#620 – April 06, 2025
Focus on high-impact tasks and don't do any unpaid work
How to work in tech when your job isn't safe
4 minutes by Sean Goedecke
In the face of rising pressure and the threat of layoffs, tech employees must adapt to new expectations. In this post Sean discusses that instead of extending work hours, engineers should focus on high-impact tasks that directly contribute to measurable output, while minimizing unpaid work like writing tests or refactoring code. By prioritizing crucial projects when attention is highest and resting during quieter periods, engineers can maintain productivity and avoid burnout, ensuring sustainable success in an increasingly demanding environment.
Why is NoSQL still failing us?
sponsored by Volt Active Data
Ten years ago, NoSQL was heralding the next phase of the database revolution. To be clear: it brought distinct advantages over relational databases. And also to be clear: after widespread adoption, NoSQL databases are still falling short in ways that make dealing with things like AI-based workloads almost impossible without compromising on your data processing in some important way.
How Apple Pay Handles 41 Million Transactions a Day Securely
6 minutes by Neo Kim
In this article Neo explains Apple Pay's architecture and security model. Instead of storing credit card details, Apple Pay generates a Device Account Number through payment networks, storing it in the iPhone's secure element. When making a payment, the system creates unique cryptograms to authorize transactions without sharing sensitive data. This process leverages secure hardware components like the secure element and enclave, biometric verification, and NFC technology to provide a safer payment alternative than traditional credit cards.
Exploring Generative AI
about 1 hour by Birgitta Böckeler
Birgitta dives deep into the impact of generative AI and LLMs on software development practices. She begins by acknowledging the uncertainty around how this technology will affect the programming profession long-term, while documenting her findings as Thoughtworks' coordinator for AI in software delivery. The memos cover topics including the AI coding toolchain landscape, practical examples of using GitHub Copilot for coding tasks, best practices for test-driven development with AI assistance, and considerations around when AI tools are most useful versus when they can create problems.
The Fifth Kind of Optimisation
16 minutes by Laurence Tratt
In this article Laurence expands on his previous article about optimization techniques, acknowledging that he mistakenly omitted "parallelization" as a major optimization strategy. He explains that modern hardware now offers significant parallelization potential with multi-core CPUs, and languages like Rust have made multi-threaded programming safer and more accessible. The article details how parallelization significantly improved his website builder's performance, making it 2-3 times faster, and dramatically reduced testing time.
The 13 software engineering laws
9 minutes by Anton Zaides
Hyrum's law, Conway’s law, Zawinski's law, and 10 others. None of those laws is a ‘real law’ - they are just great mental models. Knowing them might save you some pain in the day-to-day.
And the most popular article from the last issue was:
And to finish on a similar topic here's an: