Search the archive

Take the risk assessment (2 minutes)

THE STRANGE BEHAVIOR OF THE VOID TYPE IN TYPESCRIPT

In TypeScript, the `void` type signifies "nothing" and is mostly used to declare the return type of functions. While you can declare a variable as `void`, it's generally not recommended as it can only hold the value `undefined`, making it impractical. Functions declared with a `void` return type are expected to return `undefined`, either explicitly or implicitly. However, due to how TypeScript handles function type declarations and shorthand syntax, functions with a `void` return type can actually return non-`undefined` values - it's important to note that this is a special case to accommodate common usage patterns.

THE RISE OF OPEN SOURCE TIME SERIES DATABASES

Time series databases store and query data with timestamps, which is useful for infrastructure monitoring, finance, and IoT applications. These databases are an alternative to traditional relational databases, especially when dealing with billions or trillions of data points. The future of time series databases will likely involve solutions for high cardinality, where a large number of unique data points can strain performance, and time series churn, where frequently changing data points can make historical analysis challenging.

TERENCE TAO: I HAVE PLAYED A LITTLE BIT WITH OPENAI'S NEW ITERATION OF GPT, GPT-O1

Terence Tao, a famous mathematician and UCLA professor, says that OpenAI's GPT-o1 is getting closer to helping with advanced math tasks and research. It successfully identified and applied Cramer's theorem to solve a complex mathematical query, something that previous models struggled with. However, it still needed lots of hints and prodding for complex analysis problems.

MONOLITH FIRST

A "monolith-first" strategy is better for building new applications, even if microservices are anticipated later. Starting with a monolith is more practical and less risky than starting with microservices due to the significant overhead associated with managing a distributed system. Also, it's hard to define stable service boundaries upfront, making a monolith a better platform for experimentation and refining boundaries before moving to a microservices architecture.

WHY YOU SHOULD "DESIGN IT TWICE"?

Brainstorming at least two distinct designs when creating software allows you to compare and identify the best solution and potentially combine ideas from both. This requires more upfront effort, but prevents costly rework later and improves design quality.

locally so you don't break prod. Start free.

VIOLENTMONKEY

Violentmonkey provides userscripts support for browsers. It works on browsers with WebExtensions support.

AMPLIFY UI

Amplify UI is an open-source library that offers a range of customizable, accessible, and cloud-connected UI components. It has connected components for streamlined cloud workflows, primitive components for consistent branding, data-bound components for dynamic data display, and various theming options.

CONFORM

Conform is a type-safe form validation library for HTML forms. It integrates with server frameworks like Remix and Next.js.

EX-CROWDSTRIKE EMPLOYEES DETAIL RISING TECHNICAL ERRORS BEFORE JULY OUTAGE

A report says that CrowdStrike prioritizes speed over quality, which may be a reason behind the huge failure it had recently, which paralyzed airlines and caused significant financial losses. Former employees claim they repeatedly raised concerns about rushed deadlines, insufficient training, and increasing technical problems, but their warnings were ignored. The company disputes these claims, stating that it is committed to quality control and that the information came from disgruntled former employees.

GROUNDING AI IN REALITY WITH A LITTLE HELP FROM DATA COMMONS

Google's DataGemma models address the issue of hallucinations in LLMs by grounding them in real-world data from the Data Commons knowledge graph. Two approaches are used: Retrieval Interleaved Generation (RIG) and Retrieval Augmented Generation (RAG). RIG fine-tunes the model to identify statistics and verify them against Data Commons, while RAG retrieves relevant information before the LLM generates text.

EXPLORING THE SCALABLE MATRIX EXTENSION OF THE APPLE M4 PROCESSOR

Apple M4 chip's Scalable Matrix Extension (SME) unit can achieve impressive peak compute rates for matrix multiplication. However, its performance for vector operations is much lower due to limitations in data movement. This is due to the architecture's reliance on tile storage, where data needs to be moved from the CPU registers to the ZA storage for optimal performance.

LEARN GIT BRANCHING

Learn Git Branching is an interactive visual website that teaches Git through game levels.

FOUNDER MODE, HACKERS, AND BEING BORED BY TECH

The cult of the founder in Silicon Valley has created a one-dimensional tech landscape dominated by hype cycles and self-important personalities.

BLOG OS

Source code for writing an OS in Rust.

WHY KUBERNETES NEEDS AN LTS

Kubernetes should adopt a Long-Term Support (LTS) model with a 24-month support window to address the challenges of frequent upgrades and encourage best practices for cluster management.

WHAT'S NEW IN REACT 19

React 19 introduces major updates like Server Components, improving performance by rendering on the server and reducing client-side JavaScript. New features include Actions for handling events and enhanced hooks like useActionState for managing form states, streamlining both server and client workflows.

THE UNDENIABLE UTILITY OF CSS :HAS

The CSS :has pseudo-class provides a new way to style elements based on their relationships to other elements in the DOM. It is a powerful tool for creating more complex and dynamic CSS styles. However, it is important to consider browser support and the complexity of your use case before using :has.

NOTES ON OPENAI'S NEW O1 CHAIN-OF-THOUGHT MODELS

OpenAI has released two new "chain-of-thought" models, o1-preview and o1-mini, which prioritize reasoning over speed and cost. These models are trained to think step-by-step, enabling them to handle more complex prompts requiring backtracking and deeper analysis. While the reasoning process is hidden from users due to safety and competitive advantage concerns, it allows for improved results in tasks like generating Bash scripts, solving crossword puzzles, and validating data.

