Programming

Years ago, something clicked when I read about Anthony Bourdain’s approach to feeding crowds. It was his stunt turkey technique, where he cooks an extra turkey the week before the actual event. It made me think about how we release software. It seemed absurd at first: why would you cook two turkeys for one dinner? Then I realized what he was really doing: load balancing across time. That extra turkey unlocks a cascade of optimizations that aren’t obvious until you think through the dependencies. Turkey stock made a week ahead means better gravy and more flavourful stuffing. It also means practice runs; most people only roast a turkey once or twice a year, so that stunt turkey is essentially a fire drill. Having stock and rendered fat ready lets you prep other components days in advance. Suddenly your Thanksgiving-day workflow changes from frantic juggling to orchestrated execution. ...

Remember autoconf? If you were building software in the late ’90s or early 2000s, you probably have scars. You’d write some configuration rules, run the tool, and out popped a perfectly functional makefile. It felt like magic, until it didn’t. The moment autoconf broke, you were debugging generated shell scripts that stretched for thousands of lines. The abstraction that seemed so elegant suddenly became opaque hell. You needed to understand not just what you wanted, but how the magic worked, where it could fail, and why. ...

Evolution has a weird obsession with crabs. At least five times, nature has taken some random crustacean and said, “you know what would work better? A crab.” But it’s not just crabs, evolution continues to reinvent the wheel. Trees evolved independently more than 40 times. Many different types of eyes appeared in dozens of unrelated lineages. Flight emerged in insects, pterosaurs, birds, and bats, each independently. These aren’t coincidences. They’re examples that good design is inevitable. ...

Your AI coding assistant hasn’t gotten worse; you’ve just hit the ceiling of what your current approach can handle. “My LLM coding tools have been serving me hot garbage lately” is a complaint I see weekly on Reddit. Same tool, same developer, suddenly worse results. These threads are full of developers convinced their AI assistant has degraded overnight. Same prompts that worked last month suddenly fall over. The tool must be broken, right? ...

There’s a familiar anxiety in the air among software architects and engineering leaders. AI coding tools are going to make us lazy programmers. They’ll erode our fundamental skills. We’ll lose the ability to think through problems. Sound familiar? It should. I’ve heard this exact song before, just with different verses. In the ’90s, visual development environments were going to create a generation of developers who couldn’t “really” code. Before that, higher-level languages were supposedly making us soft. Hell, I remember when debuggers were contentious—Linus famously ranted that real kernel developers should be able to think through their code without stepping through it line by line. ...