Big tech companies often produce bad code, and the reasons boil down to a few key factors. Many engineers in these firms are relatively inexperienced with the codebases they work on. The typical tenure for engineers is just a year or two, largely due to compensation structures that discourage long-term commitment. After four years, many face significant pay cuts, prompting them to seek better opportunities elsewhere. Frequent reorganizations further exacerbate the issue, as engineers switch teams and projects, leading to a constant influx of “beginners” making changes to older, complex systems. While “old hands” — experienced engineers who have been around longer — can mitigate some of these issues through code reviews, their expertise is often underutilized. These engineers are typically overburdened, balancing their own responsibilities while trying to maintain oversight on systems they’ve mastered. The result is a workforce that is competent but spread too thin, racing against deadlines and juggling multiple projects without deep familiarity with the code they’re working on. The company culture prioritizes agility and adaptability over quality control. Big tech firms favor a model that allows for rapid redeployment of engineers, which inherently sacrifices the development of long-term expertise in specific areas. This approach leads to code that may work in the short term but is often poorly constructed. Engineers are not given the choice of pure or impure work, and even highly skilled individuals can struggle in unfamiliar environments. As a consequence, even with stronger engineers, poor code is likely to persist because of the organizational dynamics that force them into rapid, high-pressure scenarios with little time to truly understand the systems they manipulate.