#634 – July 13, 2025
You should know how good your code needs to be
How I build software quickly
11 minutes by Evan Hahn
Evan shares his principles for shipping quickly in a small team with an established product. You should know how good your code needs to be for the task at hand. Start with a rough draft/spike. Try to soften requirements if you can. Don’t get distracted. Make small changes. And practice specific skills.
How engineers at Nubank and Ramp spend more time on shipping features (vs drowning in their backlog)
sponsored by Devin AI
What do elite engineering teams at Ramp, Nubank, and Gumroad have in common? They all save 10k+ hours monthly by delegating tedious work (bug triage, refactors, scoping, testing) to Devin so they can focus on hard problems and new features. Devin is a SWE agent built for real-world work. It analyzes your codebase to craft context-grounded plans before implementing and testing. Starting at $20.
The most mysterious bug I solved at work
17 minutes by Cadence
Cadence shares how she tracked down a mysterious bug that appeared every 2-4 weeks in an e-referrals application. The issue involved referrals failing due to an illegal character (0x2) appearing in doctors' referral letters. After extensive investigation, she discovered the culprit: when doctors copied text from previous referral PDFs using Microsoft Edge's PDF viewer, hyphens at line breaks were incorrectly converted to the 0x2 control character, causing XML validation failures.
Why do we “call” functions?
9 minutes by Arthur O’Dwyer
Calling a function is like calling on a friend — we go, we stay a while, we come back. Calling a function is like calling for a servant — a summoning to perform a task. Calling a function is like making a phone call — we ask a question and get an answer from outside ourselves. Which one is it?
How to become passionate about delivering shareholder value
7 minutes by Sean Goedecke
Delivering shareholder value is underrated among software engineers. There is many smart, competent engineers who believed that their job was to write good code, and if any shareholder value happened along the way that was the company’s business. But they should to be focusing on shareholder value.
Stop hiding my controls
15 minutes by Philip Kortum
Control discoverability is still an important interface design principle, and the increasing prevalence of hidden controls on new interfaces is, paradoxically, a regression to a time when computers were harder to use because their functions were not visible.
And the most popular article from the last issue was: