Programmer Weekly (Issue 262 July 17 2025)

Welcome to issue 262 of Programmer Weekly. Let's get straight to the links this week.

Learn how to make AI work for you

AI won’t take your job, but a person using AI might. That’s why 1,000,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.


Quote of the Week

“Configuration is coding in a poorly designed programming language without tests, version control, or documentation.” — Gregor Hohpe


Reading List

A custom web crawler design, using 12 high-powered EC2 nodes, achieved crawling a billion pages in just over 24 hours for under $500 by optimizing for CPU-bound parsing and fault tolerance, despite modern web complexity and the dominance of SSL, while respecting robots.txt and politeness policies. The experiment revealed that parsing (not network or DNS) is now the main bottleneck, and that horizontal scaling with independent, domain-sharded nodes is more effective than vertical scaling for this workload.

The post details the design and optimization techniques used to create a high-performance lexer for the Purple Garden language, focusing on minimalistic token representation and using advanced methods like threaded lexing (computed gotos) to replace switch statements for speed gains. It also discusses abstracting memory allocation through an Allocator interface to support flexible and efficient memory management in the lexer and virtual machine components.

While migrating Pinterest’s large-scale, memory-intensive search system (Manas) to Kubernetes, the team encountered rare but severe latency spikes—affecting one in a million requests—traced to an unexpected interaction between their search process and cAdvisor’s intrusive memory monitoring, which caused contention in the Linux kernel page tables. By disabling the problematic cAdvisor feature and sharing their findings, they resolved the issue and unlocked smoother migration for other critical workloads to Pinterest’s Kubernetes platform.

It started with a simple thought: four bytes in a hash table bucket look just like an integer. Luckily, this one idea led to a deep dive into bit-twiddling and a 2x performance boost.

Dropbox’s advanced key management system uses a three-tier encryption hierarchy (team, namespace, and block keys) anchored in hardware security modules, enabling fast, secure file sharing and copying for teams with sensitive data while maintaining strong isolation and quick revocation if needed. This approach balances robust security with high performance by minimizing upfront re-encryption for shared files and gradually synchronizing keys in the background, ensuring both enterprise-grade protection and a smooth user experience.

Branch instructions are the primary means by which a program running on a CPU makes a decision.

Leading at scale means empowering teams with autonomy and trust instead of controlling every detail, enabling faster, better decisions and innovation through decentralized authority and outcome-focused leadership. The article warns that true organizational growth comes from reducing friction and fostering collaboration, not from micromanagement or excessive control.

AI tools were expected to boost developer productivity, but a randomized trial found they actually increased completion time by 19% for experienced open-source developers. This contradicts forecasts from developers and experts, suggesting that under real-world conditions, AI can slow down code work despite high expectations.

The author explores how LLMs handle tasks with varying levels of complexity, often excelling at surface-level predictions but struggling with deeper, context-rich problems. He argues that treating LLMs as if they have human-like understanding or intent is misleading, as they are fundamentally just pattern prediction engines shaped by data.

This is for managers who want to delegate AND maintain a high bar for quality of work. Here are the 5 tactics I learned over time that made delegating easier.

This glossary is an overview of the concepts that you’ll need to think about distributed systems reliability. We’re writing chiefly for industry practitioners – software developers who are learning about distributed systems testing at any stage of their careers.

This post details the process of using Bun to handle a 1 billion row file (13.8GB) in under 10 seconds. It covers the Bun APIs considered, byte processing and chunking strategies, plus analysis of IO versus CPU workloads.


Watch, Listen

Lennart Augustsson presents microHs, a minimal Haskell compiler that translates Haskell code into combinators for execution via a compact C runtime, resulting in tiny, portable binaries suitable even for microcontrollers. The talk covers installation, architecture, performance comparisons with GHC, and live demos—including running Haskell on a microcontroller—highlighting microHs’s focus on simplicity, portability, and educational value.

This course teaches developers and data scientists practical strategies for securing AI systems, covering how AI threats differ from traditional software, building effective threat models, and defending against attacks like prompt injection, data poisoning, and input manipulation through real-world examples and hands-on exercises.

This video teaches Bash scripting basics through a real-world log analysis example, demonstrating automation with variables, loops, and file operations to turn manual work into efficient, reproducible scripts. It covers environment setup, script writing, and best practices for making scripts executable and useful in daily tasks.


Interesting Projects, Tools and Libraries

A high-performance inference engine for AI models on Apple Silicon.

A desktop and mobile UI for Claude Code, Anthropic's official CLI for AI-assisted coding. Works locally or remotely with a proper interface everywhere.

A decentralized peer-to-peer messaging app that works over Bluetooth mesh networks. No internet required, no servers, no phone numbers.

Kanban board to manage your AI coding agents.

The front-end to your dev env.

MVPBlocks is an open-source component library for developers to copy, customize, and launch MVPs lightning-fast—no installations, just pure productivity.

Rill is a tool for effortlessly transforming data sets into powerful, opinionated dashboards using SQL. BI-as-code.

The open-source CapCut alternative.

Kafka TUI client.

Bundler-free build tool for TypeScript libraries. Powered by tsc.


Our Other Newsletters

Python Weekly - A free weekly newsletter featuring the best hand curated news, articles, tools and libraries, new releases, jobs etc related to Python.


Founder Weekly - A free weekly newsletter for entrepreneurs featuring best curated content, must read articles, how to guides, tips and tricks, resources, events and more.