HOW TO COMMUNICATE TRADEOFFS SO LEADERS WILL LISTEN

It's important to consistently communicate team priorities and frame tradeoffs clearly when handling high-priority requests from leadership. The "steelmanning" technique, which involves thoroughly understanding and presenting the strongest case for new requests to make more informed decisions, can also be useful.

with Dan

REDIRECT CHECKER

URL redirects are a way to send users and search engines from one URL to another. This website provides tools to analyze your redirects like detailed path analysis, bulk redirect checking, and more.

LEGEND MOTION

Legend Motion is a declarative animations library for React Native that makes it easy to transition between styles without needing to manage animations.

LEARNING TO REASON WITH LLMS

OpenAI o1 is a new large language model trained to reason by generating a chain of thought before responding. This model demonstrates significant advancements in reasoning capabilities, achieving impressive performance in various tasks, including competitive programming, math Olympiads, and scientific problem-solving. OpenAI o1-preview is now available for use in ChatGPT and through the API, allowing users to explore its reasoning abilities.

AN UNPOPULAR PERSPECTIVE ON THE SSO TAX

The so-called "SSO Tax" in SaaS pricing is not an unfair practice but rather a form of price discrimination that is both logical and beneficial for both vendors and customers. Price discrimination is inherent in any market, as buyers have different preferences and willingness to pay. Vendors need to segment their customers to maximize their revenue. For example, airlines have different pricing tiers that allow vendors to offer more affordable basic tiers while still catering to the needs of high-paying customers.

DOES YOUR STARTUP REALLY NEED COMPLEX CLOUD INFRASTRUCTURE?

Startups often overcomplicate their cloud infrastructure, neglecting product-market fit in favor of complex deployments

WE SPENT $20 TO ACHIEVE RCE AND ACCIDENTALLY BECAME THE ADMINS OF .MOBI

These researchers inadvertently became the admins of the .MOBI TLD after registering an expired domain, which allowed them to undermine the CA process for the entire .mobi TLD.

UNDERSTANDING CONCURRENCY, PARALLELISM, AND JS

This article explains the difference between concurrency and parallelism, how they are implemented in JavaScript, and the potential pitfalls of using them in Node.js.

code snippet. Create your free account now.

DEBUGGING TILL DAWN: HOW GIT BISECT SAVED MY DEMO

SQLITE ON RAILS: THE HOW AND WHY OF OPTIMAL PERFORMANCE

SQLite out-of-the-box on Rails isn't production ready. It needs some performance improvements with the help of some Ruby gems (packages) that enable immediate transactions, manage write query queues, and release the Global VM lock while queries wait for database responses. SQLite should be used in the write-ahead-log (WAL) mode and should have separate connection pools for reading and writing operations.

WHY COPILOT IS MAKING PROGRAMMERS WORSE AT PROGRAMMING

AI tools like GitHub Copilot are making programmers worse at programming. These tools can erode fundamental programming skills and create a false sense of expertise. Relying on them without a deep understanding of the code and the ability to problem-solve independently will make developers dependent on AI.

WHY NOT COMMENTS

Comments that explain why a certain approach wasn't taken in the code are valuable. While identifiers can describe "what" the code does, they cannot always properly convey the reasoning behind choosing a less optimal but simpler solution.

WHY SOME OF US LIKE "INTERDIFF" CODE REVIEW

Traditional code review systems, like GitHub, suffer from a "diff soup" problem, where multiple commits are crammed together and reviewed as a single diff. This makes it difficult to follow the evolution of code changes and leads to issues with blaming, bisecting, and incremental review. An "interdiff" approach, where revisions of the original commits are published instead of new commits, lets reviewers see incremental changes and lets authors avoid cluttering history with "fix review" commits.

Schedule a demo to learn more 🐺

PLATE

Plate is a rich-text editor framework for React built on top of Slate. Its key components include a plugin system, unstyled UI primitives, and pre-built styled components.

AWS AI STACK

The AWS AI Stack is a full-stack serverless boilerplate project for building AI applications on AWS. It helps devs quickly create AI chatbots, authentication services, business logic, and asynchronous workers using AWS Lambda, API Gateway, DynamoDB, and EventBridge.

VISUALDB

Visual DB is a web-based tool that simplifies database management. It offers a drag-and-drop interface for creating forms, sheets, and reports, and uses AI to assist with form layout and natural language query translation.

SIMULATOR-BASED REINFORCEMENT LEARNING FOR DATA CENTER COOLING OPTIMIZATION

Meta has successfully used reinforcement learning to optimize energy and water consumption in its data centers. Its approach uses historical data and a physics-based simulator to train a model that determines the optimal airflow setpoint for various weather conditions.

LEADING EFFECTIVE ENGINEERING TEAMS: A DEEPDIVE

DAVID CHANG ON THE LONG, HARD, STUPID WAY

Doing things "the long, hard, stupid way" means maintaining high standards, even in casual settings. This approach, though inefficient, fosters a culture of excellence and ultimately contributes to success.

GOOGLE WILL NOW LINK TO THE INTERNET ARCHIVE TO ADD MORE CONTEXT TO SEARCH RESULTS

Google has partnered with The Internet Archive to add links to the Wayback Machine in its search results, providing users with historical context for web pages.

THE DEV TOOLS PERFORMANCE MONITOR PANEL

This developer describes how the Performance Monitor panel in DevTools helped identify and fix a performance issue in an animated spinner component, ultimately reducing CPU usage and improving the overall performance of the website.

POSTGRES TO CLICKHOUSE: DATA MODELING TIPS

This is a comprehensive guide on data modeling techniques for migrating from PostgreSQL to ClickHouse that focuses on handling duplicates, understanding data types, and choosing the optimal ordering key for efficient query performance.