TLDR WEB DEV 2024-07-19

* Enhancing distributed tracing and monitoring

* Fostering a culture of observability * Ensuring scalability and flexibility

* Implementing AI and machine learning

* Managing cost without sacrificing visibility

RSVP for free ↗️

HOW A SINGLE LINE OF CODE BROUGHT DOWN A BILLION DOLLAR ROCKET

This article goes over major software bugs in the real world. For example, the Ariane 5 rocket exploded in 1996 due to a software error, causing a major financial loss and delaying scientific research. An airline was grounded for 3 hours due to an uncaught SQLException in its core system, highlighting the importance of error handling and system isolation. The Boeing 737 MAX crashes were partially caused by a flawed software system (MCAS).

ODIN: UBER'S STATEFUL PLATFORM

Uber created Odin to automate and simplify the management of its giant data storage infrastructure. Odin is designed to be declarative and self-healing, continuously working to align the actual state of workloads with their intended goal states. Odin's control plane is divided into global and host-level components, with two agents - one technology-agnostic and the other technology-specific - working together to manage workload scheduling, cluster topology, and resource allocation.

SNEAKY REACT MEMORY LEAKS: HOW USECALLBACK AND CLOSURES CAN BITE YOU

Closures combined with the useCallback hook in React can cause hidden memory leaks by holding onto large objects in the component's context, preventing garbage collection. Memoized functions may inadvertently reference each other and shared data, creating an endless chain of closures. To avoid such issues, keep closure scopes small, avoid unnecessary memoization, and consider using useRef for large objects.

POST-ARCHITECTURE: PREMATURE ABSTRACTION IS THE ROOT OF ALL EVIL

Post-Architecture is an approach to software development that prioritizes incremental design over premature abstraction. It focuses on procedural programming with a functional flavor, focusing on data structures rather than code patterns, and identifying data flows to guide development.

TRUST AS A BOTTLENECK TO GROWING TEAMS QUICKLY

Trust is necessary for effective collaboration as teams grow. A lack of trust can hinder productivity. Trust allows for efficient delegation and decision-making, especially in high-pressure situations. However, building trust isn't easy because it requires proactive effort, being reliable, sharing mental models, communicating openly, and giving feedback.

EMBRACING FUNCTIONAL CSS: HAVE WE MOVED ON?

Functional CSS, or utility-first CSS, simplifies styling by using small, reusable classes for specific purposes. Tailwind CSS and Tachyons are examples of functional CSS frameworks. They're still popular today because of how easy they are to pick up and use.

Join live to build with the Sentry team

DEVZAT

Devzat is a custom SSH server that takes you to a chat instead of a shell prompt.

NATIVELINK

NativeLink is an efficient build cache and remote execution system that makes software compilation and testing faster. It optimizes build processes for projects of all sizes by caching build artifacts and distributing tasks across multiple machines.

FLOWCHART FUN

Flowchart Fun allows you to turn text into visual flowcharts.

WHAT WOULD IT TAKE TO RECREATE BELL LABS?

Bell Labs was a highly successful industrial research lab that made numerous groundbreaking inventions and discoveries in the 20th century. Its success was due to several factors, including being a subsidiary of AT&T, a government-sanctioned monopoly, which allowed for long-term research projects and a broad research scope. The lab also benefited from historical circumstances like the advancement of quantum mechanics and the influx of talented European physicists after WWII. However, the lab's decline began after the breakup of AT&T in 1982, as it was repeatedly reorganized and faced increasing financial pressure.

OVERCOMING THE LIMITS OF CURRENT LLMS

LLMs like ChatGPT have limitations, like hallucinations, lack of confidence estimates, and the absence of citations. One approach to address these limitations is to bootstrap consistent LLMs by training them on carefully curated datasets that exclude contradicting information. This could be achieved by starting with a small, high-quality corpus and gradually expanding it using the base model to classify new text documents as consistent or inconsistent.

THE ECONOMICS OF A POSTGRES FREE TIER

Xata offers a sustainable free tier for its Postgres and OpenSearch database services by utilizing shared clusters that efficiently host multiple databases. It manages costs effectively through careful cluster management and automation, distinguishing between active and inactive databases to optimize resource allocation.

NVIDIA TRANSITIONS FULLY TOWARDS OPEN-SOURCE GPU KERNEL MODULES

NVIDIA is transitioning fully towards open-source GPU kernel modules in the upcoming R650 driver release.

USPS SHARED CUSTOMER POSTAL ADDRESSES WITH META, LINKEDIN, AND SNAP

USPS shared postal addresses with Meta, LinkedIn, and Snap through hidden tracking pixels, but it stopped after an investigation by TechCrunch, claiming the company was unaware of it.

WHY CODING EXERCISES IN TECHNICAL INTERVIEWS ARE A TERRIBLE IDEA

Coding exercises in technical interviews can be unfair to those without ample practice time outside of their jobs, like those with families and kids.

MICROSOFT BUILDS TOOL TO LET LLMS ANALYZE SPREADSHEETS

Microsoft researchers created SpreadsheetLLM, a tool that uses large language models to analyze and manipulate spreadsheet data, and SheetCompressor to address the token constraints of LLMs.