It’s fitting that the mass exodus of Twitter engineers coincided with the death of Fred Brooks, author of The Mythical Man-Month. There is no more influential book about software engineering, and I really can’t say enough about how Brooks influenced my thinking about my chosen profession. The key insight that everyone repeats from the book is that adding engineers to a project won’t make it go faster, and might make it go slower. For me, though, one of the big lessons from the book is how Brooks takes an obvious fact — some engineers are far, far more productive than others — and tries to use that insight as a basis to form software teams. Brooks used the analogy of the “surgical team” where one surgeon (the best engineer) performs the most critical work, to try to leverage this disparity in ability.
Last night, a bunch of surgeons left the building at Twitter: this Twitter thread of Twitter engineers signing off is full of people with multi-year tenures. There is no recipe for building a major real-time system like Twitter, and there is no off-the-shelf software that can be purchased to keep it running. Every long-tenured engineer leaving the building is leaving with some special knowledge of a bespoke system. In normal times, those engineers would have been replaced and the rest of the team would educate the new engineer on the ins and outs of whatever system they’re maintaining. Yesterday, entire teams left. The knowledge they took with them is irreplaceable, in this context.
Like every big tech company that’s been around for years, I’m sure Twitter had more than a few senior engineers who were phoning it in. Being constructively lazy is one of the important characteristics of a good engineer, but some of them just build a sinecure and collect their big paychecks. So, over time, an owner of Twitter who had clue fucking one about software engineering might have undertaken strategic layoffs to lean down the organization without losing too much knowledge. Musk’s ultimatum: be hardcore or take three months’ severance, was a stupid, hamfisted effort to selectively cull the lazy that turned out to be a no-brainer for pretty much anyone who isn’t restricted by a visa or an absolute need to have a paycheck.
I’m not going to make predictions about when or how Twitter will fail. We may see major outages, or maybe Musk will hire enough contractor troubleshooters to keep it limping along. But the simple fact is that a man who died yesterday wrote a book almost 50 years ago that encapsulated everything that Elon doesn’t know about software, and we’re all going to see how violating those rules turns $44 billion into a tiny fraction of that number in less than a month.