Links

This does a terrific job of expanding what I’ve always thought of as “automated code review,” placing it in the context of systems engineering. I fully agree with everything written.

The two important observations here — that human code actually contains far more bugs than you realize, and that AI is superior to humans at code review itself — I anecdotally find to be true in my personal experience and in the experience of colleagues. Many in my network are still really against this entire idea, and while I empathize with aspects of the code-as-craft disappearing, I find it impossible to side with the anti-AI crowd now. They don’t look informed anymore. They look out of touch. And scared. Not a good look, and not a good way to think about your career going forward.

I wrote a reaction essay about this in 2018 and the position has aged well. The hardest version of “choose boring” is when you are the boring choice — when the right move is to keep using what your team already knows instead of what’s interesting to you personally. The friction this generates inside engineering orgs is real, and it doesn’t get easier with seniority. McKinley’s “innovation tokens” framing is the part I still find myself reaching for years later.

I’ve spent a fair bit of time thinking about which software laws are “real” and thus still applicable in the age of AI. I haven’t spent a ton of time thinking through the analogous topic for management. This post does a good job of outlining the small number of really important things that general managers do that AI won’t immediately make obsolete.