (RSCs) individually fetch data and render entirely on the server, streaming the resulting HTML into the client-side React component tree and interleaving with other Server and Client Components as necessary. This process eliminates the need for client-side re-rendering, thereby improving performance. When an RSC needs to be re-rendered, due to state change, it refreshes on the server and seamlessly merges into the existing DOM without a hard refresh.

vercel.com10 min read

HOW TO PERFORM TYPE CASTING IN TYPESCRIPT

This is a comprehensive and advanced guide to TypeScript type casting. It covers the problems and limitations with using the ‘as’ keyword to type cast and how to go about solving these problems with code examples. The guide also explains foundational concepts like subtype and supertypes, type widening/narrowing, and how the satisfies operator works.

blog.logrocket.com7 min read

02OPINIONS & ADVICE

HOW TO BECOME A STAFF-LEVEL ENGINEER AT BIG TECH

A Staff Software Engineer at Meta details his tips to grow fast throughout the levels. He breaks down what each level does, what the role of the staff engineer is, and the uniqueness of Meta’s engineering culture.

newsletter.techworld-with-milan.com8 min read

BAD CODE STALLS DEVELOPER VELOCITY

Without standards for code quality, and with mounting pressure to complete the project at hand, it can be easier to let bad code linger in the codebase for the sake of speed. But that approach is ultimately self-defeating. It slows down the entire development cycle while also hurting product quality and creating security issues. Instead, embracing a Clean as You Code approach is the best way for development teams to deliver quality, bug-free code while still moving swiftly.

thenewstack.io6 min read

03LAUNCHES & TOOLS

Sponsored

Imagine having a direct line to over 150 senior cloud architects

doit.comSponsor

IMESSAGE-EXPORTER

imessage-exporter provides both a library to interact with iMessage data as well as a binary that can perform some useful read-only operations using that data. It supports multi-part messages, replies/threads, attachments, expressives, and more.

github.comGitHub repository

BASARAN

Basaran is an open-source alternative to the OpenAI text completion API. It provides a compatible streaming API for your Hugging Face Transformers-based text generation models.

github.comGitHub repository

04MISCELLANEOUS

DRIVE OPEN STANDARDS FOR 3D CONTENT

Pixar, Adobe, Apple, Autodesk, and NVIDIA, together with the Joint Development Foundation, an affiliate of the Linux Foundation, today announced the Alliance for OpenUSD to promote the standardization, development, evolution, and growth of Pixar’s Universal Scene Description technology. The alliance seeks to standardize the 3D ecosystem by advancing the capabilities of Open Universal Scene Description. By promoting greater interoperability of 3D tools and data, the alliance will enable developers and content creators to describe, compose, and simulate large-scale 3D projects and build an ever-widening range of 3D-enabled products and services.

apple.com9 min read

WHAT IS GOOD FREE-TO-PAID CONVERSION?

This article establishes a baseline for a good free-to-paid conversion rate for different types of products. For a freemium self-serve product, a good conversion rate is 3-5% and a great rate is 6-8%, while for a freemium product with a sales-assist motion, a good conversion rate is 5-7% and a great rate is 10-15%. Products that offer a free trial have higher conversion rates on average, with 8-12% being good and 15-25% being great. The article explains why conversation rates differ for different models and customer base and shows tips for how you can increase the conversation rate for your site.

open.substack.com4 min read

HOW WE BUILT BETTER GENAI WITH PROGRAMMATIC DATA DEVELOPMENT

Snorkel AI used weak supervision to programmatically curate an open source prompt/responses dataset with a 24% increase in response win-rate against ChatGPT. Because the curation was done programmatically, it could be done by two devs in one day as opposed to hundreds of annotators over the course of weeks/months. This article details the problem with manual annotations and the techniques used for dataset filtering and evaluation in a scalable way.

snorkel.ai12 min read