#656 – December 14, 2025
focus on solving user problems, work well with people, and prioritize clarity
Hello! 👋
I can't believe it's almost the end of 2025. This year, I worked on the foundations and the less visible parts of the newsletter. It is back on a homebrewed self-hosted platform with improved deliverability and performance. Emails are more accessible and cleaner. And I'm following almost 1000 different authors and websites every day to find the most interesting articles.
I've put together a short survey so if you'd like to help me make the newsletter better in 2026 please fill it in. It shouldn't take more than 5 minutes. And you can participate in a $30 book voucher draw as a thank you for being an awesome reader. Here's the survey.
21 lessons from 14 years at Google
12 minutes by Addy Osmani
Addy shares key lessons he's learned during his 14 years at Google. He argues the most successful engineers focus on solving user problems, work well with people, and prioritize clarity over cleverness. They ship imperfect solutions quickly to learn from real feedback rather than debating perfect designs. Success comes from building relationships, removing unnecessary work, and understanding that good engineering is ultimately about people.
Developers earned $523K in November building tools on Apify. Now there's $1M more in prizes.
sponsored by Apify
Build web scrapers, API wrappers, or automation tools. Publish on Apify Store and earn from every user who runs them. $2K per tool based on monthly active users. $30K/$20K/$10K grand prizes. $2K weekly spotlights. Plus ongoing revenue after the challenge ends. Register now, build and publish tools, attract users by January 31.
Why does Cloudflare use lava lamps to help with encryption?
8 minutes by Cloudflare
Cloudflare uses about 100 lava lamps in their lobby to create strong encryption keys. A camera takes photos of the lamps and converts them into random numbers that computers need for secure encryption. Since computers are designed to be predictable, they cannot generate truly random data on their own. The unpredictable movements of lava lamp contents provide the randomness needed to protect encrypted data from attackers.
Build a working game of Tetris in Conway's game of life
10 minutes by Totally Human
This discussion explores building a Tetris game within Conway's Game of Life using cellular automaton rules. The solution must accept input by manually changing cell states to represent game controls like piece movement and rotation. Performance is judged on five criteria: smallest bounding box area, fewest manual cell changes needed for input, fastest execution speed, lowest initial live cell count, and earliest submission time.
A modern guide to SQL JOINs
32 minutes by Alexey Makhotkin
Alexey's technical guide takes a fresh approach to teaching SQL JOINs by starting with LEFT JOIN instead of INNER JOIN. It emphasizes using only ID equality comparisons in ON conditions and avoiding the confusing 1:N case of LEFT JOIN. He presents JOINs through practical examples with an employee payment database.
How Pinterest built a real‑time radar for violative content using AI
10 minutes by Faisal Farooq et al.
Pinterest created an AI system to measure how often users see harmful content on their platform. The system samples content based on user views and uses AI to label it, providing daily reports on policy violations. This replaced expensive human reviews that took six months, giving Pinterest real-time insights to quickly detect problems and improve safety measures.
And the most popular article from the last issue was: