software-engineering

The gap between senior and junior AI usage isn't about knowing more prompts. It's the same gap that existed with Google: seniors know what they need, juniors hope the tool figures it out for them.

Most developers treat commit messages as an afterthought. But when you use AI to write them deliberately — with reasons, not just changes — your git history becomes a powerful context source for every future AI interaction.

Matt Pocock at AI Engineer Europe argues that bad code is now more expensive than ever — and that the engineers winning with AI are the ones who went back to basics.

Birgitta Böckeler's QCon London 2026 talk breaks down context engineering, harness engineering, and why blindly trusting autonomous AI agents is a security and cost disaster waiting to happen.

UC San Diego and Cornell tracked 112 professional programmers using AI agents at work. Not one of them actually vibe coded. Here's what they found — and what it means.

Everyone says AI makes coding obsolete. They're wrong — and the reason why reveals something important about what coding actually teaches you.

A step-by-step workflow for using AI coding assistants effectively — from understanding a ticket with Rovo to reviewing changes before committing with Claude.

Every era of software engineering is really about one thing: building higher abstractions to hide complexity and let developers focus on solutions, not machines. Here's the full arc from machine code to AI.

A practical breakdown of memory types in AI coding tools — what survives between sessions, what gets wiped, and how to build persistent AI teammates that actually know your codebase.

Understanding how context works in AI coding assistants — what goes in, why it runs out, and how to manage it so the AI stays helpful instead of confused.